Model parameters: d_model 2560 ffw_size 10240 kv_size 128 n_heads 20 n_layers 34 Megatron-DeepSpeed/pretrain_gpt.py --tensor-model-parallel-size 1 --pipeline-model-parallel-size 1 --num-layers 34 --hidden-size 2560 --num-attention-heads 20 --kv-channels 128 --ffn-hidden-size 10240 --seq-length 2048 --max-position-embeddings 2048 --micro-batch-size 2 --global-batch-size 512 --train-samples 26_855_469 --vocab-file gpt2/vocab.json --merge-file gpt2/merges.txt --clip-grad 1.0 --kill-switch-path kill-switch-2b855brealtasky --bf16 --optimizer adam --adam-beta1 0.9 --adam-beta2 0.95 --adam-eps 1e-8 --lr 2e-4 --min-lr 2e-5 --lr-decay-style cosine --lr-decay-samples 26_855_469 --lr-warmup-samples 268_555 --clip-grad 1.0 --weight-decay 1e-1 --log-interval 10 --save-interval 10000 --eval-interval 100 --eval-iters 100 --tensorboard-dir tensorboard_2b855brealtasky --tensorboard-queue-size 5 --log-timers-to-tensorboard --log-batch-size-to-tensorboard --log-validation-ppl-to-tensorboard --save checkpoints_2b855brealtasky --load checkpoints_2b855brealtasky --train-weighted-split-paths-path trainrealtasky.txt --valid-weighted-split-paths-path valc4oscar.txt --data-impl mmap --deepspeed --deepspeed_config ds_configs/2295714.json --zero-stage 0 START 2295714: Sun Dec 25 08:56:45 EET 2022 0: 0: 0: ======================= ROCm System Management Interface ======================= 0: ================================= Concise Info ================================= 0: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 0: 0 45.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 0: 1 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 0: 2 40.0c 82.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 0: 3 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 0: 4 40.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 0: 5 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 0: 6 40.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 0: 7 40.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 0: ================================================================================ 0: ============================= End of ROCm SMI Log ============================== 16: 16: 16: ======================= ROCm System Management Interface ======================= 16: ================================= Concise Info ================================= 16: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 16: 0 45.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 16: 1 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 16: 2 40.0c 84.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 16: 3 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 16: 4 46.0c 82.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 16: 5 52.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 16: 6 42.0c 93.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 16: 7 42.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 16: ================================================================================ 16: ============================= End of ROCm SMI Log ============================== 31: 31: 31: ======================= ROCm System Management Interface ======================= 31: ================================= Concise Info ================================= 31: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 31: 0 48.0c 95.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 31: 1 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 31: 2 42.0c 81.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 31: 3 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 31: 4 46.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 31: 5 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 31: 6 40.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 31: 7 36.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 31: ================================================================================ 31: ============================= End of ROCm SMI Log ============================== 8: 8: 8: ======================= ROCm System Management Interface ======================= 8: ================================= Concise Info ================================= 8: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 8: 0 44.0c 95.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 8: 1 49.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 8: 2 43.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 8: 3 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 8: 4 43.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 8: 5 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 8: 6 40.0c 93.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 8: 7 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 8: ================================================================================ 8: ============================= End of ROCm SMI Log ============================== 29: 29: 29: ======================= ROCm System Management Interface ======================= 29: ================================= Concise Info ================================= 29: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 29: 0 46.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 29: 1 48.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 29: 2 44.0c 85.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 29: 3 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 29: 4 40.0c 84.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 29: 5 50.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 29: 6 44.0c 83.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 29: 7 42.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 29: ================================================================================ 29: ============================= End of ROCm SMI Log ============================== 4: 4: 4: ======================= ROCm System Management Interface ======================= 4: ================================= Concise Info ================================= 4: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 4: 0 43.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 4: 1 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 4: 2 44.0c 94.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 4: 3 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 4: 4 47.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 4: 5 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 4: 6 44.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 4: 7 42.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 4: ================================================================================ 4: ============================= End of ROCm SMI Log ============================== 7: 7: 7: ======================= ROCm System Management Interface ======================= 7: ================================= Concise Info ================================= 7: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 7: 0 45.0c 97.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 7: 1 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 7: 2 44.0c 82.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 7: 3 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 7: 4 43.0c 86.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 7: 5 51.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 7: 6 34.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 7: 7 42.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 7: ================================================================================ 7: ============================= End of ROCm SMI Log ============================== 9: 9: 9: ======================= ROCm System Management Interface ======================= 9: ================================= Concise Info ================================= 9: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 9: 0 47.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 9: 1 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 9: 2 45.0c 96.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 9: 3 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 9: 4 45.0c 94.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 9: 5 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 9: 6 42.0c 88.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 9: 7 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 9: ================================================================================ 9: ============================= End of ROCm SMI Log ============================== 18: 18: 18: ======================= ROCm System Management Interface ======================= 18: ================================= Concise Info ================================= 18: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 18: 0 45.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 18: 1 49.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 18: 2 37.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 18: 3 51.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 18: 4 41.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 18: 5 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 18: 6 39.0c 86.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 18: 7 41.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 18: ================================================================================ 18: ============================= End of ROCm SMI Log ============================== 5: 5: 5: ======================= ROCm System Management Interface ======================= 5: ================================= Concise Info ================================= 5: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 5: 0 49.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 5: 1 48.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 5: 2 37.0c 86.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 5: 3 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 5: 4 45.0c 94.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 5: 5 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 5: 6 42.0c 85.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 5: 7 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 5: ================================================================================ 5: ============================= End of ROCm SMI Log ============================== 3: 3: 3: ======================= ROCm System Management Interface ======================= 3: ================================= Concise Info ================================= 3: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 3: 0 41.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 3: 1 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 3: 2 46.0c 85.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 3: 3 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 3: 4 40.0c 83.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 3: 5 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 3: 6 44.0c 86.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 3: 7 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 3: ================================================================================ 3: ============================= End of ROCm SMI Log ============================== 30: 30: 30: ======================= ROCm System Management Interface ======================= 30: ================================= Concise Info ================================= 30: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 30: 0 56.0c 94.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 30: 1 49.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 30: 2 39.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 30: 3 41.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 30: 4 47.0c 94.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 30: 5 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 30: 6 46.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 30: 7 53.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 30: ================================================================================ 30: ============================= End of ROCm SMI Log ============================== 21: 21: 21: ======================= ROCm System Management Interface ======================= 21: ================================= Concise Info ================================= 21: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 21: 0 49.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 21: 1 48.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 21: 2 43.0c 81.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 21: 3 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 21: 4 42.0c 81.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 21: 5 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 21: 6 41.0c 78.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 21: 7 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 21: ================================================================================ 21: ============================= End of ROCm SMI Log ============================== 19: 19: 19: ======================= ROCm System Management Interface ======================= 19: ================================= Concise Info ================================= 19: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 19: 0 47.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 19: 1 48.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 19: 2 47.0c 84.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 19: 3 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 19: 4 36.0c 83.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 19: 5 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 19: 6 35.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 19: 7 42.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 19: ================================================================================ 19: ============================= End of ROCm SMI Log ============================== 13: 13: 13: ======================= ROCm System Management Interface ======================= 13: ================================= Concise Info ================================= 13: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 13: 0 49.0c 96.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 13: 1 56.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 13: 2 43.0c 81.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 13: 3 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 13: 4 40.0c 86.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 13: 5 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 13: 6 42.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 13: 7 41.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 13: ================================================================================ 13: ============================= End of ROCm SMI Log ============================== 24: 24: 24: ======================= ROCm System Management Interface ======================= 24: ================================= Concise Info ================================= 24: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 24: 0 45.0c 95.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 24: 1 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 24: 2 40.0c 81.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 24: 3 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 24: 4 46.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 24: 5 52.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 24: 6 40.0c 83.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 24: 7 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 24: ================================================================================ 24: ============================= End of ROCm SMI Log ============================== 28: 28: 28: ======================= ROCm System Management Interface ======================= 28: ================================= Concise Info ================================= 28: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 28: 0 44.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 28: 1 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 28: 2 45.0c 83.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 28: 3 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 28: 4 46.0c 82.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 28: 5 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 28: 6 45.0c 88.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 28: 7 41.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 28: ================================================================================ 28: ============================= End of ROCm SMI Log ============================== 11: 11: 11: ======================= ROCm System Management Interface ======================= 11: ================================= Concise Info ================================= 11: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 11: 0 45.0c 95.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 11: 1 50.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 11: 2 43.0c 85.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 11: 3 48.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 11: 4 44.0c 83.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 11: 5 49.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 11: 6 39.0c 88.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 11: 7 42.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 11: ================================================================================ 11: ============================= End of ROCm SMI Log ============================== 14: 14: 14: ======================= ROCm System Management Interface ======================= 14: ================================= Concise Info ================================= 14: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 14: 0 49.0c 88.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 14: 1 48.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 14: 2 41.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 14: 3 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 14: 4 45.0c 88.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 14: 5 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 14: 6 38.0c 95.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 14: 7 42.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 14: ================================================================================ 14: ============================= End of ROCm SMI Log ============================== 12: 12: 12: ======================= ROCm System Management Interface ======================= 12: ================================= Concise Info ================================= 12: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 12: 0 46.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 12: 1 48.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 12: 2 42.0c 86.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 12: 3 50.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 12: 4 46.0c 85.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 12: 5 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 12: 6 38.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 12: 7 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 12: ================================================================================ 12: ============================= End of ROCm SMI Log ============================== 17: 17: 17: ======================= ROCm System Management Interface ======================= 17: ================================= Concise Info ================================= 17: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 17: 0 43.0c 96.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 17: 1 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 17: 2 45.0c 88.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 17: 3 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 17: 4 42.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 17: 5 48.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 17: 6 39.0c 85.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 17: 7 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 17: ================================================================================ 17: ============================= End of ROCm SMI Log ============================== 25: 25: 25: ======================= ROCm System Management Interface ======================= 25: ================================= Concise Info ================================= 25: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 25: 0 48.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 25: 1 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 25: 2 41.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 25: 3 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 25: 4 41.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 25: 5 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 25: 6 43.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 25: 7 39.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 25: ================================================================================ 25: ============================= End of ROCm SMI Log ============================== 23: 23: 23: ======================= ROCm System Management Interface ======================= 23: ================================= Concise Info ================================= 23: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 23: 0 50.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 23: 1 50.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 23: 2 44.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 23: 3 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 23: 4 38.0c 83.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 23: 5 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 23: 6 33.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 23: 7 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 23: ================================================================================ 23: ============================= End of ROCm SMI Log ============================== 26: 26: 26: ======================= ROCm System Management Interface ======================= 26: ================================= Concise Info ================================= 26: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 26: 0 47.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 26: 1 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 26: 2 38.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 26: 3 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 26: 4 48.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 26: 5 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 26: 6 41.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 26: 7 40.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 26: ================================================================================ 26: ============================= End of ROCm SMI Log ============================== 20: 20: 20: ======================= ROCm System Management Interface ======================= 20: ================================= Concise Info ================================= 20: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 20: 0 42.0c 95.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 20: 1 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 20: 2 39.0c 85.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 20: 3 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 20: 4 44.0c 85.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 20: 5 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 20: 6 39.0c 88.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 20: 7 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 20: ================================================================================ 20: ============================= End of ROCm SMI Log ============================== 10: 10: 10: ======================= ROCm System Management Interface ======================= 10: ================================= Concise Info ================================= 10: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 10: 0 47.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 10: 1 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 10: 2 40.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 10: 3 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 10: 4 43.0c 81.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 10: 5 50.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 10: 6 40.0c 95.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 10: 7 41.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 10: ================================================================================ 10: ============================= End of ROCm SMI Log ============================== 1: 1: 1: ======================= ROCm System Management Interface ======================= 1: ================================= Concise Info ================================= 1: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 1: 0 47.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 1: 1 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 1: 2 40.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 1: 3 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 1: 4 40.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 1: 5 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 1: 6 40.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 1: 7 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 1: ================================================================================ 1: ============================= End of ROCm SMI Log ============================== 15: 15: 15: ======================= ROCm System Management Interface ======================= 15: ================================= Concise Info ================================= 15: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 15: 0 38.0c 93.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 15: 1 48.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 15: 2 39.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 15: 3 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 15: 4 44.0c 86.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 15: 5 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 15: 6 37.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 15: 7 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 15: ================================================================================ 15: ============================= End of ROCm SMI Log ============================== 22: 22: 22: ======================= ROCm System Management Interface ======================= 22: ================================= Concise Info ================================= 22: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 22: 0 48.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 22: 1 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 22: 2 36.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 22: 3 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 22: 4 43.0c 86.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 22: 5 46.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 22: 6 37.0c 88.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 22: 7 50.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 22: ================================================================================ 22: ============================= End of ROCm SMI Log ============================== 27: 27: 27: ======================= ROCm System Management Interface ======================= 27: ================================= Concise Info ================================= 27: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 27: 0 46.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 27: 1 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 27: 2 46.0c 84.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 27: 3 39.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 27: 4 43.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 27: 5 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 27: 6 45.0c 83.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 27: 7 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 27: ================================================================================ 27: ============================= End of ROCm SMI Log ============================== 6: 6: 6: ======================= ROCm System Management Interface ======================= 6: ================================= Concise Info ================================= 6: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 6: 0 48.0c 84.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 6: 1 45.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 6: 2 42.0c 88.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 6: 3 44.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 6: 4 46.0c 87.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 6: 5 42.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 6: 6 38.0c 90.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 6: 7 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 6: ================================================================================ 6: ============================= End of ROCm SMI Log ============================== 2: 2: 2: ======================= ROCm System Management Interface ======================= 2: ================================= Concise Info ================================= 2: GPU Temp AvgPwr SCLK MCLK Fan Perf PwrCap VRAM% GPU% 2: 0 49.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 2: 1 49.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 2: 2 43.0c 92.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 2: 3 43.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 2: 4 44.0c 91.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 2: 5 40.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 2: 6 40.0c 89.0W 800Mhz 1600Mhz 0% auto 560.0W 0% 0% 2: 7 47.0c N/A 800Mhz 1600Mhz 0% auto 0.0W 0% 0% 2: ================================================================================ 2: ============================= End of ROCm SMI Log ============================== 28: Launching on nid007102 (28/32), master nid005584 port 9999, GPUs 8, CUDA: True 29: Launching on nid007103 (29/32), master nid005584 port 9999, GPUs 8, CUDA: True 7: Launching on nid006455 (7/32), master nid005584 port 9999, GPUs 8, CUDA: True 17: Launching on nid006629 (17/32), master nid005584 port 9999, GPUs 8, CUDA: True 12: Launching on nid006609 (12/32), master nid005584 port 9999, GPUs 8, CUDA: True 8: Launching on nid006488 (8/32), master nid005584 port 9999, GPUs 8, CUDA: True 11: Launching on nid006593 (11/32), master nid005584 port 9999, GPUs 8, CUDA: True 23: Launching on nid006877 (23/32), master nid005584 port 9999, GPUs 8, CUDA: True 31: Launching on nid007220 (31/32), master nid005584 port 9999, GPUs 8, CUDA: True 1: Launching on nid005702 (1/32), master nid005584 port 9999, GPUs 8, CUDA: True 10: Launching on nid006592 (10/32), master nid005584 port 9999, GPUs 8, CUDA: True 15: Launching on nid006624 (15/32), master nid005584 port 9999, GPUs 8, CUDA: True 4: Launching on nid006342 (4/32), master nid005584 port 9999, GPUs 8, CUDA: True 2: Launching on nid005903 (2/32), master nid005584 port 9999, GPUs 8, CUDA: True 3: Launching on nid006339 (3/32), master nid005584 port 9999, GPUs 8, CUDA: True 19: Launching on nid006677 (19/32), master nid005584 port 9999, GPUs 8, CUDA: True 27: Launching on nid007098 (27/32), master nid005584 port 9999, GPUs 8, CUDA: True 26: Launching on nid007092 (26/32), master nid005584 port 9999, GPUs 8, CUDA: True 14: Launching on nid006619 (14/32), master nid005584 port 9999, GPUs 8, CUDA: True 25: Launching on nid006943 (25/32), master nid005584 port 9999, GPUs 8, CUDA: True 16: Launching on nid006625 (16/32), master nid005584 port 9999, GPUs 8, CUDA: True 18: Launching on nid006635 (18/32), master nid005584 port 9999, GPUs 8, CUDA: True 0: Launching on nid005584 (0/32), master nid005584 port 9999, GPUs 8, CUDA: True 20: Launching on nid006678 (20/32), master nid005584 port 9999, GPUs 8, CUDA: True 9: Launching on nid006585 (9/32), master nid005584 port 9999, GPUs 8, CUDA: True 30: Launching on nid007114 (30/32), master nid005584 port 9999, GPUs 8, CUDA: True 5: Launching on nid006344 (5/32), master nid005584 port 9999, GPUs 8, CUDA: True 24: Launching on nid006940 (24/32), master nid005584 port 9999, GPUs 8, CUDA: True 21: Launching on nid006701 (21/32), master nid005584 port 9999, GPUs 8, CUDA: True 22: Launching on nid006820 (22/32), master nid005584 port 9999, GPUs 8, CUDA: True 13: Launching on nid006617 (13/32), master nid005584 port 9999, GPUs 8, CUDA: True 6: Launching on nid006348 (6/32), master nid005584 port 9999, GPUs 8, CUDA: True 0: using world size: 256, data-parallel-size: 256, tensor-model-parallel size: 1, pipeline-model-parallel size: 1 0: accumulate and all-reduce gradients in fp32 for bfloat16 data type. 0: using torch.bfloat16 for parameters ... 0: ------------------------ arguments ------------------------ 0: abort_on_unmet_fused_kernel_constraints ......... False 0: accumulate_allreduce_grads_in_fp32 .............. True 0: adam_beta1 ...................................... 0.9 0: adam_beta2 ...................................... 0.95 0: adam_eps ........................................ 1e-08 0: adlr_autoresume ................................. False 0: adlr_autoresume_interval ........................ 1000 0: apply_query_key_layer_scaling ................... True 0: apply_residual_connection_post_layernorm ........ False 0: attention_dropout ............................... 0.1 0: attention_softmax_in_fp32 ....................... False 0: bert_binary_head ................................ True 0: bert_load ....................................... None 0: bf16 ............................................ True 0: bias_dropout_fusion ............................. True 0: bias_gelu_fusion ................................ True 0: biencoder_projection_dim ........................ 0 0: biencoder_shared_query_context_model ............ False 0: block_data_path ................................. None 0: checkpoint_activations .......................... False 0: checkpoint_in_cpu ............................... False 0: checkpoint_num_layers ........................... 1 0: clip_grad ....................................... 1.0 0: codecarbon_dir .................................. None 0: consumed_train_samples .......................... 0 0: consumed_train_tokens ........................... 0 0: consumed_valid_samples .......................... 0 0: contigious_checkpointing ........................ False 0: cpu_optimizer ................................... False 0: cpu_torch_adam .................................. False 0: curriculum_learning ............................. False 0: data_impl ....................................... mmap 0: data_parallel_size .............................. 256 0: data_path ....................................... None 0: dataloader_type ................................. single 0: DDP_impl ........................................ local 0: decoder_seq_length .............................. None 0: deepscale ....................................... False 0: deepscale_config ................................ None 0: deepspeed ....................................... True 0: deepspeed_activation_checkpointing .............. False 0: deepspeed_config ................................ ds_configs/2295714.json 0: deepspeed_mpi ................................... False 0: distribute_checkpointed_activations ............. False 0: distributed_backend ............................. nccl 0: embed_layernorm ................................. False 0: embedding_path .................................. None 0: encoder_seq_length .............................. 2048 0: eod_mask_loss ................................... False 0: eval_interval ................................... 100 0: eval_iters ...................................... 100 0: eval_only ....................................... None 0: evidence_data_path .............................. None 0: exit_duration_in_mins ........................... None 0: exit_interval ................................... None 0: ffn_hidden_size ................................. 10240 0: finetune ........................................ False 0: fp16 ............................................ False 0: fp16_lm_cross_entropy ........................... False 0: fp32_residual_connection ........................ False 0: gigaflos_no_embeds .............................. 0 0: global_batch_size ............................... 512 0: glu_activation .................................. None 0: hidden_dropout .................................. 0.1 0: hidden_size ..................................... 2560 0: hysteresis ...................................... 2 0: ict_head_size ................................... None 0: ict_load ........................................ None 0: img_dim ......................................... 224 0: indexer_batch_size .............................. 128 0: indexer_log_interval ............................ 1000 0: inference ....................................... False 0: init_method_std ................................. 0.02 0: init_method_xavier_uniform ...................... False 0: initial_loss_scale .............................. 4294967296 0: kill_switch_path ................................ kill-switch-2b855brealtasky 0: kv_channels ..................................... 128 0: layer_norm_fusion ............................... True 0: layernorm_epsilon ............................... 1e-05 0: lazy_mpu_init ................................... None 0: load ............................................ checkpoints_2b855brealtasky 0: local_rank ...................................... None 0: log_batch_size_to_tensorboard ................... True 0: log_interval .................................... 10 0: log_learning_rate_to_tensorboard ................ True 0: log_level ....................................... None 0: log_level_replica ............................... None 0: log_loss_scale_to_tensorboard ................... True 0: log_num_zeros_in_grad ........................... False 0: log_params_norm ................................. False 0: log_path ........................................ None 0: log_timers_to_tensorboard ....................... True 0: log_validation_ppl_to_tensorboard ............... True 0: loss_on_targets_only ............................ False 0: loss_scale ...................................... None 0: loss_scale_window ............................... 1000 0: lr .............................................. 0.0002 0: lr_decay_iters .................................. None 0: lr_decay_samples ................................ 26855469 0: lr_decay_style .................................. cosine 0: lr_decay_tokens ................................. None 0: lr_warmup_fraction .............................. None 0: lr_warmup_iters ................................. 0 0: lr_warmup_samples ............................... 268555 0: make_vocab_size_divisible_by .................... 128 0: mask_prob ....................................... 0.15 0: masked_softmax_fusion ........................... True 0: max_position_embeddings ......................... 2048 0: mean_noise_span_length .......................... None 0: memory_centric_tiled_linear ..................... False 0: merge_file ...................................... gpt2/merges.txt 0: micro_batch_size ................................ 2 0: min_loss_scale .................................. 1.0 0: min_lr .......................................... 2e-05 0: mmap_warmup ..................................... False 0: no_load_optim ................................... None 0: no_load_rng ..................................... None 0: no_save_optim ................................... None 0: no_save_rng ..................................... None 0: noise_density ................................... None 0: num_attention_heads ............................. 20 0: num_channels .................................... 3 0: num_classes ..................................... 1000 0: num_layers ...................................... 34 0: num_layers_per_virtual_pipeline_stage ........... None 0: num_workers ..................................... 2 0: onnx_safe ....................................... None 0: openai_gelu ..................................... False 0: optimizer ....................................... adam 0: optimizer_fusion ................................ True 0: override_lr_scheduler ........................... False 0: pad_vocab_size_to ............................... None 0: params_dtype .................................... torch.bfloat16 0: partition_activations ........................... False 0: patch_dim ....................................... 16 0: pipeline_model_parallel_size .................... 1 0: position_embedding_type ......................... PositionEmbeddingType.absolute 0: pp_partition_method ............................. None 0: profile_backward ................................ False 0: query_in_block_prob ............................. 0.1 0: rampup_batch_size ............................... None 0: rank ............................................ 0 0: remote_device ................................... none 0: reset_attention_mask ............................ False 0: reset_position_ids .............................. False 0: retriever_report_topk_accuracies ................ [] 0: retriever_score_scaling ......................... False 0: retriever_seq_length ............................ 256 0: reweight_loss_based_on_position_frequency ....... False 0: sample_rate ..................................... 1.0 0: save ............................................ checkpoints_2b855brealtasky 0: save_interval ................................... 10000 0: scatter_gather_tensors_in_pipeline .............. True 0: scattered_embeddings ............................ False 0: seed ............................................ 1234 0: seq_length ...................................... 2048 0: sgd_momentum .................................... 0.9 0: short_seq_prob .................................. 0.1 0: skip_train_iteration_range ...................... None 0: split ........................................... None 0: split_transformers .............................. False 0: sync_tp_duplicated_parameters ................... False 0: synchronize_each_layer .......................... False 0: tensor_model_parallel_size ...................... 1 0: tensorboard_dir ................................. tensorboard_2b855brealtasky 0: tensorboard_log_interval ........................ 1 0: tensorboard_queue_size .......................... 5 0: test_weighted_split_paths ....................... None 0: test_weighted_split_paths_path .................. None 0: tile_factor ..................................... 1 0: titles_data_path ................................ None 0: tokenizer_name_or_path .......................... None 0: tokenizer_type .................................. GPT2BPETokenizer 0: train_iters ..................................... None 0: train_samples ................................... 26855469 0: train_tokens .................................... None 0: train_weighted_split_names ...................... ['train'] 0: train_weighted_split_paths ...................... [['/scratch/project_462000119/data/realtasky/gpt2tok_realtasky_text_document']] 0: train_weighted_split_paths_path ................. None 0: train_weighted_split_splits ..................... [['0:1']] 0: train_weighted_split_weights .................... [['1.0']] 0: universal_checkpoint ............................ False 0: use_bnb_optimizer ............................... False 0: use_checkpoint_lr_scheduler ..................... False 0: use_contiguous_buffers_in_ddp ................... True 0: use_cpu_initialization .......................... None 0: use_one_sent_docs ............................... False 0: use_pin_memory .................................. False 0: valid_num_workers ............................... 2 0: valid_weighted_split_names ...................... ['validation_c4', 'validation_oscar'] 0: valid_weighted_split_paths ...................... [['/scratch/project_462000119/data/c4_validation/gpt2tok_c4validation_rerun_text_document'], ['/scratch/project_462000119/data/oscar_validation/gpt2tok_oscarvalidation_text_document']] 0: valid_weighted_split_paths_path ................. None 0: valid_weighted_split_splits ..................... [['0:1'], ['0:1']] 0: valid_weighted_split_weights .................... [['1.0'], ['1.0']] 0: virtual_pipeline_model_parallel_size ............ None 0: vocab_extra_ids ................................. 0 0: vocab_file ...................................... gpt2/vocab.json 0: weight_decay .................................... 0.1 0: world_size ...................................... 256 0: zero_allgather_bucket_size ...................... 0.0 0: zero_contigious_gradients ....................... False 0: zero_reduce_bucket_size ......................... 0.0 0: zero_reduce_scatter ............................. False 0: zero_stage ...................................... 0 0: -------------------- end of arguments --------------------- 0: setting number of micro-batches to constant 1 0: > building GPT2BPETokenizer tokenizer ... 0: > padded vocab (size: 50257) with 47 dummy tokens (new size: 50304) 0: DeepSpeed general environment info: 0: torch install path ............... ['/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/venv/lib/python3.9/site-packages/torch'] 0: torch version .................... 1.13.0+rocm5.2 0: torch cuda version ............... None 0: torch hip version ................ 5.2.21151-afdc89f8 0: nvcc version ..................... None 0: deepspeed install path ........... ['/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/venv/lib/python3.9/site-packages/deepspeed'] 0: deepspeed info ................... 0.7.5, unknown, unknown 0: deepspeed wheel compiled w. ...... torch 1.13, hip 5.1 31: > setting tensorboard ... 0: **** Git info for Megatron: git_hash=unknown git_branch=unknown **** 0: > initializing torch distributed ... 0: [2022-12-25 08:58:00,792] [INFO] [comm.py:633:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl 0: > initializing tensor model parallel with size 1 0: > initializing pipeline model parallel with size 1 0: > setting random seeds to 1234 ... 0: > initializing model parallel cuda seeds on global rank 0, model parallel rank 0, and data parallel rank 0 with model parallel seed: 3952 and data parallel seed: 1234 0: > compiling dataset index builder ... 0: make: Entering directory '/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/data' 0: make: Nothing to be done for 'default'. 0: make: Leaving directory '/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/data' 0: >>> done with dataset index builder. Compilation time: 0.108 seconds 0: > compiling and loading fused kernels ... 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax.cpp -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax_hip.cpp [skipped, already hipified] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax_hip.h [skipped, already hipified] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/compat.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/compat.h [skipped, no changes] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/type_shim.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/type_shim.h [skipped, no changes] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax_cuda.cu -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax_hip.hip [skipped, already hipified] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/type_shim.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/type_shim.h [skipped, no changes] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/compat.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/compat.h [skipped, no changes] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax_hip.h [skipped, already hipified] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax_hip.h [skipped, already hipified] 0: Total number of unsupported CUDA function calls: 0 0: 0: 0: Total number of replaced kernel launches: 87 0: [1/1] c++ scaled_upper_triang_masked_softmax_hip.cuda.o scaled_upper_triang_masked_softmax_hip.o -shared -L/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/venv/lib/python3.9/site-packages/torch/lib -lc10 -lc10_hip -ltorch_cpu -ltorch_hip -ltorch -ltorch_python -L/pfs/lustrep2/projappl/project_462000125/samantao-public/rocm/rocm-5.2.3/lib -lamdhip64 -o scaled_upper_triang_masked_softmax_cuda.so 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax.cpp -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax_hip.cpp [skipped, already hipified] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax_cuda.cu -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax_hip.hip [skipped, already hipified] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/type_shim.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/type_shim.h [skipped, no changes] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/compat.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/compat.h [skipped, no changes] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax_hip.h [skipped, already hipified] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax_hip.h [skipped, already hipified] 0: Total number of unsupported CUDA function calls: 0 0: 0: 0: Total number of replaced kernel launches: 63 0: [1/1] c++ scaled_masked_softmax_hip.cuda.o scaled_masked_softmax_hip.o -shared -L/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/venv/lib/python3.9/site-packages/torch/lib -lc10 -lc10_hip -ltorch_cpu -ltorch_hip -ltorch -ltorch_python -L/pfs/lustrep2/projappl/project_462000125/samantao-public/rocm/rocm-5.2.3/lib -lamdhip64 -o scaled_masked_softmax_cuda.so 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/layer_norm_cuda.cpp -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/layer_norm_cuda.cpp [skipped, no changes] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/layer_norm_cuda_kernel.cu -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/layer_norm_hip_kernel.hip [skipped, already hipified] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/type_shim.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/type_shim.h [skipped, no changes] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/compat.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/compat.h [skipped, no changes] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_upper_triang_masked_softmax_hip.h [skipped, already hipified] 0: /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax.h -> /pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/Megatron-DeepSpeed/megatron/fused_kernels/scaled_masked_softmax_hip.h [skipped, already hipified] 0: Total number of unsupported CUDA function calls: 0 0: 0: 0: Total number of replaced kernel launches: 67 0: ninja: no work to do. 0: >>> done with compiling and loading fused kernels. Compilation time: 20.962 seconds 0: time to initialize megatron (seconds): 26.439 0: [after megatron is initialized] datetime: 2022-12-25 08:58:34 0: building GPT model ... 0: [2022-12-25 08:58:34,569] [INFO] [utils.py:827:see_memory_usage] Before Building Model 0: [2022-12-25 08:58:34,570] [INFO] [utils.py:828:see_memory_usage] MA 0.0 GB Max_MA 0.0 GB CA 0.0 GB Max_CA 0 GB 0: [2022-12-25 08:58:34,570] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 55.35 GB, percent = 11.0% 0: SEED_LAYERS=False BASE_SEED=1234 SEED_FN=None 0: Using topology: {ProcessCoord(pipe=0, data=0, model=0): 0, ProcessCoord(pipe=0, data=1, model=0): 1, ProcessCoord(pipe=0, data=2, model=0): 2, ProcessCoord(pipe=0, data=3, model=0): 3, ProcessCoord(pipe=0, data=4, model=0): 4, ProcessCoord(pipe=0, data=5, model=0): 5, ProcessCoord(pipe=0, data=6, model=0): 6, ProcessCoord(pipe=0, data=7, model=0): 7, ProcessCoord(pipe=0, data=8, model=0): 8, ProcessCoord(pipe=0, data=9, model=0): 9, ProcessCoord(pipe=0, data=10, model=0): 10, ProcessCoord(pipe=0, data=11, model=0): 11, ProcessCoord(pipe=0, data=12, model=0): 12, ProcessCoord(pipe=0, data=13, model=0): 13, ProcessCoord(pipe=0, data=14, model=0): 14, ProcessCoord(pipe=0, data=15, model=0): 15, ProcessCoord(pipe=0, data=16, model=0): 16, ProcessCoord(pipe=0, data=17, model=0): 17, ProcessCoord(pipe=0, data=18, model=0): 18, ProcessCoord(pipe=0, data=19, model=0): 19, ProcessCoord(pipe=0, data=20, model=0): 20, ProcessCoord(pipe=0, data=21, model=0): 21, ProcessCoord(pipe=0, data=22, model=0): 22, ProcessCoord(pi 0: pe=0, data=23, model=0): 23, ProcessCoord(pipe=0, data=24, model=0): 24, ProcessCoord(pipe=0, data=25, model=0): 25, ProcessCoord(pipe=0, data=26, model=0): 26, ProcessCoord(pipe=0, data=27, model=0): 27, ProcessCoord(pipe=0, data=28, model=0): 28, ProcessCoord(pipe=0, data=29, model=0): 29, ProcessCoord(pipe=0, data=30, model=0): 30, ProcessCoord(pipe=0, data=31, model=0): 31, ProcessCoord(pipe=0, data=32, model=0): 32, ProcessCoord(pipe=0, data=33, model=0): 33, ProcessCoord(pipe=0, data=34, model=0): 34, ProcessCoord(pipe=0, data=35, model=0): 35, ProcessCoord(pipe=0, data=36, model=0): 36, ProcessCoord(pipe=0, data=37, model=0): 37, ProcessCoord(pipe=0, data=38, model=0): 38, ProcessCoord(pipe=0, data=39, model=0): 39, ProcessCoord(pipe=0, data=40, model=0): 40, ProcessCoord(pipe=0, data=41, model=0): 41, ProcessCoord(pipe=0, data=42, model=0): 42, ProcessCoord(pipe=0, data=43, model=0): 43, ProcessCoord(pipe=0, data=44, model=0): 44, ProcessCoord(pipe=0, data=45, model=0): 45, ProcessCoord(pipe=0, data=4 0: 6, model=0): 46, ProcessCoord(pipe=0, data=47, model=0): 47, ProcessCoord(pipe=0, data=48, model=0): 48, ProcessCoord(pipe=0, data=49, model=0): 49, ProcessCoord(pipe=0, data=50, model=0): 50, ProcessCoord(pipe=0, data=51, model=0): 51, ProcessCoord(pipe=0, data=52, model=0): 52, ProcessCoord(pipe=0, data=53, model=0): 53, ProcessCoord(pipe=0, data=54, model=0): 54, ProcessCoord(pipe=0, data=55, model=0): 55, ProcessCoord(pipe=0, data=56, model=0): 56, ProcessCoord(pipe=0, data=57, model=0): 57, ProcessCoord(pipe=0, data=58, model=0): 58, ProcessCoord(pipe=0, data=59, model=0): 59, ProcessCoord(pipe=0, data=60, model=0): 60, ProcessCoord(pipe=0, data=61, model=0): 61, ProcessCoord(pipe=0, data=62, model=0): 62, ProcessCoord(pipe=0, data=63, model=0): 63, ProcessCoord(pipe=0, data=64, model=0): 64, ProcessCoord(pipe=0, data=65, model=0): 65, ProcessCoord(pipe=0, data=66, model=0): 66, ProcessCoord(pipe=0, data=67, model=0): 67, ProcessCoord(pipe=0, data=68, model=0): 68, ProcessCoord(pipe=0, data=69, model=0): 0: 69, ProcessCoord(pipe=0, data=70, model=0): 70, ProcessCoord(pipe=0, data=71, model=0): 71, ProcessCoord(pipe=0, data=72, model=0): 72, ProcessCoord(pipe=0, data=73, model=0): 73, ProcessCoord(pipe=0, data=74, model=0): 74, ProcessCoord(pipe=0, data=75, model=0): 75, ProcessCoord(pipe=0, data=76, model=0): 76, ProcessCoord(pipe=0, data=77, model=0): 77, ProcessCoord(pipe=0, data=78, model=0): 78, ProcessCoord(pipe=0, data=79, model=0): 79, ProcessCoord(pipe=0, data=80, model=0): 80, ProcessCoord(pipe=0, data=81, model=0): 81, ProcessCoord(pipe=0, data=82, model=0): 82, ProcessCoord(pipe=0, data=83, model=0): 83, ProcessCoord(pipe=0, data=84, model=0): 84, ProcessCoord(pipe=0, data=85, model=0): 85, ProcessCoord(pipe=0, data=86, model=0): 86, ProcessCoord(pipe=0, data=87, model=0): 87, ProcessCoord(pipe=0, data=88, model=0): 88, ProcessCoord(pipe=0, data=89, model=0): 89, ProcessCoord(pipe=0, data=90, model=0): 90, ProcessCoord(pipe=0, data=91, model=0): 91, ProcessCoord(pipe=0, data=92, model=0): 92, Process 0: Coord(pipe=0, data=93, model=0): 93, ProcessCoord(pipe=0, data=94, model=0): 94, ProcessCoord(pipe=0, data=95, model=0): 95, ProcessCoord(pipe=0, data=96, model=0): 96, ProcessCoord(pipe=0, data=97, model=0): 97, ProcessCoord(pipe=0, data=98, model=0): 98, ProcessCoord(pipe=0, data=99, model=0): 99, ProcessCoord(pipe=0, data=100, model=0): 100, ProcessCoord(pipe=0, data=101, model=0): 101, ProcessCoord(pipe=0, data=102, model=0): 102, ProcessCoord(pipe=0, data=103, model=0): 103, ProcessCoord(pipe=0, data=104, model=0): 104, ProcessCoord(pipe=0, data=105, model=0): 105, ProcessCoord(pipe=0, data=106, model=0): 106, ProcessCoord(pipe=0, data=107, model=0): 107, ProcessCoord(pipe=0, data=108, model=0): 108, ProcessCoord(pipe=0, data=109, model=0): 109, ProcessCoord(pipe=0, data=110, model=0): 110, ProcessCoord(pipe=0, data=111, model=0): 111, ProcessCoord(pipe=0, data=112, model=0): 112, ProcessCoord(pipe=0, data=113, model=0): 113, ProcessCoord(pipe=0, data=114, model=0): 114, ProcessCoord(pipe=0, data=115, mo 0: del=0): 115, ProcessCoord(pipe=0, data=116, model=0): 116, ProcessCoord(pipe=0, data=117, model=0): 117, ProcessCoord(pipe=0, data=118, model=0): 118, ProcessCoord(pipe=0, data=119, model=0): 119, ProcessCoord(pipe=0, data=120, model=0): 120, ProcessCoord(pipe=0, data=121, model=0): 121, ProcessCoord(pipe=0, data=122, model=0): 122, ProcessCoord(pipe=0, data=123, model=0): 123, ProcessCoord(pipe=0, data=124, model=0): 124, ProcessCoord(pipe=0, data=125, model=0): 125, ProcessCoord(pipe=0, data=126, model=0): 126, ProcessCoord(pipe=0, data=127, model=0): 127, ProcessCoord(pipe=0, data=128, model=0): 128, ProcessCoord(pipe=0, data=129, model=0): 129, ProcessCoord(pipe=0, data=130, model=0): 130, ProcessCoord(pipe=0, data=131, model=0): 131, ProcessCoord(pipe=0, data=132, model=0): 132, ProcessCoord(pipe=0, data=133, model=0): 133, ProcessCoord(pipe=0, data=134, model=0): 134, ProcessCoord(pipe=0, data=135, model=0): 135, ProcessCoord(pipe=0, data=136, model=0): 136, ProcessCoord(pipe=0, data=137, model=0): 137, 0: ProcessCoord(pipe=0, data=138, model=0): 138, ProcessCoord(pipe=0, data=139, model=0): 139, ProcessCoord(pipe=0, data=140, model=0): 140, ProcessCoord(pipe=0, data=141, model=0): 141, ProcessCoord(pipe=0, data=142, model=0): 142, ProcessCoord(pipe=0, data=143, model=0): 143, ProcessCoord(pipe=0, data=144, model=0): 144, ProcessCoord(pipe=0, data=145, model=0): 145, ProcessCoord(pipe=0, data=146, model=0): 146, ProcessCoord(pipe=0, data=147, model=0): 147, ProcessCoord(pipe=0, data=148, model=0): 148, ProcessCoord(pipe=0, data=149, model=0): 149, ProcessCoord(pipe=0, data=150, model=0): 150, ProcessCoord(pipe=0, data=151, model=0): 151, ProcessCoord(pipe=0, data=152, model=0): 152, ProcessCoord(pipe=0, data=153, model=0): 153, ProcessCoord(pipe=0, data=154, model=0): 154, ProcessCoord(pipe=0, data=155, model=0): 155, ProcessCoord(pipe=0, data=156, model=0): 156, ProcessCoord(pipe=0, data=157, model=0): 157, ProcessCoord(pipe=0, data=158, model=0): 158, ProcessCoord(pipe=0, data=159, model=0): 159, ProcessCoor 0: d(pipe=0, data=160, model=0): 160, ProcessCoord(pipe=0, data=161, model=0): 161, ProcessCoord(pipe=0, data=162, model=0): 162, ProcessCoord(pipe=0, data=163, model=0): 163, ProcessCoord(pipe=0, data=164, model=0): 164, ProcessCoord(pipe=0, data=165, model=0): 165, ProcessCoord(pipe=0, data=166, model=0): 166, ProcessCoord(pipe=0, data=167, model=0): 167, ProcessCoord(pipe=0, data=168, model=0): 168, ProcessCoord(pipe=0, data=169, model=0): 169, ProcessCoord(pipe=0, data=170, model=0): 170, ProcessCoord(pipe=0, data=171, model=0): 171, ProcessCoord(pipe=0, data=172, model=0): 172, ProcessCoord(pipe=0, data=173, model=0): 173, ProcessCoord(pipe=0, data=174, model=0): 174, ProcessCoord(pipe=0, data=175, model=0): 175, ProcessCoord(pipe=0, data=176, model=0): 176, ProcessCoord(pipe=0, data=177, model=0): 177, ProcessCoord(pipe=0, data=178, model=0): 178, ProcessCoord(pipe=0, data=179, model=0): 179, ProcessCoord(pipe=0, data=180, model=0): 180, ProcessCoord(pipe=0, data=181, model=0): 181, ProcessCoord(pipe=0, da 0: ta=182, model=0): 182, ProcessCoord(pipe=0, data=183, model=0): 183, ProcessCoord(pipe=0, data=184, model=0): 184, ProcessCoord(pipe=0, data=185, model=0): 185, ProcessCoord(pipe=0, data=186, model=0): 186, ProcessCoord(pipe=0, data=187, model=0): 187, ProcessCoord(pipe=0, data=188, model=0): 188, ProcessCoord(pipe=0, data=189, model=0): 189, ProcessCoord(pipe=0, data=190, model=0): 190, ProcessCoord(pipe=0, data=191, model=0): 191, ProcessCoord(pipe=0, data=192, model=0): 192, ProcessCoord(pipe=0, data=193, model=0): 193, ProcessCoord(pipe=0, data=194, model=0): 194, ProcessCoord(pipe=0, data=195, model=0): 195, ProcessCoord(pipe=0, data=196, model=0): 196, ProcessCoord(pipe=0, data=197, model=0): 197, ProcessCoord(pipe=0, data=198, model=0): 198, ProcessCoord(pipe=0, data=199, model=0): 199, ProcessCoord(pipe=0, data=200, model=0): 200, ProcessCoord(pipe=0, data=201, model=0): 201, ProcessCoord(pipe=0, data=202, model=0): 202, ProcessCoord(pipe=0, data=203, model=0): 203, ProcessCoord(pipe=0, data=204, mode 0: l=0): 204, ProcessCoord(pipe=0, data=205, model=0): 205, ProcessCoord(pipe=0, data=206, model=0): 206, ProcessCoord(pipe=0, data=207, model=0): 207, ProcessCoord(pipe=0, data=208, model=0): 208, ProcessCoord(pipe=0, data=209, model=0): 209, ProcessCoord(pipe=0, data=210, model=0): 210, ProcessCoord(pipe=0, data=211, model=0): 211, ProcessCoord(pipe=0, data=212, model=0): 212, ProcessCoord(pipe=0, data=213, model=0): 213, ProcessCoord(pipe=0, data=214, model=0): 214, ProcessCoord(pipe=0, data=215, model=0): 215, ProcessCoord(pipe=0, data=216, model=0): 216, ProcessCoord(pipe=0, data=217, model=0): 217, ProcessCoord(pipe=0, data=218, model=0): 218, ProcessCoord(pipe=0, data=219, model=0): 219, ProcessCoord(pipe=0, data=220, model=0): 220, ProcessCoord(pipe=0, data=221, model=0): 221, ProcessCoord(pipe=0, data=222, model=0): 222, ProcessCoord(pipe=0, data=223, model=0): 223, ProcessCoord(pipe=0, data=224, model=0): 224, ProcessCoord(pipe=0, data=225, model=0): 225, ProcessCoord(pipe=0, data=226, model=0): 226, P 0: rocessCoord(pipe=0, data=227, model=0): 227, ProcessCoord(pipe=0, data=228, model=0): 228, ProcessCoord(pipe=0, data=229, model=0): 229, ProcessCoord(pipe=0, data=230, model=0): 230, ProcessCoord(pipe=0, data=231, model=0): 231, ProcessCoord(pipe=0, data=232, model=0): 232, ProcessCoord(pipe=0, data=233, model=0): 233, ProcessCoord(pipe=0, data=234, model=0): 234, ProcessCoord(pipe=0, data=235, model=0): 235, ProcessCoord(pipe=0, data=236, model=0): 236, ProcessCoord(pipe=0, data=237, model=0): 237, ProcessCoord(pipe=0, data=238, model=0): 238, ProcessCoord(pipe=0, data=239, model=0): 239, ProcessCoord(pipe=0, data=240, model=0): 240, ProcessCoord(pipe=0, data=241, model=0): 241, ProcessCoord(pipe=0, data=242, model=0): 242, ProcessCoord(pipe=0, data=243, model=0): 243, ProcessCoord(pipe=0, data=244, model=0): 244, ProcessCoord(pipe=0, data=245, model=0): 245, ProcessCoord(pipe=0, data=246, model=0): 246, ProcessCoord(pipe=0, data=247, model=0): 247, ProcessCoord(pipe=0, data=248, model=0): 248, ProcessCoord( 0: pipe=0, data=249, model=0): 249, ProcessCoord(pipe=0, data=250, model=0): 250, ProcessCoord(pipe=0, data=251, model=0): 251, ProcessCoord(pipe=0, data=252, model=0): 252, ProcessCoord(pipe=0, data=253, model=0): 253, ProcessCoord(pipe=0, data=254, model=0): 254, ProcessCoord(pipe=0, data=255, model=0): 255} 0: [2022-12-25 08:58:43,036] [INFO] [module.py:366:_partition_layers] Partitioning pipeline stages with method type:transformer 0: stage=0 layers=41 0: 0: _to_float16 0: 1: EmbeddingPipe 0: 2: 0: 3: ParallelTransformerLayerPipe 0: 4: ParallelTransformerLayerPipe 0: 5: ParallelTransformerLayerPipe 0: 6: ParallelTransformerLayerPipe 0: 7: ParallelTransformerLayerPipe 0: 8: ParallelTransformerLayerPipe 0: 9: ParallelTransformerLayerPipe 0: 10: ParallelTransformerLayerPipe 0: 11: ParallelTransformerLayerPipe 0: 12: ParallelTransformerLayerPipe 0: 13: ParallelTransformerLayerPipe 0: 14: ParallelTransformerLayerPipe 0: 15: ParallelTransformerLayerPipe 0: 16: ParallelTransformerLayerPipe 0: 17: ParallelTransformerLayerPipe 0: 18: ParallelTransformerLayerPipe 0: 19: ParallelTransformerLayerPipe 0: 20: ParallelTransformerLayerPipe 0: 21: ParallelTransformerLayerPipe 0: 22: ParallelTransformerLayerPipe 0: 23: ParallelTransformerLayerPipe 0: 24: ParallelTransformerLayerPipe 0: 25: ParallelTransformerLayerPipe 0: 26: ParallelTransformerLayerPipe 0: 27: ParallelTransformerLayerPipe 0: 28: ParallelTransformerLayerPipe 0: 29: ParallelTransformerLayerPipe 0: 30: ParallelTransformerLayerPipe 0: 31: ParallelTransformerLayerPipe 0: 32: ParallelTransformerLayerPipe 0: 33: ParallelTransformerLayerPipe 0: 34: ParallelTransformerLayerPipe 0: 35: ParallelTransformerLayerPipe 0: 36: ParallelTransformerLayerPipe 0: 37: undo 0: 38: MixedFusedLayerNorm 0: 39: EmbeddingPipe 0: 40: float16_to_fp32 0: loss: CrossEntropy 0: [2022-12-25 08:58:43,299] [INFO] [utils.py:827:see_memory_usage] After Building Model 0: [2022-12-25 08:58:43,300] [INFO] [utils.py:828:see_memory_usage] MA 5.26 GB Max_MA 5.26 GB CA 5.31 GB Max_CA 5 GB 0: [2022-12-25 08:58:43,300] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 55.39 GB, percent = 11.0% 0: setting training iterations to 52452 0: > learning rate decay style: cosine 0: DeepSpeed is enabled. 0: [2022-12-25 08:58:43,303] [INFO] [logging.py:68:log_dist] [Rank 0] DeepSpeed info: version=0.7.5, git-hash=unknown, git-branch=unknown 0: [2022-12-25 08:59:05,478] [INFO] [logging.py:68:log_dist] [Rank 0] DeepSpeed Flops Profiler Enabled: False 0: [2022-12-25 08:59:05,478] [INFO] [logging.py:68:log_dist] [Rank 0] Removing param_group that has no 'params' in the client Optimizer 0: [2022-12-25 08:59:05,478] [INFO] [logging.py:68:log_dist] [Rank 0] Using client Optimizer as basic optimizer 0: [2022-12-25 08:59:05,496] [INFO] [logging.py:68:log_dist] [Rank 0] DeepSpeed Basic Optimizer = FusedAdam 0: [2022-12-25 08:59:05,496] [INFO] [logging.py:68:log_dist] [Rank 0] Creating BF16 optimizer 0: [2022-12-25 08:59:05,601] [INFO] [utils.py:827:see_memory_usage] begin bf16_optimizer 0: [2022-12-25 08:59:05,601] [INFO] [utils.py:828:see_memory_usage] MA 5.25 GB Max_MA 5.27 GB CA 5.32 GB Max_CA 5 GB 0: [2022-12-25 08:59:05,602] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.08 GB, percent = 11.1% 0: ninja: no work to do. 0: Time to load utils op: 0.2997281551361084 seconds 0: Time to load utils op: 0.20277070999145508 seconds 13: Time to load utils op: 0.318281888961792 seconds 5: Time to load utils op: 0.3286752700805664 seconds 10: Time to load utils op: 0.32410526275634766 seconds 6: Time to load utils op: 0.3285040855407715 seconds 0: [2022-12-25 08:59:05,915] [INFO] [utils.py:827:see_memory_usage] before initializing group 0 0: [2022-12-25 08:59:05,915] [INFO] [utils.py:828:see_memory_usage] MA 5.25 GB Max_MA 5.25 GB CA 5.32 GB Max_CA 5 GB 0: [2022-12-25 08:59:05,915] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.09 GB, percent = 11.1% 0: ninja: no work to do. 0: Time to load utils op: 0.1878347396850586 seconds 0: [2022-12-25 08:59:06,041] [INFO] [utils.py:827:see_memory_usage] after initializing group 0 0: [2022-12-25 08:59:06,042] [INFO] [utils.py:828:see_memory_usage] MA 10.64 GB Max_MA 10.64 GB CA 13.39 GB Max_CA 13 GB 0: [2022-12-25 08:59:06,042] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.09 GB, percent = 11.1% 0: Time to load utils op: 0.2029247283935547 seconds 0: Time to load utils op: 0.20276451110839844 secondsTime to load utils op: 0.2021932601928711 seconds 0: 0: Time to load utils op: 0.20264434814453125 seconds 0: Time to load utils op: 0.20200037956237793 seconds 0: Time to load utils op: 0.0006778240203857422 seconds 5: Time to load utils op: 0.20403099060058594 seconds 6: Time to load utils op: 0.20310163497924805 seconds 6: Time to load utils op: 0.20313477516174316 seconds 5: Time to load utils op: 0.20430684089660645 seconds 0: Time to load utils op: 0.0004794597625732422 seconds 5: Time to load utils op: 0.20447611808776855 seconds 5: Time to load utils op: 0.2052137851715088 seconds 5: Time to load utils op: 0.2052924633026123 seconds 6: Time to load utils op: 0.20345330238342285 secondsTime to load utils op: 0.20227646827697754 seconds 6: 6: Time to load utils op: 0.20264363288879395 secondsTime to load utils op: 0.20367050170898438 secondsTime to load utils op: 0.20314788818359375 seconds 6: 6: 5: Time to load utils op: 0.20550537109375 secondsTime to load utils op: 0.2051856517791748 seconds 5: 10: Time to load utils op: 0.20285964012145996 seconds 10: Time to load utils op: 0.20200014114379883 seconds 10: Time to load utils op: 0.20261883735656738 seconds 10: Time to load utils op: 0.20243573188781738 seconds 10: Time to load utils op: 0.20250654220581055 seconds 10: Time to load utils op: 0.20257091522216797 seconds 10: Time to load utils op: 0.20267558097839355 seconds 13: Time to load utils op: 0.20218753814697266 seconds 13: Time to load utils op: 0.20346903800964355 seconds 13: Time to load utils op: 0.20259428024291992 seconds 13: Time to load utils op: 0.20235776901245117 seconds 13: Time to load utils op: 0.20384693145751953 seconds 13: Time to load utils op: 0.20369696617126465 seconds 13: Time to load utils op: 0.20411920547485352 seconds 6: Time to load utils op: 0.0005533695220947266 seconds 10: Time to load utils op: 0.0005452632904052734 seconds 5: Time to load utils op: 0.0006985664367675781 seconds 0: Time to load utils op: 0.0004582405090332031 seconds 0: Time to load utils op: 0.00040531158447265625 seconds 0: Time to load utils op: 0.00042128562927246094 seconds 0: Time to load utils op: 0.0004165172576904297 seconds 0: Time to load utils op: 0.0004220008850097656 seconds 13: Time to load utils op: 0.0005023479461669922 seconds 6: Time to load utils op: 0.0003647804260253906 seconds 6: Time to load utils op: 0.0003600120544433594 seconds 5: Time to load utils op: 0.0003578662872314453 seconds 5: Time to load utils op: 0.0003497600555419922 seconds 5: Time to load utils op: 0.0004112720489501953 seconds 5: Time to load utils op: 0.00036215782165527344 seconds 5: Time to load utils op: 0.00036144256591796875 seconds 6: Time to load utils op: 0.0003936290740966797 seconds 5: Time to load utils op: 0.0003712177276611328 seconds 5: Time to load utils op: 0.0003485679626464844 seconds 6: Time to load utils op: 0.0003578662872314453 seconds 6: Time to load utils op: 0.00034689903259277344 seconds 6: Time to load utils op: 0.0003771781921386719 seconds 6: Time to load utils op: 0.00032830238342285156 seconds 10: Time to load utils op: 0.0003294944763183594 seconds 10: Time to load utils op: 0.0003790855407714844 seconds 10: Time to load utils op: 0.0004038810729980469 seconds 10: Time to load utils op: 0.0004093647003173828 seconds 10: Time to load utils op: 0.00035572052001953125 seconds 10: Time to load utils op: 0.00035858154296875 seconds 10: Time to load utils op: 0.00037741661071777344 seconds 1: Time to load utils op: 0.23421216011047363 seconds 1: Time to load utils op: 0.23427820205688477 seconds 2: Time to load utils op: 0.23291301727294922 seconds 1: Time to load utils op: 0.2342846393585205 seconds 2: Time to load utils op: 0.23296046257019043 seconds 1: Time to load utils op: 0.234344482421875 secondsTime to load utils op: 0.2343158721923828 secondsTime to load utils op: 0.23431921005249023 seconds 1: 1: 1: Time to load utils op: 0.23432588577270508 seconds 2: Time to load utils op: 0.23296737670898438 seconds 2: Time to load utils op: 0.23298335075378418 seconds 1: Time to load utils op: 0.23434829711914062 seconds 2: Time to load utils op: 0.23303818702697754 seconds 2: Time to load utils op: 0.23303699493408203 seconds 2: Time to load utils op: 0.23303008079528809 secondsTime to load utils op: 0.23303747177124023 seconds 2: 4: Time to load utils op: 0.23479938507080078 seconds 4: Time to load utils op: 0.2351236343383789 seconds 4: Time to load utils op: 0.2349996566772461 secondsTime to load utils op: 0.23498129844665527 seconds 4: 4: Time to load utils op: 0.23490476608276367 secondsTime to load utils op: 0.2351818084716797 seconds 4: 4: Time to load utils op: 0.23519396781921387 seconds 4: Time to load utils op: 0.23523426055908203 seconds 13: Time to load utils op: 0.0003972053527832031 seconds 13: Time to load utils op: 0.0003330707550048828 seconds 13: Time to load utils op: 0.00040435791015625 seconds 13: Time to load utils op: 0.00038695335388183594 seconds 13: Time to load utils op: 0.0003974437713623047 seconds 13: Time to load utils op: 0.0003752708435058594 seconds 3: Time to load utils op: 0.23125243186950684 seconds 3: Time to load utils op: 0.23124980926513672 seconds 3: Time to load utils op: 0.23125457763671875 seconds 3: Time to load utils op: 0.231276273727417 seconds 3: Time to load utils op: 0.23120856285095215 secondsTime to load utils op: 0.2312784194946289 seconds 3: 3: Time to load utils op: 0.23128771781921387 secondsTime to load utils op: 0.2312941551208496 seconds 3: 13: Time to load utils op: 0.000385284423828125 seconds 28: Time to load utils op: 0.21480298042297363 seconds 20: Time to load utils op: 0.2254495620727539 secondsTime to load utils op: 0.2254798412322998 seconds 20: 28: Time to load utils op: 0.5399811267852783 secondsTime to load utils op: 0.20543694496154785 seconds 28: 20: Time to load utils op: 0.22546982765197754 seconds 20: Time to load utils op: 0.2255401611328125 secondsTime to load utils op: 0.22553610801696777 seconds 20: 28: Time to load utils op: 0.20632433891296387 secondsTime to load utils op: 0.21631669998168945 seconds 28: 20: Time to load utils op: 0.22550082206726074 secondsTime to load utils op: 0.2255408763885498 secondsTime to load utils op: 0.22554540634155273 seconds 28: Time to load utils op: 0.20300602912902832 seconds 20: 20: 28: Time to load utils op: 0.2172257900238037 secondsTime to load utils op: 0.20364689826965332 seconds 28: 23: Time to load utils op: 0.2282090187072754 seconds 23: Time to load utils op: 0.22815704345703125 seconds 23: Time to load utils op: 0.22710514068603516 seconds 9: Time to load utils op: 0.242828369140625 seconds 9: Time to load utils op: 0.24283242225646973 seconds 23: Time to load utils op: 0.2290785312652588 secondsTime to load utils op: 0.22832417488098145 seconds 23: 23: Time to load utils op: 0.22902798652648926 secondsTime to load utils op: 0.22730207443237305 seconds 23: 9: Time to load utils op: 0.23246145248413086 seconds 23: Time to load utils op: 0.2277061939239502 seconds 9: Time to load utils op: 0.24219560623168945 seconds 9: Time to load utils op: 0.2422323226928711 seconds 9: Time to load utils op: 0.24240779876708984 seconds 9: Time to load utils op: 0.2356572151184082 secondsTime to load utils op: 0.2354114055633545 seconds 9: 18: Time to load utils op: 0.23271679878234863 seconds 18: Time to load utils op: 0.23272275924682617 seconds 18: Time to load utils op: 0.23273277282714844 seconds 18: Time to load utils op: 0.23279023170471191 seconds 18: Time to load utils op: 0.23280787467956543 secondsTime to load utils op: 0.2327580451965332 seconds 18: Time to load utils op: 0.23274827003479004 seconds 18: Time to load utils op: 0.2328169345855713 seconds 18: 15: Time to load utils op: 0.23714756965637207 secondsTime to load utils op: 0.23714399337768555 seconds 15: Time to load utils op: 0.2371523380279541 seconds 15: 15: Time to load utils op: 0.23715543746948242 seconds 15: Time to load utils op: 0.2371969223022461 secondsTime to load utils op: 0.23719263076782227 seconds 15: 15: Time to load utils op: 0.23714876174926758 seconds 15: Time to load utils op: 0.23717880249023438 seconds 7: Time to load utils op: 0.2486705780029297 seconds 7: Time to load utils op: 0.24868392944335938 seconds 7: Time to load utils op: 0.24868321418762207 seconds 7: Time to load utils op: 0.24875617027282715 seconds 7: Time to load utils op: 0.24873852729797363 seconds 7: Time to load utils op: 0.2487943172454834 secondsTime to load utils op: 0.24878406524658203 secondsTime to load utils op: 0.2487928867340088 seconds 7: 7: 11: Time to load utils op: 0.24539685249328613 seconds 11: Time to load utils op: 0.24541735649108887 secondsTime to load utils op: 0.24541950225830078 seconds 11: 11: Time to load utils op: 0.2454516887664795 seconds 11: Time to load utils op: 0.24545073509216309 seconds 11: Time to load utils op: 0.2454686164855957 seconds 11: Time to load utils op: 0.2454698085784912 seconds 11: Time to load utils op: 0.24547767639160156 seconds 25: Time to load utils op: 0.22903203964233398 seconds 25: Time to load utils op: 0.22902989387512207 seconds 25: Time to load utils op: 0.22911667823791504 seconds 25: Time to load utils op: 0.22906184196472168 seconds 25: Time to load utils op: 0.22911334037780762 seconds 25: Time to load utils op: 0.22912001609802246 secondsTime to load utils op: 0.229081392288208 seconds 25: Time to load utils op: 0.2291100025177002 seconds 25: 14: Time to load utils op: 0.24260950088500977 seconds 14: Time to load utils op: 0.24262213706970215 seconds 14: Time to load utils op: 0.24257850646972656 secondsTime to load utils op: 0.24257826805114746 secondsTime to load utils op: 0.2425839900970459 seconds 14: 14: 14: Time to load utils op: 0.24257755279541016 secondsTime to load utils op: 0.24260449409484863 seconds 14: 14: Time to load utils op: 0.24266862869262695 seconds 26: Time to load utils op: 0.2270967960357666 seconds 17: Time to load utils op: 0.23952031135559082 seconds 26: Time to load utils op: 0.22713065147399902 seconds 17: Time to load utils op: 0.23954439163208008 seconds 26: Time to load utils op: 0.22714900970458984 seconds 17: Time to load utils op: 0.23956608772277832 seconds 17: Time to load utils op: 0.23958468437194824 seconds 17: Time to load utils op: 0.23958754539489746 secondsTime to load utils op: 0.23958778381347656 seconds 17: 17: Time to load utils op: 0.23960566520690918 seconds 26: Time to load utils op: 0.22716736793518066 seconds 17: Time to load utils op: 0.23962116241455078 seconds 26: Time to load utils op: 0.22720909118652344 secondsTime to load utils op: 0.22722125053405762 secondsTime to load utils op: 0.22721362113952637 seconds 26: 26: 26: Time to load utils op: 0.22719264030456543 seconds 16: Time to load utils op: 0.24216127395629883 secondsTime to load utils op: 0.24215126037597656 seconds 16: 16: Time to load utils op: 0.24215245246887207 secondsTime to load utils op: 0.24226069450378418 seconds 16: Time to load utils op: 0.24216198921203613 seconds 16: 16: Time to load utils op: 0.24216532707214355 seconds 16: Time to load utils op: 0.24216556549072266 seconds 16: Time to load utils op: 0.24225211143493652 seconds 12: Time to load utils op: 0.24688124656677246 seconds 12: Time to load utils op: 0.24692845344543457 seconds 12: Time to load utils op: 0.2469310760498047 seconds 12: Time to load utils op: 0.24698233604431152 secondsTime to load utils op: 0.2469635009765625 seconds 12: 12: Time to load utils op: 0.24698591232299805 seconds 12: Time to load utils op: 0.24698352813720703 seconds 12: Time to load utils op: 0.24698162078857422 seconds 19: Time to load utils op: 0.23755431175231934 seconds 19: Time to load utils op: 0.23757147789001465 seconds 19: Time to load utils op: 0.23760247230529785 seconds 19: Time to load utils op: 0.2376079559326172 seconds 19: Time to load utils op: 0.2376255989074707 seconds 19: Time to load utils op: 0.23763227462768555 seconds 19: Time to load utils op: 0.2376711368560791 seconds 19: Time to load utils op: 0.237656831741333 seconds 21: Time to load utils op: 0.23584461212158203 secondsTime to load utils op: 0.23586583137512207 seconds 21: 8: Time to load utils op: 0.2526531219482422 secondsTime to load utils op: 0.25264954566955566 seconds 8: 21: Time to load utils op: 0.23590874671936035 seconds 21: Time to load utils op: 0.23591232299804688 seconds 8: Time to load utils op: 0.2527132034301758 seconds 21: Time to load utils op: 0.23594355583190918 secondsTime to load utils op: 0.23592114448547363 seconds 21: 8: Time to load utils op: 0.25272035598754883 seconds 21: Time to load utils op: 0.23593688011169434 secondsTime to load utils op: 0.23594450950622559 seconds 21: 8: Time to load utils op: 0.25269246101379395 secondsTime to load utils op: 0.25273799896240234 seconds 8: 8: Time to load utils op: 0.25273823738098145 secondsTime to load utils op: 0.25276660919189453 seconds 8: 29: Time to load utils op: 0.2269275188446045 seconds 29: Time to load utils op: 0.22688961029052734 seconds 29: Time to load utils op: 0.2268984317779541 seconds 29: Time to load utils op: 0.2269277572631836 seconds 29: Time to load utils op: 0.22700762748718262 secondsTime to load utils op: 0.2269437313079834 seconds 29: 29: Time to load utils op: 0.22694754600524902 secondsTime to load utils op: 0.2270219326019287 seconds 29: 27: Time to load utils op: 0.2289886474609375 secondsTime to load utils op: 0.2289893627166748 seconds 27: 27: Time to load utils op: 0.22899198532104492 seconds 27: Time to load utils op: 0.22902441024780273 seconds 27: Time to load utils op: 0.22907519340515137 secondsTime to load utils op: 0.22909879684448242 seconds 27: Time to load utils op: 0.2290799617767334 seconds 27: 27: Time to load utils op: 0.22911334037780762 seconds 24: Time to load utils op: 0.23287725448608398 seconds 24: Time to load utils op: 0.2329254150390625 seconds 24: Time to load utils op: 0.23294425010681152 seconds 24: Time to load utils op: 0.2329864501953125 secondsTime to load utils op: 0.23299288749694824 seconds 24: 24: Time to load utils op: 0.23301410675048828 seconds 24: Time to load utils op: 0.23297882080078125 seconds 24: Time to load utils op: 0.23305463790893555 seconds 22: Time to load utils op: 0.23567581176757812 secondsTime to load utils op: 0.23567962646484375 seconds 22: 22: Time to load utils op: 0.2356882095336914 seconds 22: Time to load utils op: 0.2357320785522461 secondsTime to load utils op: 0.23573994636535645 seconds 22: 22: Time to load utils op: 0.23571300506591797 seconds 22: Time to load utils op: 0.2357494831085205 seconds 22: Time to load utils op: 0.2357490062713623 seconds 30: Time to load utils op: 0.22635388374328613 seconds 30: Time to load utils op: 0.2263495922088623 seconds 31: Time to load utils op: 0.2250504493713379 secondsTime to load utils op: 0.22504758834838867 seconds 31: Time to load utils op: 0.22515416145324707 seconds 31: 31: Time to load utils op: 0.22505593299865723 secondsTime to load utils op: 0.22506070137023926 secondsTime to load utils op: 0.22505855560302734 seconds 31: 31: 31: Time to load utils op: 0.22505974769592285 secondsTime to load utils op: 0.22506475448608398 seconds 31: 30: Time to load utils op: 0.22635674476623535 secondsTime to load utils op: 0.22635960578918457 seconds 30: 30: Time to load utils op: 0.22637438774108887 seconds 30: Time to load utils op: 0.22638678550720215 secondsTime to load utils op: 0.22639155387878418 seconds 30: 30: Time to load utils op: 0.22639083862304688 seconds 2: Time to load utils op: 0.0009987354278564453 seconds 2: Time to load utils op: 0.0011584758758544922 seconds 2: Time to load utils op: 0.0014505386352539062 seconds 2: Time to load utils op: 0.001461029052734375 seconds 2: Time to load utils op: 0.0014233589172363281 seconds 2: Time to load utils op: 0.0013172626495361328 seconds 2: Time to load utils op: 0.0014104843139648438 seconds 2: Time to load utils op: 0.0014643669128417969 seconds 4: Time to load utils op: 0.0008494853973388672 seconds 4: Time to load utils op: 0.0007739067077636719 seconds 3: Time to load utils op: 0.0009584426879882812 secondsTime to load utils op: 0.0009791851043701172 seconds 3: 1: Time to load utils op: 0.0009438991546630859 seconds 4: Time to load utils op: 0.0010187625885009766 seconds 3: Time to load utils op: 0.001065969467163086 seconds 1: Time to load utils op: 0.0009129047393798828 seconds 1: Time to load utils op: 0.0010216236114501953 seconds 3: Time to load utils op: 0.0012583732604980469 secondsTime to load utils op: 0.001222372055053711 seconds 3: 4: Time to load utils op: 0.0011937618255615234 seconds 4: Time to load utils op: 0.0012483596801757812 seconds 3: Time to load utils op: 0.001298666000366211 seconds 4: Time to load utils op: 0.0011627674102783203 seconds 3: Time to load utils op: 0.0012454986572265625 seconds 4: Time to load utils op: 0.0012869834899902344 seconds 3: Time to load utils op: 0.001287221908569336 seconds 4: Time to load utils op: 0.0012753009796142578 seconds 1: Time to load utils op: 0.0012149810791015625 seconds 1: Time to load utils op: 0.001131296157836914 seconds 1: Time to load utils op: 0.0012023448944091797 seconds 1: Time to load utils op: 0.001234292984008789 seconds 1: Time to load utils op: 0.0013699531555175781 seconds 20: Time to load utils op: 0.0010225772857666016 seconds 20: Time to load utils op: 0.001293182373046875 seconds 20: Time to load utils op: 0.0014319419860839844 seconds 20: Time to load utils op: 0.0013704299926757812 secondsTime to load utils op: 0.0014984607696533203 seconds 20: 20: Time to load utils op: 0.0013816356658935547 seconds 20: Time to load utils op: 0.0013899803161621094 seconds 20: Time to load utils op: 0.0013391971588134766 seconds 23: Time to load utils op: 0.00045800209045410156 secondsTime to load utils op: 0.0005080699920654297 seconds 23: 23: Time to load utils op: 0.0004177093505859375 secondsTime to load utils op: 0.0004508495330810547 seconds 23: 23: Time to load utils op: 0.0004057884216308594 seconds 23: Time to load utils op: 0.0005273818969726562 seconds 23: Time to load utils op: 0.000545501708984375 seconds 23: Time to load utils op: 0.000423431396484375 seconds 28: Time to load utils op: 0.0003604888916015625 seconds 28: Time to load utils op: 0.0005619525909423828 secondsTime to load utils op: 0.00037860870361328125 secondsTime to load utils op: 0.0006010532379150391 seconds 28: 28: 28: Time to load utils op: 0.00040721893310546875 seconds 28: Time to load utils op: 0.0004119873046875 seconds 28: Time to load utils op: 0.0003762245178222656 seconds 28: Time to load utils op: 0.00041484832763671875 seconds 7: Time to load utils op: 0.0008816719055175781 seconds 7: Time to load utils op: 0.0009951591491699219 seconds 7: Time to load utils op: 0.001033782958984375 seconds 7: Time to load utils op: 0.0010051727294921875 seconds 7: Time to load utils op: 0.0010831356048583984 seconds 7: Time to load utils op: 0.0010421276092529297 seconds 7: Time to load utils op: 0.0010802745819091797 seconds 7: Time to load utils op: 0.001119852066040039 seconds 15: Time to load utils op: 0.0012288093566894531 seconds 15: Time to load utils op: 0.0013699531555175781 seconds 9: Time to load utils op: 0.0004248619079589844 secondsTime to load utils op: 0.0005674362182617188 seconds 9: 15: Time to load utils op: 0.001514434814453125 seconds 15: Time to load utils op: 0.001478433609008789 seconds 15: Time to load utils op: 0.0015530586242675781 secondsTime to load utils op: 0.0015552043914794922 seconds 15: 9: Time to load utils op: 0.00036525726318359375 secondsTime to load utils op: 0.0005180835723876953 seconds 9: 15: Time to load utils op: 0.0015473365783691406 seconds 15: Time to load utils op: 0.0015397071838378906 seconds 9: Time to load utils op: 0.00040221214294433594 seconds 9: Time to load utils op: 0.00042939186096191406 seconds 9: Time to load utils op: 0.00041365623474121094 secondsTime to load utils op: 0.0004131793975830078 seconds 9: 14: Time to load utils op: 0.0007159709930419922 seconds 0: [2022-12-25 08:59:06,156] [INFO] [utils.py:827:see_memory_usage] before initializing group 1 14: Time to load utils op: 0.0009515285491943359 seconds 14: Time to load utils op: 0.0010271072387695312 seconds 14: Time to load utils op: 0.0011920928955078125 seconds 14: Time to load utils op: 0.0012476444244384766 seconds 14: Time to load utils op: 0.0010776519775390625 seconds 14: Time to load utils op: 0.0011010169982910156 seconds 0: [2022-12-25 08:59:06,157] [INFO] [utils.py:828:see_memory_usage] MA 10.64 GB Max_MA 10.64 GB CA 13.39 GB Max_CA 13 GB 14: Time to load utils op: 0.001131296157836914 seconds 0: [2022-12-25 08:59:06,157] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.24 GB, percent = 11.2% 25: Time to load utils op: 0.0010788440704345703 seconds 16: Time to load utils op: 0.0006966590881347656 seconds 16: Time to load utils op: 0.0005052089691162109 secondsTime to load utils op: 0.0005099773406982422 seconds 16: 16: Time to load utils op: 0.0005035400390625 secondsTime to load utils op: 0.0005154609680175781 secondsTime to load utils op: 0.0004870891571044922 seconds 16: 16: 16: Time to load utils op: 0.00048828125 seconds 16: Time to load utils op: 0.0006608963012695312 seconds 25: Time to load utils op: 0.001432657241821289 seconds 25: Time to load utils op: 0.0013630390167236328 seconds 25: Time to load utils op: 0.0013506412506103516 secondsTime to load utils op: 0.001283407211303711 seconds 25: 25: Time to load utils op: 0.0013432502746582031 seconds 25: Time to load utils op: 0.001321554183959961 seconds 25: Time to load utils op: 0.0014073848724365234 seconds 24: Time to load utils op: 0.0006110668182373047 seconds 22: Time to load utils op: 0.000896453857421875 seconds 24: Time to load utils op: 0.0006461143493652344 seconds 26: Time to load utils op: 0.0012745857238769531 seconds 17: Time to load utils op: 0.0008554458618164062 seconds 22: Time to load utils op: 0.0013594627380371094 seconds 22: Time to load utils op: 0.0013232231140136719 seconds 22: Time to load utils op: 0.0013287067413330078 secondsTime to load utils op: 0.001369476318359375 seconds 22: 26: Time to load utils op: 0.0013539791107177734 seconds 26: Time to load utils op: 0.001514434814453125 seconds 22: Time to load utils op: 0.001346588134765625 seconds 22: Time to load utils op: 0.0013480186462402344 seconds 24: Time to load utils op: 0.0013959407806396484 seconds 22: Time to load utils op: 0.001444101333618164 seconds 24: Time to load utils op: 0.0014216899871826172 secondsTime to load utils op: 0.0013363361358642578 seconds 24: 26: Time to load utils op: 0.0014431476593017578 seconds 26: Time to load utils op: 0.0015156269073486328 seconds 26: Time to load utils op: 0.0014638900756835938 seconds 24: Time to load utils op: 0.0014300346374511719 secondsTime to load utils op: 0.0014488697052001953 seconds 24: 17: Time to load utils op: 0.0010738372802734375 seconds 26: Time to load utils op: 0.0014622211456298828 seconds 17: Time to load utils op: 0.0011124610900878906 seconds 24: Time to load utils op: 0.0013968944549560547 seconds 26: Time to load utils op: 0.0015420913696289062 seconds 17: Time to load utils op: 0.001226186752319336 seconds 17: Time to load utils op: 0.0012278556823730469 secondsTime to load utils op: 0.001306772232055664 seconds 17: 17: Time to load utils op: 0.0012362003326416016 seconds 17: Time to load utils op: 0.0012958049774169922 seconds 27: Time to load utils op: 0.0006356239318847656 seconds 18: Time to load utils op: 0.001054525375366211 seconds 18: Time to load utils op: 0.0011773109436035156 seconds 27: Time to load utils op: 0.0009791851043701172 seconds 27: Time to load utils op: 0.0011191368103027344 seconds 29: Time to load utils op: 0.0006716251373291016 seconds 27: Time to load utils op: 0.0012476444244384766 seconds 27: Time to load utils op: 0.0012023448944091797 seconds 18: Time to load utils op: 0.0014050006866455078 seconds 27: Time to load utils op: 0.0011479854583740234 seconds 18: Time to load utils op: 0.0014171600341796875 seconds 27: Time to load utils op: 0.0010950565338134766 seconds 18: Time to load utils op: 0.0013079643249511719 secondsTime to load utils op: 0.0013430118560791016 secondsTime to load utils op: 0.0014927387237548828 seconds 18: 18: 27: Time to load utils op: 0.0011284351348876953 seconds 18: Time to load utils op: 0.0014209747314453125 seconds 29: Time to load utils op: 0.0006449222564697266 seconds 29: Time to load utils op: 0.000644683837890625 secondsTime to load utils op: 0.0006318092346191406 seconds 29: 29: Time to load utils op: 0.0006318092346191406 seconds 29: Time to load utils op: 0.0006222724914550781 seconds 29: Time to load utils op: 0.0006251335144042969 seconds 29: Time to load utils op: 0.0008916854858398438 seconds 31: Time to load utils op: 0.001172780990600586 seconds 8: Time to load utils op: 0.000993490219116211 seconds 19: Time to load utils op: 0.0006532669067382812 seconds 31: Time to load utils op: 0.0012524127960205078 secondsTime to load utils op: 0.0013248920440673828 seconds 31: 19: Time to load utils op: 0.0004134178161621094 seconds 31: Time to load utils op: 0.0012769699096679688 seconds 31: Time to load utils op: 0.001283407211303711 seconds 31: Time to load utils op: 0.0012922286987304688 seconds 31: Time to load utils op: 0.0013842582702636719 seconds 8: Time to load utils op: 0.0012340545654296875 seconds 31: Time to load utils op: 0.001409769058227539 seconds 19: Time to load utils op: 0.0006451606750488281 seconds 19: Time to load utils op: 0.000644683837890625 seconds 19: Time to load utils op: 0.0006279945373535156 seconds 8: Time to load utils op: 0.0011332035064697266 seconds 19: Time to load utils op: 0.0009779930114746094 seconds 8: Time to load utils op: 0.0014615058898925781 seconds 19: Time to load utils op: 0.0008082389831542969 seconds 8: Time to load utils op: 0.00139617919921875 seconds 8: Time to load utils op: 0.0014066696166992188 seconds 8: Time to load utils op: 0.0014750957489013672 seconds 19: Time to load utils op: 0.0010502338409423828 seconds 8: Time to load utils op: 0.0013799667358398438 seconds 21: Time to load utils op: 0.0013022422790527344 secondsTime to load utils op: 0.0012540817260742188 seconds 21: 21: Time to load utils op: 0.0015759468078613281 secondsTime to load utils op: 0.001447439193725586 seconds 21: 21: Time to load utils op: 0.001558065414428711 seconds 21: Time to load utils op: 0.0014872550964355469 seconds 21: Time to load utils op: 0.001596689224243164 seconds 21: Time to load utils op: 0.0015997886657714844 seconds 30: Time to load utils op: 0.0008556842803955078 seconds 30: Time to load utils op: 0.0009558200836181641 secondsTime to load utils op: 0.00095367431640625 seconds 30: 30: Time to load utils op: 0.0010464191436767578 seconds 30: Time to load utils op: 0.0011992454528808594 seconds 30: Time to load utils op: 0.0012738704681396484 seconds 30: Time to load utils op: 0.0012314319610595703 seconds 30: Time to load utils op: 0.001310586929321289 seconds 11: Time to load utils op: 0.0009119510650634766 seconds 11: Time to load utils op: 0.0008089542388916016 seconds 11: Time to load utils op: 0.001239776611328125 seconds 11: Time to load utils op: 0.001180410385131836 seconds 11: Time to load utils op: 0.0011608600616455078 secondsTime to load utils op: 0.0011539459228515625 seconds 11: 11: Time to load utils op: 0.0012309551239013672 seconds 11: Time to load utils op: 0.0011799335479736328 seconds 12: Time to load utils op: 0.0008022785186767578 seconds 12: Time to load utils op: 0.0006120204925537109 seconds 12: Time to load utils op: 0.0007774829864501953 secondsTime to load utils op: 0.0006577968597412109 seconds 12: 12: Time to load utils op: 0.0006475448608398438 secondsTime to load utils op: 0.0006711483001708984 seconds 12: 12: Time to load utils op: 0.0010249614715576172 seconds 12: Time to load utils op: 0.0007886886596679688 seconds 0: [2022-12-25 08:59:06,257] [INFO] [utils.py:827:see_memory_usage] after initializing group 1 0: [2022-12-25 08:59:06,257] [INFO] [utils.py:828:see_memory_usage] MA 15.73 GB Max_MA 15.73 GB CA 21.01 GB Max_CA 21 GB 0: [2022-12-25 08:59:06,257] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.25 GB, percent = 11.2% 0: [2022-12-25 08:59:06,350] [INFO] [utils.py:827:see_memory_usage] before initializing group 2 0: [2022-12-25 08:59:06,350] [INFO] [utils.py:828:see_memory_usage] MA 15.73 GB Max_MA 15.73 GB CA 21.01 GB Max_CA 21 GB 0: [2022-12-25 08:59:06,350] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.25 GB, percent = 11.2% 0: [2022-12-25 08:59:06,447] [INFO] [utils.py:827:see_memory_usage] after initializing group 2 0: [2022-12-25 08:59:06,448] [INFO] [utils.py:828:see_memory_usage] MA 15.74 GB Max_MA 15.74 GB CA 21.01 GB Max_CA 21 GB 0: [2022-12-25 08:59:06,448] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.25 GB, percent = 11.2% 0: [2022-12-25 08:59:06,539] [INFO] [utils.py:827:see_memory_usage] before initialize_optimizer 0: [2022-12-25 08:59:06,539] [INFO] [utils.py:828:see_memory_usage] MA 15.74 GB Max_MA 15.74 GB CA 21.01 GB Max_CA 21 GB 0: [2022-12-25 08:59:06,540] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.25 GB, percent = 11.2% 0: [2022-12-25 08:59:06,636] [INFO] [utils.py:827:see_memory_usage] end initialize_optimizer 0: [2022-12-25 08:59:06,636] [INFO] [utils.py:828:see_memory_usage] MA 15.82 GB Max_MA 15.82 GB CA 21.01 GB Max_CA 21 GB 0: [2022-12-25 08:59:06,636] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.25 GB, percent = 11.2% 0: [2022-12-25 08:59:06,728] [INFO] [utils.py:827:see_memory_usage] end bf16_optimizer 0: [2022-12-25 08:59:06,728] [INFO] [utils.py:828:see_memory_usage] MA 15.82 GB Max_MA 15.82 GB CA 21.01 GB Max_CA 21 GB 0: [2022-12-25 08:59:06,728] [INFO] [utils.py:836:see_memory_usage] CPU Virtual Memory: used = 56.25 GB, percent = 11.2% 0: [2022-12-25 08:59:06,728] [INFO] [logging.py:68:log_dist] [Rank 0] DeepSpeed Final Optimizer = FusedAdam 0: [2022-12-25 08:59:06,728] [INFO] [logging.py:68:log_dist] [Rank 0] DeepSpeed using client LR scheduler 0: [2022-12-25 08:59:06,729] [INFO] [logging.py:68:log_dist] [Rank 0] DeepSpeed LR Scheduler = 0: [2022-12-25 08:59:06,729] [INFO] [logging.py:68:log_dist] [Rank 0] step=0, skipped=0, lr=[0.0, 0.0, 0.0], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 0: [2022-12-25 08:59:06,729] [INFO] [config.py:1007:print] DeepSpeedEngine configuration: 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] activation_checkpointing_config { 0: "partition_activations": false, 0: "contiguous_memory_optimization": false, 0: "cpu_checkpointing": false, 0: "number_checkpoints": null, 0: "synchronize_checkpoint_boundary": false, 0: "profile": false 0: } 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] aio_config ................... {'block_size': 1048576, 'queue_depth': 8, 'thread_count': 1, 'single_submit': False, 'overlap_events': True} 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] amp_enabled .................. False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] amp_params ................... False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] autotuning_config ............ { 0: "enabled": false, 0: "start_step": null, 0: "end_step": null, 0: "metric_path": null, 0: "arg_mappings": null, 0: "metric": "throughput", 0: "model_info": null, 0: "results_dir": "/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/autotuning_results", 0: "exps_dir": "/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/autotuning_exps", 0: "overwrite": true, 0: "fast": true, 0: "start_profile_step": 3, 0: "end_profile_step": 5, 0: "tuner_type": "gridsearch", 0: "tuner_early_stopping": 5, 0: "tuner_num_trials": 50, 0: "model_info_path": null, 0: "mp_size": 1, 0: "max_train_batch_size": null, 0: "min_train_batch_size": 1, 0: "max_train_micro_batch_size_per_gpu": 1.024000e+03, 0: "min_train_micro_batch_size_per_gpu": 1, 0: "num_tuning_micro_batch_sizes": 3 0: } 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] bfloat16_enabled ............. True 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] checkpoint_parallel_write_pipeline False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] checkpoint_tag_validation_enabled True 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] checkpoint_tag_validation_fail False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] comms_config ................. 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] communication_data_type ...... None 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] compression_config ........... {'weight_quantization': {'shared_parameters': {'enabled': False, 'quantizer_kernel': False, 'schedule_offset': 0, 'quantize_groups': 1, 'quantize_verbose': False, 'quantization_type': 'symmetric', 'quantize_weight_in_forward': False, 'rounding': 'nearest', 'fp16_mixed_quantize': False, 'quantize_change_ratio': 0.001}, 'different_groups': {}}, 'activation_quantization': {'shared_parameters': {'enabled': False, 'quantization_type': 'symmetric', 'range_calibration': 'dynamic', 'schedule_offset': 1000}, 'different_groups': {}}, 'sparse_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'row_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'head_pruning': {'shared_parameters': {'enabled': False, 'method': 'topk', 'schedule_offset': 1000}, 'different_groups': {}}, 'channel_pruning': {'shared_pa 0: rameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'layer_reduction': {'enabled': False}} 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] curriculum_enabled ........... False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] curriculum_params ............ False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] dataloader_drop_last ......... False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] disable_allgather ............ False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] dump_state ................... False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] dynamic_loss_scale_args ...... None 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] eigenvalue_enabled ........... False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] eigenvalue_gas_boundary_resolution 1 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] eigenvalue_layer_name ........ bert.encoder.layer 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] eigenvalue_layer_num ......... 0 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] eigenvalue_max_iter .......... 100 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] eigenvalue_stability ......... 1e-06 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] eigenvalue_tol ............... 0.01 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] eigenvalue_verbose ........... False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] elasticity_enabled ........... False 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] flops_profiler_config ........ { 0: "enabled": false, 0: "profile_step": 1, 0: "module_depth": -1, 0: "top_modules": 1, 0: "detailed": true, 0: "output_file": null 0: } 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] fp16_auto_cast ............... None 0: [2022-12-25 08:59:06,730] [INFO] [config.py:1011:print] fp16_enabled ................. False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] fp16_master_weights_and_gradients False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] global_rank .................. 0 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] gradient_accumulation_steps .. 1 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] gradient_clipping ............ 1.0 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] gradient_predivide_factor .... 1.0 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] initial_dynamic_scale ........ 1 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] load_universal_checkpoint .... False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] loss_scale ................... 1.0 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] memory_breakdown ............. False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] monitor_config ............... 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] nebula_config ................ { 0: "enabled": false, 0: "persistent_storage_path": null, 0: "persistent_time_interval": 100, 0: "num_of_version_in_retention": 2, 0: "enable_nebula_load": true, 0: "load_path": null 0: } 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] optimizer_legacy_fusion ...... False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] optimizer_name ............... None 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] optimizer_params ............. None 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] pipeline ..................... {'stages': 'auto', 'partition': 'best', 'seed_layers': False, 'activation_checkpoint_interval': 0} 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] pld_enabled .................. False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] pld_params ................... False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] prescale_gradients ........... False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] scheduler_name ............... None 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] scheduler_params ............. None 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] sparse_attention ............. None 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] sparse_gradients_enabled ..... False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] steps_per_print .............. 2000 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] train_batch_size ............. 512 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] train_micro_batch_size_per_gpu 2 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] use_node_local_storage ....... False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] wall_clock_breakdown ......... False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] world_size ................... 256 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] zero_allow_untested_optimizer False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] zero_config .................. stage=0 contiguous_gradients=True reduce_scatter=True reduce_bucket_size=500000000 allgather_partitions=True allgather_bucket_size=500000000 overlap_comm=False load_from_fp32_weights=True elastic_checkpoint=False offload_param=None offload_optimizer=None sub_group_size=1000000000 cpu_offload_param=None cpu_offload_use_pin_memory=None cpu_offload=None prefetch_bucket_size=50000000 param_persistence_threshold=100000 model_persistence_threshold=9223372036854775807 max_live_parameters=1000000000 max_reuse_distance=1000000000 gather_16bit_weights_on_model_save=False stage3_gather_fp16_weights_on_model_save=False ignore_unused_parameters=True legacy_stage1=False round_robin_gradients=False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] zero_enabled ................. False 0: [2022-12-25 08:59:06,731] [INFO] [config.py:1011:print] zero_optimization_stage ...... 0 0: [2022-12-25 08:59:06,731] [INFO] [config.py:996:print_user_config] json = { 0: "train_micro_batch_size_per_gpu": 2, 0: "train_batch_size": 512, 0: "gradient_clipping": 1.0, 0: "zero_optimization": { 0: "stage": 0 0: }, 0: "bf16": { 0: "enabled": true 0: }, 0: "steps_per_print": 2.000000e+03, 0: "wall_clock_breakdown": false 0: } 0: Time to load utils op: 0.0004220008850097656 seconds 0: [2022-12-25 08:59:06,732] [INFO] [engine.py:87:__init__] CONFIG: micro_batches=1 micro_batch_size=2 0: [2022-12-25 08:59:06,806] [INFO] [engine.py:145:__init__] RANK=0 STAGE=0 LAYERS=41 [0, 41) STAGE_PARAMS=2809026560 (2809.027M) TOTAL_PARAMS=2809026560 (2809.027M) UNIQUE_PARAMS=2809026560 (2809.027M) 0: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 28: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 28: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: WARNING: could not find the metadata file checkpoints_2b855brealtasky 31: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 31: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: will not load any checkpoints and will start from random 24: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 24: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 24: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 28: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 30: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 30: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 31: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 31: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 16: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 16: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 16: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 8: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 8: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 24: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 28: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 30: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 27: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 27: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 16: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 30: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 29: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 29: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 20: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 20: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 14: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 12: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 12: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 12: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 8: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 26: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 26: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 29: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 23: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 20: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 27: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 22: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 22: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 15: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 15: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 15: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 14: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 12: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 8: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 26: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 29: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 23: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 20: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 27: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 22: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 15: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 14: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 18: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 18: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 13: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 13: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 25: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 25: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 31: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 11: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 11: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 23: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 23: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 6: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 6: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 16: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 14: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 4: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 4: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 4: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 13: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 13: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 24: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 21: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 21: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 30: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 30: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 26: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 25: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 31: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 11: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 10: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 10: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 22: [2022-12-25 08:59:06,811] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 6: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 3: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 3: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 18: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 5: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 5: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 4: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 9: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 9: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 28: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 28: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 21: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 21: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 19: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 19: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 19: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 10: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 10: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 6: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 16: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 24: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 17: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 17: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 19: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 25: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 11: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 20: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 14: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 18: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 5: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 5: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 9: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 17: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 27: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 27: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 14: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 12: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 12: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 3: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 8: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 7: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 7: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 7: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 17: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 26: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 26: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 29: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 29: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 23: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 20: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 2: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 2: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 22: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 22: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 1: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 1: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 15: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 15: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 16: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 16: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 3: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 8: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 4: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 7: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 9: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 24: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 28: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 30: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 30: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 31: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 23: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 2: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 1: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 6: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 18: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 18: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 4: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 7: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 7: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 28: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 21: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 21: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 19: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 19: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 25: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 31: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 11: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 10: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 20: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 2: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 6: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 14: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 14: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 12: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 12: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 8: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 24: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 29: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 23: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 20: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 27: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 27: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 1: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 15: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 15: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 3: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 3: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 5: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 9: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 26: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 25: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 29: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 11: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 10: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 23: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 22: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 8: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 5: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 17: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 17: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 2: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 2: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 9: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 26: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 4: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 21: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 19: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 19: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 21: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 11: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 11: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 10: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 22: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 6: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 6: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 18: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 4: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 7: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 25: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 1: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 10: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 18: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 25: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 7: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 1: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 5: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 3: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 5: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 2: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 2: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 9: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 9: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 17: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 17: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 3: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 1: [2022-12-25 08:59:06,813] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 1: [2022-12-25 08:59:06,813] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 31: time (ms) | load-checkpoint: 9.80 13: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 13: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 13: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 13: [2022-12-25 08:59:06,812] [WARNING] [engine.py:2581:load_checkpoint] Unable to find latest file at checkpoints_2b855brealtasky/latest, if trying to load latest checkpoint please ensure this file exists or pass an explicit checkpoint tag when loading a checkpoint. 0: estimated model parameters: 2.80902656 0: estimated model parameters without embeddings: 2.67500544 0: [after model, optimizer, and learning rate scheduler are built] datetime: 2022-12-25 08:59:07 0: > building train, validation, and test datasets ... 0: > datasets target sizes (minimum size): 0: train: 26855469 0: validation: 26880000 0: test: 51200 0: > building train, validation, and test datasets for GPT ... 0: > building dataset index ... 0: reading sizes... 0: reading pointers... 0: reading document index... 0: creating numpy buffer of mmap... 0: creating memory view of numpy buffer... 0: > finished creating indexed dataset in 0.035602 seconds 0: number of documents: 49378792 0: > dataset split: 0: train: 0: document indices in [0, 49378792) total of 49378792 documents 0: > WARNING: could not find index map files, building the indices on rank 0 ... 0: > last epoch number of samples (4922666) is smaller than 95.0% of number of samples per epoch (7310934), setting separate_last_epoch to True 0: > elasped time to build and save doc-idx mapping (seconds): 13.775576 0: using: 0: number of documents: 49378792 0: number of epochs: 4 0: sequence length: 2048 0: total number of samples: 29243738 0: > elasped time to build and save sample-idx mapping (seconds): 1.932505 0: > building shuffle index with split [0, 21932803) and [21932803, 29243738) ... 0: > elasped time to build and save shuffle-idx mapping (seconds): 1.030737 0: > loading doc-idx mapping from /scratch/project_462000119/data/realtasky/gpt2tok_realtasky_text_document_train_indexmap_26855469ns_2048sl_1234s_doc_idx.npy 0: > loading sample-idx mapping from /scratch/project_462000119/data/realtasky/gpt2tok_realtasky_text_document_train_indexmap_26855469ns_2048sl_1234s_sample_idx.npy 0: > loading shuffle-idx mapping from /scratch/project_462000119/data/realtasky/gpt2tok_realtasky_text_document_train_indexmap_26855469ns_2048sl_1234s_shuffle_idx.npy 0: loaded indexed file in 0.055 seconds 0: total number of samples: 29243739 0: total number of epochs: 4 0: > building dataset index ... 0: reading sizes... 0: reading pointers... 0: reading document index... 0: creating numpy buffer of mmap... 0: creating memory view of numpy buffer... 0: > finished creating indexed dataset in 0.000637 seconds 0: number of documents: 364608 0: > dataset split: 0: validation_c4: 0: document indices in [0, 364608) total of 364608 documents 0: > loading doc-idx mapping from /scratch/project_462000119/data/c4_validation/gpt2tok_c4validation_rerun_text_document_validation_c4_indexmap_26880000ns_2048sl_1234s_doc_idx.npy 0: > loading sample-idx mapping from /scratch/project_462000119/data/c4_validation/gpt2tok_c4validation_rerun_text_document_validation_c4_indexmap_26880000ns_2048sl_1234s_sample_idx.npy 0: > loading shuffle-idx mapping from /scratch/project_462000119/data/c4_validation/gpt2tok_c4validation_rerun_text_document_validation_c4_indexmap_26880000ns_2048sl_1234s_shuffle_idx.npy 0: loaded indexed file in 0.061 seconds 0: total number of samples: 26937865 0: total number of epochs: 317 0: > building dataset index ... 0: reading sizes... 0: reading pointers... 0: reading document index... 0: creating numpy buffer of mmap... 0: creating memory view of numpy buffer... 0: > finished creating indexed dataset in 0.009497 seconds 0: number of documents: 364608 0: > dataset split: 0: validation_oscar: 0: document indices in [0, 364608) total of 364608 documents 0: > loading doc-idx mapping from /scratch/project_462000119/data/oscar_validation/gpt2tok_oscarvalidation_text_document_validation_oscar_indexmap_26880000ns_2048sl_1234s_doc_idx.npy 0: > loading sample-idx mapping from /scratch/project_462000119/data/oscar_validation/gpt2tok_oscarvalidation_text_document_validation_oscar_indexmap_26880000ns_2048sl_1234s_sample_idx.npy 0: > loading shuffle-idx mapping from /scratch/project_462000119/data/oscar_validation/gpt2tok_oscarvalidation_text_document_validation_oscar_indexmap_26880000ns_2048sl_1234s_shuffle_idx.npy 0: loaded indexed file in 0.055 seconds 0: total number of samples: 26954343 0: total number of epochs: 172 0: > finished creating GPT datasets ... 0: [after dataloaders are built] datetime: 2022-12-25 08:59:42 0: done with setup ... 0: training ... 0: Number of parameters: [tensor rank - pipeline rank] w/ and w/o embeddings: 31: time (ms) | model-and-optimizer-setup: 32723.54 | train/valid/test-data-iterators-setup: 34066.77 0: [000-000] 2.8090B / 2.6750B 0: [before the start of training step] datetime: 2022-12-25 08:59:42 0: [Rank 0] (after 10 iterations) memory (MB) | allocated: 22352.48388671875 | max allocated: 62349.7021484375 | reserved: 62532.0 | max reserved: 63334.0 31: iteration 10/ 52452 | consumed samples: 5120 | consumed tokens: 10485760 | elapsed time per iteration (s): 4.28 | learning rate: 3.813E-06 | global batch size: 512 | lm loss: 1.060410E+01 | grad norm: 17.872 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 119.643 | TFLOPs: 17.96 | 31: iteration 20/ 52452 | consumed samples: 10240 | consumed tokens: 20971520 | elapsed time per iteration (s): 2.13 | learning rate: 7.626E-06 | global batch size: 512 | lm loss: 9.189635E+00 | grad norm: 5.832 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 239.928 | TFLOPs: 36.01 | 31: iteration 30/ 52452 | consumed samples: 15360 | consumed tokens: 31457280 | elapsed time per iteration (s): 1.89 | learning rate: 1.144E-05 | global batch size: 512 | lm loss: 8.777203E+00 | grad norm: 8.462 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.518 | TFLOPs: 40.60 | 31: iteration 40/ 52452 | consumed samples: 20480 | consumed tokens: 41943040 | elapsed time per iteration (s): 1.88 | learning rate: 1.525E-05 | global batch size: 512 | lm loss: 8.475147E+00 | grad norm: 3.503 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.372 | TFLOPs: 40.88 | 31: iteration 50/ 52452 | consumed samples: 25600 | consumed tokens: 52428800 | elapsed time per iteration (s): 1.88 | learning rate: 1.906E-05 | global batch size: 512 | lm loss: 8.052976E+00 | grad norm: 7.545 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.856 | TFLOPs: 40.95 | 31: iteration 60/ 52452 | consumed samples: 30720 | consumed tokens: 62914560 | elapsed time per iteration (s): 1.86 | learning rate: 2.288E-05 | global batch size: 512 | lm loss: 7.653759E+00 | grad norm: 3.138 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.879 | TFLOPs: 41.26 | 31: iteration 70/ 52452 | consumed samples: 35840 | consumed tokens: 73400320 | elapsed time per iteration (s): 1.85 | learning rate: 2.669E-05 | global batch size: 512 | lm loss: 7.344408E+00 | grad norm: 3.065 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.092 | TFLOPs: 41.44 | 31: iteration 80/ 52452 | consumed samples: 40960 | consumed tokens: 83886080 | elapsed time per iteration (s): 1.91 | learning rate: 3.050E-05 | global batch size: 512 | lm loss: 7.089422E+00 | grad norm: 3.015 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.787 | TFLOPs: 40.19 | 31: iteration 90/ 52452 | consumed samples: 46080 | consumed tokens: 94371840 | elapsed time per iteration (s): 1.85 | learning rate: 3.432E-05 | global batch size: 512 | lm loss: 6.867936E+00 | grad norm: 3.986 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.143 | TFLOPs: 41.60 | 31: iteration 100/ 52452 | consumed samples: 51200 | consumed tokens: 104857600 | elapsed time per iteration (s): 2.03 | learning rate: 3.813E-05 | global batch size: 512 | lm loss: 6.742662E+00 | grad norm: 4.912 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 252.830 | TFLOPs: 37.95 | 31: ------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 100 | lm loss value: 7.275565E+00 | lm loss PPL: 1.444567E+03 | 31: ------------------------------------------------------------------------------------------------- 31: ---------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 100 | lm loss value: 7.345697E+00 | lm loss PPL: 1.549514E+03 | 31: ---------------------------------------------------------------------------------------------------- 31: iteration 110/ 52452 | consumed samples: 56320 | consumed tokens: 115343360 | elapsed time per iteration (s): 9.22 | learning rate: 4.194E-05 | global batch size: 512 | lm loss: 6.588409E+00 | grad norm: 1.688 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.504 | TFLOPs: 8.33 | 31: iteration 120/ 52452 | consumed samples: 61440 | consumed tokens: 125829120 | elapsed time per iteration (s): 1.88 | learning rate: 4.576E-05 | global batch size: 512 | lm loss: 6.512106E+00 | grad norm: 2.863 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.520 | TFLOPs: 40.90 | 31: iteration 130/ 52452 | consumed samples: 66560 | consumed tokens: 136314880 | elapsed time per iteration (s): 1.89 | learning rate: 4.957E-05 | global batch size: 512 | lm loss: 6.389154E+00 | grad norm: 2.781 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.130 | TFLOPs: 40.70 | 31: iteration 140/ 52452 | consumed samples: 71680 | consumed tokens: 146800640 | elapsed time per iteration (s): 1.86 | learning rate: 5.338E-05 | global batch size: 512 | lm loss: 6.278399E+00 | grad norm: 2.126 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.193 | TFLOPs: 41.31 | 31: iteration 150/ 52452 | consumed samples: 76800 | consumed tokens: 157286400 | elapsed time per iteration (s): 1.82 | learning rate: 5.719E-05 | global batch size: 512 | lm loss: 6.184066E+00 | grad norm: 2.926 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.435 | TFLOPs: 42.24 | 31: iteration 160/ 52452 | consumed samples: 81920 | consumed tokens: 167772160 | elapsed time per iteration (s): 1.86 | learning rate: 6.101E-05 | global batch size: 512 | lm loss: 6.113217E+00 | grad norm: 3.269 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.923 | TFLOPs: 41.26 | 31: iteration 170/ 52452 | consumed samples: 87040 | consumed tokens: 178257920 | elapsed time per iteration (s): 1.83 | learning rate: 6.482E-05 | global batch size: 512 | lm loss: 6.045766E+00 | grad norm: 3.117 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.864 | TFLOPs: 42.01 | 31: iteration 180/ 52452 | consumed samples: 92160 | consumed tokens: 188743680 | elapsed time per iteration (s): 2.07 | learning rate: 6.863E-05 | global batch size: 512 | lm loss: 5.969768E+00 | grad norm: 1.973 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 247.479 | TFLOPs: 37.15 | 31: iteration 190/ 52452 | consumed samples: 97280 | consumed tokens: 199229440 | elapsed time per iteration (s): 1.83 | learning rate: 7.245E-05 | global batch size: 512 | lm loss: 5.911575E+00 | grad norm: 1.847 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.514 | TFLOPs: 41.95 | 31: iteration 200/ 52452 | consumed samples: 102400 | consumed tokens: 209715200 | elapsed time per iteration (s): 1.86 | learning rate: 7.626E-05 | global batch size: 512 | lm loss: 5.824476E+00 | grad norm: 1.945 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.626 | TFLOPs: 41.22 | 31: ------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 200 | lm loss value: 6.645358E+00 | lm loss PPL: 7.692054E+02 | 31: ------------------------------------------------------------------------------------------------- 31: ---------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 200 | lm loss value: 6.735088E+00 | lm loss PPL: 8.414178E+02 | 31: ---------------------------------------------------------------------------------------------------- 31: iteration 210/ 52452 | consumed samples: 107520 | consumed tokens: 220200960 | elapsed time per iteration (s): 9.25 | learning rate: 8.007E-05 | global batch size: 512 | lm loss: 5.803768E+00 | grad norm: 2.097 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.354 | TFLOPs: 8.31 | 31: iteration 220/ 52452 | consumed samples: 112640 | consumed tokens: 230686720 | elapsed time per iteration (s): 1.86 | learning rate: 8.389E-05 | global batch size: 512 | lm loss: 5.687029E+00 | grad norm: 1.630 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.133 | TFLOPs: 41.30 | 31: iteration 230/ 52452 | consumed samples: 117760 | consumed tokens: 241172480 | elapsed time per iteration (s): 1.83 | learning rate: 8.770E-05 | global batch size: 512 | lm loss: 5.632665E+00 | grad norm: 1.277 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.138 | TFLOPs: 41.90 | 31: iteration 240/ 52452 | consumed samples: 122880 | consumed tokens: 251658240 | elapsed time per iteration (s): 1.89 | learning rate: 9.151E-05 | global batch size: 512 | lm loss: 5.602747E+00 | grad norm: 1.690 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.141 | TFLOPs: 40.70 | 31: iteration 250/ 52452 | consumed samples: 128000 | consumed tokens: 262144000 | elapsed time per iteration (s): 1.87 | learning rate: 9.532E-05 | global batch size: 512 | lm loss: 5.554870E+00 | grad norm: 1.117 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.438 | TFLOPs: 41.19 | 31: iteration 260/ 52452 | consumed samples: 133120 | consumed tokens: 272629760 | elapsed time per iteration (s): 1.86 | learning rate: 9.914E-05 | global batch size: 512 | lm loss: 5.502147E+00 | grad norm: 1.681 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.396 | TFLOPs: 41.34 | 31: iteration 270/ 52452 | consumed samples: 138240 | consumed tokens: 283115520 | elapsed time per iteration (s): 1.90 | learning rate: 1.030E-04 | global batch size: 512 | lm loss: 5.450171E+00 | grad norm: 1.261 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.511 | TFLOPs: 40.45 | 31: iteration 280/ 52452 | consumed samples: 143360 | consumed tokens: 293601280 | elapsed time per iteration (s): 2.17 | learning rate: 1.068E-04 | global batch size: 512 | lm loss: 5.368832E+00 | grad norm: 1.634 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 236.239 | TFLOPs: 35.46 | 31: iteration 290/ 52452 | consumed samples: 148480 | consumed tokens: 304087040 | elapsed time per iteration (s): 1.96 | learning rate: 1.106E-04 | global batch size: 512 | lm loss: 5.347374E+00 | grad norm: 1.333 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 261.751 | TFLOPs: 39.29 | 31: iteration 300/ 52452 | consumed samples: 153600 | consumed tokens: 314572800 | elapsed time per iteration (s): 1.85 | learning rate: 1.144E-04 | global batch size: 512 | lm loss: 5.294761E+00 | grad norm: 1.130 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.044 | TFLOPs: 41.58 | 31: ------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 300 | lm loss value: 6.400434E+00 | lm loss PPL: 6.021063E+02 | 31: ------------------------------------------------------------------------------------------------- 31: ---------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 300 | lm loss value: 6.507608E+00 | lm loss PPL: 6.702216E+02 | 31: ---------------------------------------------------------------------------------------------------- 31: iteration 310/ 52452 | consumed samples: 158720 | consumed tokens: 325058560 | elapsed time per iteration (s): 9.26 | learning rate: 1.182E-04 | global batch size: 512 | lm loss: 5.247649E+00 | grad norm: 1.113 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.310 | TFLOPs: 8.30 | 31: iteration 320/ 52452 | consumed samples: 163840 | consumed tokens: 335544320 | elapsed time per iteration (s): 1.79 | learning rate: 1.220E-04 | global batch size: 512 | lm loss: 5.170983E+00 | grad norm: 1.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.831 | TFLOPs: 42.90 | 31: iteration 330/ 52452 | consumed samples: 168960 | consumed tokens: 346030080 | elapsed time per iteration (s): 1.90 | learning rate: 1.258E-04 | global batch size: 512 | lm loss: 5.129897E+00 | grad norm: 1.356 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.305 | TFLOPs: 40.42 | 31: iteration 340/ 52452 | consumed samples: 174080 | consumed tokens: 356515840 | elapsed time per iteration (s): 1.86 | learning rate: 1.296E-04 | global batch size: 512 | lm loss: 5.093025E+00 | grad norm: 0.924 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.833 | TFLOPs: 41.25 | 31: iteration 350/ 52452 | consumed samples: 179200 | consumed tokens: 367001600 | elapsed time per iteration (s): 1.85 | learning rate: 1.335E-04 | global batch size: 512 | lm loss: 5.008747E+00 | grad norm: 0.970 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.733 | TFLOPs: 41.54 | 31: iteration 360/ 52452 | consumed samples: 184320 | consumed tokens: 377487360 | elapsed time per iteration (s): 1.87 | learning rate: 1.373E-04 | global batch size: 512 | lm loss: 4.972164E+00 | grad norm: 1.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.076 | TFLOPs: 40.99 | 31: iteration 370/ 52452 | consumed samples: 189440 | consumed tokens: 387973120 | elapsed time per iteration (s): 1.86 | learning rate: 1.411E-04 | global batch size: 512 | lm loss: 4.906990E+00 | grad norm: 0.849 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.970 | TFLOPs: 41.42 | 31: iteration 380/ 52452 | consumed samples: 194560 | consumed tokens: 398458880 | elapsed time per iteration (s): 1.87 | learning rate: 1.449E-04 | global batch size: 512 | lm loss: 4.859923E+00 | grad norm: 1.043 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.157 | TFLOPs: 41.15 | 31: iteration 390/ 52452 | consumed samples: 199680 | consumed tokens: 408944640 | elapsed time per iteration (s): 1.82 | learning rate: 1.487E-04 | global batch size: 512 | lm loss: 4.813803E+00 | grad norm: 1.134 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.944 | TFLOPs: 42.17 | 31: iteration 400/ 52452 | consumed samples: 204800 | consumed tokens: 419430400 | elapsed time per iteration (s): 1.82 | learning rate: 1.525E-04 | global batch size: 512 | lm loss: 4.763961E+00 | grad norm: 0.900 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.646 | TFLOPs: 42.12 | 31: ------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 400 | lm loss value: 6.051097E+00 | lm loss PPL: 4.245787E+02 | 31: ------------------------------------------------------------------------------------------------- 31: ---------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 400 | lm loss value: 6.169158E+00 | lm loss PPL: 4.777836E+02 | 31: ---------------------------------------------------------------------------------------------------- 31: iteration 410/ 52452 | consumed samples: 209920 | consumed tokens: 429916160 | elapsed time per iteration (s): 9.23 | learning rate: 1.563E-04 | global batch size: 512 | lm loss: 4.683485E+00 | grad norm: 0.751 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.500 | TFLOPs: 8.33 | 31: iteration 420/ 52452 | consumed samples: 215040 | consumed tokens: 440401920 | elapsed time per iteration (s): 1.88 | learning rate: 1.601E-04 | global batch size: 512 | lm loss: 4.624284E+00 | grad norm: 0.817 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.898 | TFLOPs: 40.81 | 31: iteration 430/ 52452 | consumed samples: 220160 | consumed tokens: 450887680 | elapsed time per iteration (s): 1.80 | learning rate: 1.640E-04 | global batch size: 512 | lm loss: 4.588420E+00 | grad norm: 1.049 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.720 | TFLOPs: 42.58 | 31: iteration 440/ 52452 | consumed samples: 225280 | consumed tokens: 461373440 | elapsed time per iteration (s): 1.91 | learning rate: 1.678E-04 | global batch size: 512 | lm loss: 4.540294E+00 | grad norm: 1.012 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.283 | TFLOPs: 40.27 | 31: iteration 450/ 52452 | consumed samples: 230400 | consumed tokens: 471859200 | elapsed time per iteration (s): 1.91 | learning rate: 1.716E-04 | global batch size: 512 | lm loss: 4.478101E+00 | grad norm: 0.890 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.437 | TFLOPs: 40.29 | 31: iteration 460/ 52452 | consumed samples: 235520 | consumed tokens: 482344960 | elapsed time per iteration (s): 1.85 | learning rate: 1.754E-04 | global batch size: 512 | lm loss: 4.426796E+00 | grad norm: 0.997 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.884 | TFLOPs: 41.56 | 31: iteration 470/ 52452 | consumed samples: 240640 | consumed tokens: 492830720 | elapsed time per iteration (s): 1.85 | learning rate: 1.792E-04 | global batch size: 512 | lm loss: 4.453942E+00 | grad norm: 0.707 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.482 | TFLOPs: 41.50 | 31: iteration 480/ 52452 | consumed samples: 245760 | consumed tokens: 503316480 | elapsed time per iteration (s): 1.90 | learning rate: 1.830E-04 | global batch size: 512 | lm loss: 4.356557E+00 | grad norm: 0.825 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.331 | TFLOPs: 40.43 | 31: iteration 490/ 52452 | consumed samples: 250880 | consumed tokens: 513802240 | elapsed time per iteration (s): 1.87 | learning rate: 1.868E-04 | global batch size: 512 | lm loss: 4.301677E+00 | grad norm: 0.724 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.472 | TFLOPs: 41.20 | 31: iteration 500/ 52452 | consumed samples: 256000 | consumed tokens: 524288000 | elapsed time per iteration (s): 1.82 | learning rate: 1.906E-04 | global batch size: 512 | lm loss: 4.251322E+00 | grad norm: 0.786 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.602 | TFLOPs: 42.27 | 31: ------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 500 | lm loss value: 5.687281E+00 | lm loss PPL: 2.950902E+02 | 31: ------------------------------------------------------------------------------------------------- 31: ---------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 500 | lm loss value: 5.832468E+00 | lm loss PPL: 3.411996E+02 | 31: ---------------------------------------------------------------------------------------------------- 31: iteration 510/ 52452 | consumed samples: 261120 | consumed tokens: 534773760 | elapsed time per iteration (s): 9.25 | learning rate: 1.945E-04 | global batch size: 512 | lm loss: 4.213436E+00 | grad norm: 0.917 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.338 | TFLOPs: 8.31 | 31: iteration 520/ 52452 | consumed samples: 266240 | consumed tokens: 545259520 | elapsed time per iteration (s): 2.25 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 4.211595E+00 | grad norm: 0.693 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 227.263 | TFLOPs: 34.11 | 31: iteration 530/ 52452 | consumed samples: 271360 | consumed tokens: 555745280 | elapsed time per iteration (s): 1.85 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 4.145908E+00 | grad norm: 0.638 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.076 | TFLOPs: 41.59 | 31: iteration 540/ 52452 | consumed samples: 276480 | consumed tokens: 566231040 | elapsed time per iteration (s): 2.00 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 4.069089E+00 | grad norm: 0.711 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 255.737 | TFLOPs: 38.38 | 31: iteration 550/ 52452 | consumed samples: 281600 | consumed tokens: 576716800 | elapsed time per iteration (s): 2.01 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 4.033499E+00 | grad norm: 0.737 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 254.291 | TFLOPs: 38.17 | 31: iteration 560/ 52452 | consumed samples: 286720 | consumed tokens: 587202560 | elapsed time per iteration (s): 1.83 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.999212E+00 | grad norm: 0.659 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.864 | TFLOPs: 42.01 | 31: iteration 570/ 52452 | consumed samples: 291840 | consumed tokens: 597688320 | elapsed time per iteration (s): 1.86 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.969479E+00 | grad norm: 0.629 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.050 | TFLOPs: 41.28 | 31: iteration 580/ 52452 | consumed samples: 296960 | consumed tokens: 608174080 | elapsed time per iteration (s): 1.89 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.901509E+00 | grad norm: 0.753 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.035 | TFLOPs: 40.68 | 31: iteration 590/ 52452 | consumed samples: 302080 | consumed tokens: 618659840 | elapsed time per iteration (s): 1.86 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.865519E+00 | grad norm: 0.649 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.948 | TFLOPs: 41.42 | 31: iteration 600/ 52452 | consumed samples: 307200 | consumed tokens: 629145600 | elapsed time per iteration (s): 1.86 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.812341E+00 | grad norm: 0.505 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.712 | TFLOPs: 41.23 | 31: ------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 600 | lm loss value: 5.268315E+00 | lm loss PPL: 1.940886E+02 | 31: ------------------------------------------------------------------------------------------------- 31: ---------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 600 | lm loss value: 5.427415E+00 | lm loss PPL: 2.275602E+02 | 31: ---------------------------------------------------------------------------------------------------- 31: iteration 610/ 52452 | consumed samples: 312320 | consumed tokens: 639631360 | elapsed time per iteration (s): 9.25 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.765653E+00 | grad norm: 0.803 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.344 | TFLOPs: 8.31 | 31: iteration 620/ 52452 | consumed samples: 317440 | consumed tokens: 650117120 | elapsed time per iteration (s): 1.86 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.737234E+00 | grad norm: 0.735 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.724 | TFLOPs: 41.38 | 31: iteration 630/ 52452 | consumed samples: 322560 | consumed tokens: 660602880 | elapsed time per iteration (s): 1.84 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.693987E+00 | grad norm: 0.653 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.172 | TFLOPs: 41.75 | 31: iteration 640/ 52452 | consumed samples: 327680 | consumed tokens: 671088640 | elapsed time per iteration (s): 1.89 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.621192E+00 | grad norm: 0.486 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.564 | TFLOPs: 40.61 | 31: iteration 650/ 52452 | consumed samples: 332800 | consumed tokens: 681574400 | elapsed time per iteration (s): 1.90 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.573552E+00 | grad norm: 0.728 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.568 | TFLOPs: 40.46 | 31: iteration 660/ 52452 | consumed samples: 337920 | consumed tokens: 692060160 | elapsed time per iteration (s): 1.91 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.541171E+00 | grad norm: 0.583 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.487 | TFLOPs: 40.30 | 31: iteration 670/ 52452 | consumed samples: 343040 | consumed tokens: 702545920 | elapsed time per iteration (s): 1.85 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.504540E+00 | grad norm: 0.488 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.336 | TFLOPs: 41.63 | 31: iteration 680/ 52452 | consumed samples: 348160 | consumed tokens: 713031680 | elapsed time per iteration (s): 1.97 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.465985E+00 | grad norm: 0.690 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 260.222 | TFLOPs: 39.06 | 31: iteration 690/ 52452 | consumed samples: 353280 | consumed tokens: 723517440 | elapsed time per iteration (s): 1.86 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.420718E+00 | grad norm: 0.632 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.346 | TFLOPs: 41.33 | 31: iteration 700/ 52452 | consumed samples: 358400 | consumed tokens: 734003200 | elapsed time per iteration (s): 1.86 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.381095E+00 | grad norm: 0.544 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.649 | TFLOPs: 41.22 | 31: ------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 700 | lm loss value: 4.825912E+00 | lm loss PPL: 1.247001E+02 | 31: ------------------------------------------------------------------------------------------------- 31: ---------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 700 | lm loss value: 4.976064E+00 | lm loss PPL: 1.449029E+02 | 31: ---------------------------------------------------------------------------------------------------- 31: iteration 710/ 52452 | consumed samples: 363520 | consumed tokens: 744488960 | elapsed time per iteration (s): 9.25 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.337080E+00 | grad norm: 0.484 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.352 | TFLOPs: 8.31 | 31: iteration 720/ 52452 | consumed samples: 368640 | consumed tokens: 754974720 | elapsed time per iteration (s): 1.81 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.313451E+00 | grad norm: 0.442 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.657 | TFLOPs: 42.43 | 31: iteration 730/ 52452 | consumed samples: 373760 | consumed tokens: 765460480 | elapsed time per iteration (s): 1.85 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.296896E+00 | grad norm: 0.619 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.612 | TFLOPs: 41.52 | 31: iteration 740/ 52452 | consumed samples: 378880 | consumed tokens: 775946240 | elapsed time per iteration (s): 1.89 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.250779E+00 | grad norm: 0.468 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.076 | TFLOPs: 40.69 | 31: iteration 750/ 52452 | consumed samples: 384000 | consumed tokens: 786432000 | elapsed time per iteration (s): 1.85 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.218459E+00 | grad norm: 0.465 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.146 | TFLOPs: 41.60 | 31: iteration 760/ 52452 | consumed samples: 389120 | consumed tokens: 796917760 | elapsed time per iteration (s): 1.94 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.188166E+00 | grad norm: 0.651 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 264.184 | TFLOPs: 39.65 | 31: iteration 770/ 52452 | consumed samples: 394240 | consumed tokens: 807403520 | elapsed time per iteration (s): 1.80 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.187783E+00 | grad norm: 0.427 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.906 | TFLOPs: 42.76 | 31: iteration 780/ 52452 | consumed samples: 399360 | consumed tokens: 817889280 | elapsed time per iteration (s): 1.92 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.151949E+00 | grad norm: 0.457 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.955 | TFLOPs: 40.07 | 31: iteration 790/ 52452 | consumed samples: 404480 | consumed tokens: 828375040 | elapsed time per iteration (s): 1.83 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.146856E+00 | grad norm: 0.403 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.473 | TFLOPs: 41.95 | 31: iteration 800/ 52452 | consumed samples: 409600 | consumed tokens: 838860800 | elapsed time per iteration (s): 1.93 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.112155E+00 | grad norm: 0.451 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 264.668 | TFLOPs: 39.73 | 31: ------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 800 | lm loss value: 4.590991E+00 | lm loss PPL: 9.859204E+01 | 31: ------------------------------------------------------------------------------------------------- 31: ---------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 800 | lm loss value: 4.753726E+00 | lm loss PPL: 1.160158E+02 | 31: ---------------------------------------------------------------------------------------------------- 31: iteration 810/ 52452 | consumed samples: 414720 | consumed tokens: 849346560 | elapsed time per iteration (s): 9.45 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.094068E+00 | grad norm: 0.422 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.171 | TFLOPs: 8.13 | 31: iteration 820/ 52452 | consumed samples: 419840 | consumed tokens: 859832320 | elapsed time per iteration (s): 2.01 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.081815E+00 | grad norm: 0.475 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 255.088 | TFLOPs: 38.29 | 31: iteration 830/ 52452 | consumed samples: 424960 | consumed tokens: 870318080 | elapsed time per iteration (s): 1.97 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.059820E+00 | grad norm: 0.324 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 260.378 | TFLOPs: 39.08 | 31: iteration 840/ 52452 | consumed samples: 430080 | consumed tokens: 880803840 | elapsed time per iteration (s): 1.83 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.033769E+00 | grad norm: 0.539 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.340 | TFLOPs: 42.08 | 31: iteration 850/ 52452 | consumed samples: 435200 | consumed tokens: 891289600 | elapsed time per iteration (s): 1.90 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.001630E+00 | grad norm: 0.455 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.487 | TFLOPs: 40.45 | 31: iteration 860/ 52452 | consumed samples: 440320 | consumed tokens: 901775360 | elapsed time per iteration (s): 1.82 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 3.018016E+00 | grad norm: 0.490 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.955 | TFLOPs: 42.32 | 31: iteration 870/ 52452 | consumed samples: 445440 | consumed tokens: 912261120 | elapsed time per iteration (s): 1.88 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.992616E+00 | grad norm: 0.352 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.933 | TFLOPs: 40.82 | 31: iteration 880/ 52452 | consumed samples: 450560 | consumed tokens: 922746880 | elapsed time per iteration (s): 1.95 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.970200E+00 | grad norm: 0.368 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 262.343 | TFLOPs: 39.38 | 31: iteration 890/ 52452 | consumed samples: 455680 | consumed tokens: 933232640 | elapsed time per iteration (s): 1.77 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.937753E+00 | grad norm: 0.351 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.637 | TFLOPs: 43.32 | 31: iteration 900/ 52452 | consumed samples: 460800 | consumed tokens: 943718400 | elapsed time per iteration (s): 1.84 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.933951E+00 | grad norm: 0.380 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.129 | TFLOPs: 41.75 | 31: ------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 900 | lm loss value: 4.444854E+00 | lm loss PPL: 8.518746E+01 | 31: ------------------------------------------------------------------------------------------------- 31: ---------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 900 | lm loss value: 4.603800E+00 | lm loss PPL: 9.986306E+01 | 31: ---------------------------------------------------------------------------------------------------- 31: iteration 910/ 52452 | consumed samples: 465920 | consumed tokens: 954204160 | elapsed time per iteration (s): 9.25 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.939408E+00 | grad norm: 0.360 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.344 | TFLOPs: 8.31 | 31: iteration 920/ 52452 | consumed samples: 471040 | consumed tokens: 964689920 | elapsed time per iteration (s): 1.82 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.908836E+00 | grad norm: 0.356 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.022 | TFLOPs: 42.18 | 31: iteration 930/ 52452 | consumed samples: 476160 | consumed tokens: 975175680 | elapsed time per iteration (s): 1.80 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.894007E+00 | grad norm: 0.452 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.943 | TFLOPs: 42.62 | 31: iteration 940/ 52452 | consumed samples: 481280 | consumed tokens: 985661440 | elapsed time per iteration (s): 1.96 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.881154E+00 | grad norm: 0.347 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 261.050 | TFLOPs: 39.18 | 31: iteration 950/ 52452 | consumed samples: 486400 | consumed tokens: 996147200 | elapsed time per iteration (s): 2.02 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.859365E+00 | grad norm: 0.336 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 253.774 | TFLOPs: 38.09 | 31: iteration 960/ 52452 | consumed samples: 491520 | consumed tokens: 1006632960 | elapsed time per iteration (s): 1.95 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.861636E+00 | grad norm: 0.308 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 262.583 | TFLOPs: 39.41 | 31: iteration 970/ 52452 | consumed samples: 496640 | consumed tokens: 1017118720 | elapsed time per iteration (s): 1.89 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.838286E+00 | grad norm: 0.326 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.914 | TFLOPs: 40.66 | 31: iteration 980/ 52452 | consumed samples: 501760 | consumed tokens: 1027604480 | elapsed time per iteration (s): 1.84 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.816518E+00 | grad norm: 0.356 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.860 | TFLOPs: 41.71 | 31: iteration 990/ 52452 | consumed samples: 506880 | consumed tokens: 1038090240 | elapsed time per iteration (s): 1.85 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.812188E+00 | grad norm: 0.343 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.244 | TFLOPs: 41.46 | 31: iteration 1000/ 52452 | consumed samples: 512000 | consumed tokens: 1048576000 | elapsed time per iteration (s): 1.88 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.809318E+00 | grad norm: 0.331 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.025 | TFLOPs: 40.98 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1000 | lm loss value: 4.326494E+00 | lm loss PPL: 7.567851E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1000 | lm loss value: 4.493692E+00 | lm loss PPL: 8.945108E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1010/ 52452 | consumed samples: 517120 | consumed tokens: 1059061760 | elapsed time per iteration (s): 9.39 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.794657E+00 | grad norm: 0.460 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.535 | TFLOPs: 8.19 | 31: iteration 1020/ 52452 | consumed samples: 522240 | consumed tokens: 1069547520 | elapsed time per iteration (s): 1.92 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.797105E+00 | grad norm: 0.369 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.337 | TFLOPs: 40.13 | 31: iteration 1030/ 52452 | consumed samples: 527360 | consumed tokens: 1080033280 | elapsed time per iteration (s): 1.92 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.760056E+00 | grad norm: 0.362 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.223 | TFLOPs: 39.96 | 31: iteration 1040/ 52452 | consumed samples: 532480 | consumed tokens: 1090519040 | elapsed time per iteration (s): 1.81 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.754323E+00 | grad norm: 0.369 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.218 | TFLOPs: 42.36 | 31: iteration 1050/ 52452 | consumed samples: 537600 | consumed tokens: 1101004800 | elapsed time per iteration (s): 1.95 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.745093E+00 | grad norm: 0.299 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 262.841 | TFLOPs: 39.45 | 31: iteration 1060/ 52452 | consumed samples: 542720 | consumed tokens: 1111490560 | elapsed time per iteration (s): 1.89 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.736992E+00 | grad norm: 0.366 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.936 | TFLOPs: 40.67 | 31: iteration 1070/ 52452 | consumed samples: 547840 | consumed tokens: 1121976320 | elapsed time per iteration (s): 2.01 | learning rate: 2.000E-04 | global batch size: 512 | lm loss: 2.739386E+00 | grad norm: 0.318 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 255.313 | TFLOPs: 38.32 | 31: iteration 1080/ 52452 | consumed samples: 552960 | consumed tokens: 1132462080 | elapsed time per iteration (s): 1.86 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.730316E+00 | grad norm: 0.296 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.245 | TFLOPs: 41.31 | 31: iteration 1090/ 52452 | consumed samples: 558080 | consumed tokens: 1142947840 | elapsed time per iteration (s): 1.83 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.712974E+00 | grad norm: 0.341 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.119 | TFLOPs: 41.89 | 31: iteration 1100/ 52452 | consumed samples: 563200 | consumed tokens: 1153433600 | elapsed time per iteration (s): 1.93 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.713455E+00 | grad norm: 0.356 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.182 | TFLOPs: 39.80 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1100 | lm loss value: 4.250059E+00 | lm loss PPL: 7.010952E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1100 | lm loss value: 4.412794E+00 | lm loss PPL: 8.249962E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1110/ 52452 | consumed samples: 568320 | consumed tokens: 1163919360 | elapsed time per iteration (s): 9.31 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.694572E+00 | grad norm: 0.335 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.982 | TFLOPs: 8.25 | 31: iteration 1120/ 52452 | consumed samples: 573440 | consumed tokens: 1174405120 | elapsed time per iteration (s): 1.91 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.684643E+00 | grad norm: 0.263 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.728 | TFLOPs: 40.18 | 31: iteration 1130/ 52452 | consumed samples: 578560 | consumed tokens: 1184890880 | elapsed time per iteration (s): 1.92 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.665287E+00 | grad norm: 0.366 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.337 | TFLOPs: 39.98 | 31: iteration 1140/ 52452 | consumed samples: 583680 | consumed tokens: 1195376640 | elapsed time per iteration (s): 1.91 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.660852E+00 | grad norm: 0.335 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.695 | TFLOPs: 40.33 | 31: iteration 1150/ 52452 | consumed samples: 588800 | consumed tokens: 1205862400 | elapsed time per iteration (s): 1.90 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.651842E+00 | grad norm: 0.309 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.407 | TFLOPs: 40.44 | 31: iteration 1160/ 52452 | consumed samples: 593920 | consumed tokens: 1216348160 | elapsed time per iteration (s): 1.83 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.636493E+00 | grad norm: 0.292 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.349 | TFLOPs: 42.08 | 31: iteration 1170/ 52452 | consumed samples: 599040 | consumed tokens: 1226833920 | elapsed time per iteration (s): 1.89 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.644927E+00 | grad norm: 0.313 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.953 | TFLOPs: 40.67 | 31: iteration 1180/ 52452 | consumed samples: 604160 | consumed tokens: 1237319680 | elapsed time per iteration (s): 1.84 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.635925E+00 | grad norm: 0.793 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.390 | TFLOPs: 41.78 | 31: iteration 1190/ 52452 | consumed samples: 609280 | consumed tokens: 1247805440 | elapsed time per iteration (s): 1.92 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.659677E+00 | grad norm: 0.253 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.123 | TFLOPs: 39.94 | 31: iteration 1200/ 52452 | consumed samples: 614400 | consumed tokens: 1258291200 | elapsed time per iteration (s): 1.84 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.614386E+00 | grad norm: 0.300 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.495 | TFLOPs: 41.80 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1200 | lm loss value: 4.190738E+00 | lm loss PPL: 6.607155E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1200 | lm loss value: 4.360785E+00 | lm loss PPL: 7.831863E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1210/ 52452 | consumed samples: 619520 | consumed tokens: 1268776960 | elapsed time per iteration (s): 9.32 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.599597E+00 | grad norm: 0.297 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.936 | TFLOPs: 8.25 | 31: iteration 1220/ 52452 | consumed samples: 624640 | consumed tokens: 1279262720 | elapsed time per iteration (s): 1.84 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.592484E+00 | grad norm: 0.247 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.887 | TFLOPs: 41.71 | 31: iteration 1230/ 52452 | consumed samples: 629760 | consumed tokens: 1289748480 | elapsed time per iteration (s): 1.82 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.581935E+00 | grad norm: 0.334 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.408 | TFLOPs: 42.24 | 31: iteration 1240/ 52452 | consumed samples: 634880 | consumed tokens: 1300234240 | elapsed time per iteration (s): 1.83 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.593055E+00 | grad norm: 0.291 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.517 | TFLOPs: 42.10 | 31: iteration 1250/ 52452 | consumed samples: 640000 | consumed tokens: 1310720000 | elapsed time per iteration (s): 1.82 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.567141E+00 | grad norm: 0.257 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.010 | TFLOPs: 42.18 | 31: iteration 1260/ 52452 | consumed samples: 645120 | consumed tokens: 1321205760 | elapsed time per iteration (s): 1.80 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.563013E+00 | grad norm: 0.294 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.698 | TFLOPs: 42.58 | 31: iteration 1270/ 52452 | consumed samples: 650240 | consumed tokens: 1331691520 | elapsed time per iteration (s): 1.90 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.563048E+00 | grad norm: 0.288 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.863 | TFLOPs: 40.35 | 31: iteration 1280/ 52452 | consumed samples: 655360 | consumed tokens: 1342177280 | elapsed time per iteration (s): 1.77 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.548402E+00 | grad norm: 0.274 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.957 | TFLOPs: 43.37 | 31: iteration 1290/ 52452 | consumed samples: 660480 | consumed tokens: 1352663040 | elapsed time per iteration (s): 1.92 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.550761E+00 | grad norm: 0.348 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.203 | TFLOPs: 40.11 | 31: iteration 1300/ 52452 | consumed samples: 665600 | consumed tokens: 1363148800 | elapsed time per iteration (s): 1.88 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.517755E+00 | grad norm: 0.247 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.729 | TFLOPs: 40.78 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1300 | lm loss value: 4.161338E+00 | lm loss PPL: 6.415733E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1300 | lm loss value: 4.320877E+00 | lm loss PPL: 7.525457E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1310/ 52452 | consumed samples: 670720 | consumed tokens: 1373634560 | elapsed time per iteration (s): 9.18 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.534088E+00 | grad norm: 0.287 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.768 | TFLOPs: 8.37 | 31: iteration 1320/ 52452 | consumed samples: 675840 | consumed tokens: 1384120320 | elapsed time per iteration (s): 1.84 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.513126E+00 | grad norm: 0.251 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.532 | TFLOPs: 41.81 | 31: iteration 1330/ 52452 | consumed samples: 680960 | consumed tokens: 1394606080 | elapsed time per iteration (s): 1.85 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.515134E+00 | grad norm: 0.275 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.594 | TFLOPs: 41.52 | 31: iteration 1340/ 52452 | consumed samples: 686080 | consumed tokens: 1405091840 | elapsed time per iteration (s): 1.81 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.523265E+00 | grad norm: 0.299 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.097 | TFLOPs: 42.34 | 31: iteration 1350/ 52452 | consumed samples: 691200 | consumed tokens: 1415577600 | elapsed time per iteration (s): 1.82 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.508097E+00 | grad norm: 0.267 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.733 | TFLOPs: 42.29 | 31: iteration 1360/ 52452 | consumed samples: 696320 | consumed tokens: 1426063360 | elapsed time per iteration (s): 1.78 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.521098E+00 | grad norm: 0.310 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.699 | TFLOPs: 43.18 | 31: iteration 1370/ 52452 | consumed samples: 701440 | consumed tokens: 1436549120 | elapsed time per iteration (s): 1.82 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.492543E+00 | grad norm: 0.245 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.094 | TFLOPs: 42.19 | 31: iteration 1380/ 52452 | consumed samples: 706560 | consumed tokens: 1447034880 | elapsed time per iteration (s): 1.94 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.490520E+00 | grad norm: 0.317 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 263.429 | TFLOPs: 39.54 | 31: iteration 1390/ 52452 | consumed samples: 711680 | consumed tokens: 1457520640 | elapsed time per iteration (s): 2.05 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.467697E+00 | grad norm: 0.244 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 249.492 | TFLOPs: 37.45 | 31: iteration 1400/ 52452 | consumed samples: 716800 | consumed tokens: 1468006400 | elapsed time per iteration (s): 1.85 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.471105E+00 | grad norm: 0.316 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.516 | TFLOPs: 41.50 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1400 | lm loss value: 4.178710E+00 | lm loss PPL: 6.528158E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1400 | lm loss value: 4.342018E+00 | lm loss PPL: 7.686246E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1410/ 52452 | consumed samples: 721920 | consumed tokens: 1478492160 | elapsed time per iteration (s): 9.27 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.479249E+00 | grad norm: 0.265 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.252 | TFLOPs: 8.29 | 31: iteration 1420/ 52452 | consumed samples: 727040 | consumed tokens: 1488977920 | elapsed time per iteration (s): 1.82 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.467210E+00 | grad norm: 0.237 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.663 | TFLOPs: 42.28 | 31: iteration 1430/ 52452 | consumed samples: 732160 | consumed tokens: 1499463680 | elapsed time per iteration (s): 1.82 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.455416E+00 | grad norm: 0.280 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.019 | TFLOPs: 42.18 | 31: iteration 1440/ 52452 | consumed samples: 737280 | consumed tokens: 1509949440 | elapsed time per iteration (s): 1.80 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.446665E+00 | grad norm: 0.248 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.070 | TFLOPs: 42.79 | 31: iteration 1450/ 52452 | consumed samples: 742400 | consumed tokens: 1520435200 | elapsed time per iteration (s): 1.75 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.444228E+00 | grad norm: 0.241 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.191 | TFLOPs: 43.86 | 31: iteration 1460/ 52452 | consumed samples: 747520 | consumed tokens: 1530920960 | elapsed time per iteration (s): 1.77 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.445520E+00 | grad norm: 0.256 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.674 | TFLOPs: 43.48 | 31: iteration 1470/ 52452 | consumed samples: 752640 | consumed tokens: 1541406720 | elapsed time per iteration (s): 1.76 | learning rate: 1.999E-04 | global batch size: 512 | lm loss: 2.442048E+00 | grad norm: 0.262 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.985 | TFLOPs: 43.68 | 31: iteration 1480/ 52452 | consumed samples: 757760 | consumed tokens: 1551892480 | elapsed time per iteration (s): 1.86 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.433017E+00 | grad norm: 0.250 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.248 | TFLOPs: 41.31 | 31: iteration 1490/ 52452 | consumed samples: 762880 | consumed tokens: 1562378240 | elapsed time per iteration (s): 1.79 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.406534E+00 | grad norm: 0.284 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.452 | TFLOPs: 42.99 | 31: iteration 1500/ 52452 | consumed samples: 768000 | consumed tokens: 1572864000 | elapsed time per iteration (s): 1.78 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.426611E+00 | grad norm: 0.229 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.846 | TFLOPs: 43.20 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1500 | lm loss value: 4.114090E+00 | lm loss PPL: 6.119650E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1500 | lm loss value: 4.275443E+00 | lm loss PPL: 7.191196E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1510/ 52452 | consumed samples: 773120 | consumed tokens: 1583349760 | elapsed time per iteration (s): 9.40 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.415854E+00 | grad norm: 0.254 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.475 | TFLOPs: 8.18 | 31: iteration 1520/ 52452 | consumed samples: 778240 | consumed tokens: 1593835520 | elapsed time per iteration (s): 1.76 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.406710E+00 | grad norm: 0.271 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.586 | TFLOPs: 43.77 | 31: iteration 1530/ 52452 | consumed samples: 783360 | consumed tokens: 1604321280 | elapsed time per iteration (s): 1.97 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.391866E+00 | grad norm: 0.294 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 259.384 | TFLOPs: 38.93 | 31: iteration 1540/ 52452 | consumed samples: 788480 | consumed tokens: 1614807040 | elapsed time per iteration (s): 1.79 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.396277E+00 | grad norm: 0.300 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.127 | TFLOPs: 42.95 | 31: iteration 1550/ 52452 | consumed samples: 793600 | consumed tokens: 1625292800 | elapsed time per iteration (s): 1.80 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.386953E+00 | grad norm: 0.316 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.926 | TFLOPs: 42.77 | 31: iteration 1560/ 52452 | consumed samples: 798720 | consumed tokens: 1635778560 | elapsed time per iteration (s): 1.78 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.394049E+00 | grad norm: 0.232 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.123 | TFLOPs: 43.25 | 31: iteration 1570/ 52452 | consumed samples: 803840 | consumed tokens: 1646264320 | elapsed time per iteration (s): 1.76 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.382565E+00 | grad norm: 0.231 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.584 | TFLOPs: 43.62 | 31: iteration 1580/ 52452 | consumed samples: 808960 | consumed tokens: 1656750080 | elapsed time per iteration (s): 1.80 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.378256E+00 | grad norm: 0.264 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.914 | TFLOPs: 42.76 | 31: iteration 1590/ 52452 | consumed samples: 814080 | consumed tokens: 1667235840 | elapsed time per iteration (s): 1.82 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.375176E+00 | grad norm: 0.259 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.772 | TFLOPs: 42.14 | 31: iteration 1600/ 52452 | consumed samples: 819200 | consumed tokens: 1677721600 | elapsed time per iteration (s): 1.79 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.372506E+00 | grad norm: 0.257 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.297 | TFLOPs: 42.97 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1600 | lm loss value: 4.109067E+00 | lm loss PPL: 6.088991E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1600 | lm loss value: 4.266405E+00 | lm loss PPL: 7.126498E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1610/ 52452 | consumed samples: 824320 | consumed tokens: 1688207360 | elapsed time per iteration (s): 9.21 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.363181E+00 | grad norm: 0.253 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.569 | TFLOPs: 8.34 | 31: iteration 1620/ 52452 | consumed samples: 829440 | consumed tokens: 1698693120 | elapsed time per iteration (s): 1.81 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.354122E+00 | grad norm: 0.234 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.500 | TFLOPs: 42.40 | 31: iteration 1630/ 52452 | consumed samples: 834560 | consumed tokens: 1709178880 | elapsed time per iteration (s): 1.76 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.352578E+00 | grad norm: 0.267 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.821 | TFLOPs: 43.65 | 31: iteration 1640/ 52452 | consumed samples: 839680 | consumed tokens: 1719664640 | elapsed time per iteration (s): 1.76 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.337732E+00 | grad norm: 0.241 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.785 | TFLOPs: 43.65 | 31: iteration 1650/ 52452 | consumed samples: 844800 | consumed tokens: 1730150400 | elapsed time per iteration (s): 1.79 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.332994E+00 | grad norm: 0.333 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.484 | TFLOPs: 42.85 | 31: iteration 1660/ 52452 | consumed samples: 849920 | consumed tokens: 1740636160 | elapsed time per iteration (s): 1.77 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.333482E+00 | grad norm: 0.228 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.298 | TFLOPs: 43.42 | 31: iteration 1670/ 52452 | consumed samples: 855040 | consumed tokens: 1751121920 | elapsed time per iteration (s): 1.80 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.335176E+00 | grad norm: 0.245 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.917 | TFLOPs: 42.61 | 31: iteration 1680/ 52452 | consumed samples: 860160 | consumed tokens: 1761607680 | elapsed time per iteration (s): 1.81 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.312275E+00 | grad norm: 0.236 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.399 | TFLOPs: 42.54 | 31: iteration 1690/ 52452 | consumed samples: 865280 | consumed tokens: 1772093440 | elapsed time per iteration (s): 1.74 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.322449E+00 | grad norm: 0.270 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.015 | TFLOPs: 44.13 | 31: iteration 1700/ 52452 | consumed samples: 870400 | consumed tokens: 1782579200 | elapsed time per iteration (s): 2.16 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.329347E+00 | grad norm: 0.238 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 237.431 | TFLOPs: 35.64 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1700 | lm loss value: 4.104767E+00 | lm loss PPL: 6.062861E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1700 | lm loss value: 4.254084E+00 | lm loss PPL: 7.039228E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1710/ 52452 | consumed samples: 875520 | consumed tokens: 1793064960 | elapsed time per iteration (s): 9.32 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.310240E+00 | grad norm: 0.235 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.947 | TFLOPs: 8.25 | 31: iteration 1720/ 52452 | consumed samples: 880640 | consumed tokens: 1803550720 | elapsed time per iteration (s): 1.74 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.307111E+00 | grad norm: 0.279 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.535 | TFLOPs: 44.06 | 31: iteration 1730/ 52452 | consumed samples: 885760 | consumed tokens: 1814036480 | elapsed time per iteration (s): 1.76 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.306101E+00 | grad norm: 0.293 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.498 | TFLOPs: 43.75 | 31: iteration 1740/ 52452 | consumed samples: 890880 | consumed tokens: 1824522240 | elapsed time per iteration (s): 1.80 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.301260E+00 | grad norm: 0.231 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.702 | TFLOPs: 42.58 | 31: iteration 1750/ 52452 | consumed samples: 896000 | consumed tokens: 1835008000 | elapsed time per iteration (s): 1.80 | learning rate: 1.998E-04 | global batch size: 512 | lm loss: 2.313811E+00 | grad norm: 0.256 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.083 | TFLOPs: 42.64 | 31: iteration 1760/ 52452 | consumed samples: 901120 | consumed tokens: 1845493760 | elapsed time per iteration (s): 1.77 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.283211E+00 | grad norm: 0.256 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.649 | TFLOPs: 43.47 | 31: iteration 1770/ 52452 | consumed samples: 906240 | consumed tokens: 1855979520 | elapsed time per iteration (s): 1.83 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.299396E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.666 | TFLOPs: 41.98 | 31: iteration 1780/ 52452 | consumed samples: 911360 | consumed tokens: 1866465280 | elapsed time per iteration (s): 1.78 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.287945E+00 | grad norm: 0.254 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.249 | TFLOPs: 43.11 | 31: iteration 1790/ 52452 | consumed samples: 916480 | consumed tokens: 1876951040 | elapsed time per iteration (s): 1.76 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.293478E+00 | grad norm: 0.238 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.485 | TFLOPs: 43.60 | 31: iteration 1800/ 52452 | consumed samples: 921600 | consumed tokens: 1887436800 | elapsed time per iteration (s): 1.87 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.302132E+00 | grad norm: 0.248 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.491 | TFLOPs: 41.20 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1800 | lm loss value: 4.206954E+00 | lm loss PPL: 6.715172E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1800 | lm loss value: 4.329056E+00 | lm loss PPL: 7.587261E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1810/ 52452 | consumed samples: 926720 | consumed tokens: 1897922560 | elapsed time per iteration (s): 9.18 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.287769E+00 | grad norm: 0.219 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.781 | TFLOPs: 8.37 | 31: iteration 1820/ 52452 | consumed samples: 931840 | consumed tokens: 1908408320 | elapsed time per iteration (s): 1.79 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.280946E+00 | grad norm: 0.282 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.516 | TFLOPs: 43.00 | 31: iteration 1830/ 52452 | consumed samples: 936960 | consumed tokens: 1918894080 | elapsed time per iteration (s): 1.78 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.270232E+00 | grad norm: 0.228 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.496 | TFLOPs: 43.15 | 31: iteration 1840/ 52452 | consumed samples: 942080 | consumed tokens: 1929379840 | elapsed time per iteration (s): 1.81 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.268418E+00 | grad norm: 0.212 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.884 | TFLOPs: 42.46 | 31: iteration 1850/ 52452 | consumed samples: 947200 | consumed tokens: 1939865600 | elapsed time per iteration (s): 1.91 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.267126E+00 | grad norm: 0.255 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.769 | TFLOPs: 40.19 | 31: iteration 1860/ 52452 | consumed samples: 952320 | consumed tokens: 1950351360 | elapsed time per iteration (s): 1.83 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.266653E+00 | grad norm: 0.324 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.399 | TFLOPs: 42.09 | 31: iteration 1870/ 52452 | consumed samples: 957440 | consumed tokens: 1960837120 | elapsed time per iteration (s): 1.76 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.257384E+00 | grad norm: 0.238 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.339 | TFLOPs: 43.73 | 31: iteration 1880/ 52452 | consumed samples: 962560 | consumed tokens: 1971322880 | elapsed time per iteration (s): 1.79 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.250763E+00 | grad norm: 0.233 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.846 | TFLOPs: 42.90 | 31: iteration 1890/ 52452 | consumed samples: 967680 | consumed tokens: 1981808640 | elapsed time per iteration (s): 1.79 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.251007E+00 | grad norm: 0.289 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.481 | TFLOPs: 42.85 | 31: iteration 1900/ 52452 | consumed samples: 972800 | consumed tokens: 1992294400 | elapsed time per iteration (s): 1.84 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.236802E+00 | grad norm: 0.276 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.513 | TFLOPs: 41.65 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 1900 | lm loss value: 4.187717E+00 | lm loss PPL: 6.587223E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 1900 | lm loss value: 4.314370E+00 | lm loss PPL: 7.476648E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 1910/ 52452 | consumed samples: 977920 | consumed tokens: 2002780160 | elapsed time per iteration (s): 9.38 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.237670E+00 | grad norm: 0.229 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.613 | TFLOPs: 8.20 | 31: iteration 1920/ 52452 | consumed samples: 983040 | consumed tokens: 2013265920 | elapsed time per iteration (s): 1.82 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.244890E+00 | grad norm: 0.235 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.498 | TFLOPs: 42.25 | 31: iteration 1930/ 52452 | consumed samples: 988160 | consumed tokens: 2023751680 | elapsed time per iteration (s): 1.78 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.245317E+00 | grad norm: 0.208 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.645 | TFLOPs: 43.17 | 31: iteration 1940/ 52452 | consumed samples: 993280 | consumed tokens: 2034237440 | elapsed time per iteration (s): 1.85 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.226467E+00 | grad norm: 0.257 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.693 | TFLOPs: 41.53 | 31: iteration 1950/ 52452 | consumed samples: 998400 | consumed tokens: 2044723200 | elapsed time per iteration (s): 1.77 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.216088E+00 | grad norm: 0.210 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.675 | TFLOPs: 43.33 | 31: iteration 1960/ 52452 | consumed samples: 1003520 | consumed tokens: 2055208960 | elapsed time per iteration (s): 1.82 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.215478E+00 | grad norm: 0.237 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.856 | TFLOPs: 42.15 | 31: iteration 1970/ 52452 | consumed samples: 1008640 | consumed tokens: 2065694720 | elapsed time per iteration (s): 1.83 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.213828E+00 | grad norm: 0.241 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.288 | TFLOPs: 41.92 | 31: iteration 1980/ 52452 | consumed samples: 1013760 | consumed tokens: 2076180480 | elapsed time per iteration (s): 1.78 | learning rate: 1.997E-04 | global batch size: 512 | lm loss: 2.221838E+00 | grad norm: 0.252 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.910 | TFLOPs: 43.06 | 31: iteration 1990/ 52452 | consumed samples: 1018880 | consumed tokens: 2086666240 | elapsed time per iteration (s): 1.82 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.222433E+00 | grad norm: 0.242 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.632 | TFLOPs: 42.27 | 0: [2022-12-25 10:25:37,011] [INFO] [logging.py:68:log_dist] [Rank 0] step=2000, skipped=0, lr=[0.00019964166131491873, 0.00019964166131491873, 0.00019964166131491873], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 2000/ 52452 | consumed samples: 1024000 | consumed tokens: 2097152000 | elapsed time per iteration (s): 1.79 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.220377E+00 | grad norm: 0.220 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.308 | TFLOPs: 42.97 | 0: steps: 2000 loss: 2.2417 iter time (s): 1.874 samples/sec: 273.255 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2000 | lm loss value: 4.126097E+00 | lm loss PPL: 6.193573E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2000 | lm loss value: 4.244226E+00 | lm loss PPL: 6.970182E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2010/ 52452 | consumed samples: 1029120 | consumed tokens: 2107637760 | elapsed time per iteration (s): 9.19 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.216848E+00 | grad norm: 0.228 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.716 | TFLOPs: 8.36 | 31: iteration 2020/ 52452 | consumed samples: 1034240 | consumed tokens: 2118123520 | elapsed time per iteration (s): 1.73 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.222127E+00 | grad norm: 0.234 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.589 | TFLOPs: 44.52 | 31: iteration 2030/ 52452 | consumed samples: 1039360 | consumed tokens: 2128609280 | elapsed time per iteration (s): 1.82 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.194226E+00 | grad norm: 0.241 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.292 | TFLOPs: 42.22 | 31: iteration 2040/ 52452 | consumed samples: 1044480 | consumed tokens: 2139095040 | elapsed time per iteration (s): 1.78 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.213974E+00 | grad norm: 0.226 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.918 | TFLOPs: 43.06 | 31: iteration 2050/ 52452 | consumed samples: 1049600 | consumed tokens: 2149580800 | elapsed time per iteration (s): 1.77 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.197175E+00 | grad norm: 0.309 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.483 | TFLOPs: 43.45 | 31: iteration 2060/ 52452 | consumed samples: 1054720 | consumed tokens: 2160066560 | elapsed time per iteration (s): 1.78 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.205816E+00 | grad norm: 0.231 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.226 | TFLOPs: 43.11 | 31: iteration 2070/ 52452 | consumed samples: 1059840 | consumed tokens: 2170552320 | elapsed time per iteration (s): 1.78 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.177304E+00 | grad norm: 0.217 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.238 | TFLOPs: 43.26 | 31: iteration 2080/ 52452 | consumed samples: 1064960 | consumed tokens: 2181038080 | elapsed time per iteration (s): 1.81 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.196838E+00 | grad norm: 0.250 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.913 | TFLOPs: 42.46 | 31: iteration 2090/ 52452 | consumed samples: 1070080 | consumed tokens: 2191523840 | elapsed time per iteration (s): 1.73 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.188564E+00 | grad norm: 0.227 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.534 | TFLOPs: 44.51 | 31: iteration 2100/ 52452 | consumed samples: 1075200 | consumed tokens: 2202009600 | elapsed time per iteration (s): 1.76 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.182225E+00 | grad norm: 0.243 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.120 | TFLOPs: 43.55 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2100 | lm loss value: 4.259101E+00 | lm loss PPL: 7.074638E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2100 | lm loss value: 4.394331E+00 | lm loss PPL: 8.099047E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2110/ 52452 | consumed samples: 1080320 | consumed tokens: 2212495360 | elapsed time per iteration (s): 9.22 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.168559E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.515 | TFLOPs: 8.33 | 31: iteration 2120/ 52452 | consumed samples: 1085440 | consumed tokens: 2222981120 | elapsed time per iteration (s): 1.76 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.169071E+00 | grad norm: 0.223 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.278 | TFLOPs: 43.72 | 31: iteration 2130/ 52452 | consumed samples: 1090560 | consumed tokens: 2233466880 | elapsed time per iteration (s): 1.96 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.175920E+00 | grad norm: 0.231 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 261.345 | TFLOPs: 39.23 | 31: iteration 2140/ 52452 | consumed samples: 1095680 | consumed tokens: 2243952640 | elapsed time per iteration (s): 2.23 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.171521E+00 | grad norm: 0.217 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 229.175 | TFLOPs: 34.40 | 31: iteration 2150/ 52452 | consumed samples: 1100800 | consumed tokens: 2254438400 | elapsed time per iteration (s): 1.84 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.170005E+00 | grad norm: 0.236 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.647 | TFLOPs: 41.67 | 31: iteration 2160/ 52452 | consumed samples: 1105920 | consumed tokens: 2264924160 | elapsed time per iteration (s): 1.78 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.171892E+00 | grad norm: 0.231 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.531 | TFLOPs: 43.16 | 31: iteration 2170/ 52452 | consumed samples: 1111040 | consumed tokens: 2275409920 | elapsed time per iteration (s): 1.75 | learning rate: 1.996E-04 | global batch size: 512 | lm loss: 2.159880E+00 | grad norm: 0.210 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.080 | TFLOPs: 43.99 | 31: iteration 2180/ 52452 | consumed samples: 1116160 | consumed tokens: 2285895680 | elapsed time per iteration (s): 1.78 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.164633E+00 | grad norm: 0.258 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.847 | TFLOPs: 43.05 | 31: iteration 2190/ 52452 | consumed samples: 1121280 | consumed tokens: 2296381440 | elapsed time per iteration (s): 1.89 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.173326E+00 | grad norm: 0.216 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.278 | TFLOPs: 40.72 | 31: iteration 2200/ 52452 | consumed samples: 1126400 | consumed tokens: 2306867200 | elapsed time per iteration (s): 1.75 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.147247E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.181 | TFLOPs: 43.85 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2200 | lm loss value: 4.147679E+00 | lm loss PPL: 6.328693E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2200 | lm loss value: 4.283089E+00 | lm loss PPL: 7.246391E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2210/ 52452 | consumed samples: 1131520 | consumed tokens: 2317352960 | elapsed time per iteration (s): 9.18 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.162297E+00 | grad norm: 0.245 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.756 | TFLOPs: 8.37 | 31: iteration 2220/ 52452 | consumed samples: 1136640 | consumed tokens: 2327838720 | elapsed time per iteration (s): 1.80 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.137921E+00 | grad norm: 0.221 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.097 | TFLOPs: 42.79 | 31: iteration 2230/ 52452 | consumed samples: 1141760 | consumed tokens: 2338324480 | elapsed time per iteration (s): 1.78 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.144449E+00 | grad norm: 0.236 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.293 | TFLOPs: 43.12 | 31: iteration 2240/ 52452 | consumed samples: 1146880 | consumed tokens: 2348810240 | elapsed time per iteration (s): 1.78 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.148088E+00 | grad norm: 0.220 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.802 | TFLOPs: 43.20 | 31: iteration 2250/ 52452 | consumed samples: 1152000 | consumed tokens: 2359296000 | elapsed time per iteration (s): 1.80 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.163690E+00 | grad norm: 0.240 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.387 | TFLOPs: 42.68 | 31: iteration 2260/ 52452 | consumed samples: 1157120 | consumed tokens: 2369781760 | elapsed time per iteration (s): 1.78 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.134958E+00 | grad norm: 0.233 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.144 | TFLOPs: 43.25 | 31: iteration 2270/ 52452 | consumed samples: 1162240 | consumed tokens: 2380267520 | elapsed time per iteration (s): 1.75 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.135485E+00 | grad norm: 0.258 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.341 | TFLOPs: 44.03 | 31: iteration 2280/ 52452 | consumed samples: 1167360 | consumed tokens: 2390753280 | elapsed time per iteration (s): 1.78 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.128428E+00 | grad norm: 0.289 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.422 | TFLOPs: 43.29 | 31: iteration 2290/ 52452 | consumed samples: 1172480 | consumed tokens: 2401239040 | elapsed time per iteration (s): 1.86 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.155876E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.375 | TFLOPs: 41.33 | 31: iteration 2300/ 52452 | consumed samples: 1177600 | consumed tokens: 2411724800 | elapsed time per iteration (s): 1.75 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.130609E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.365 | TFLOPs: 43.88 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2300 | lm loss value: 4.212324E+00 | lm loss PPL: 6.751324E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2300 | lm loss value: 4.336645E+00 | lm loss PPL: 7.645059E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2310/ 52452 | consumed samples: 1182720 | consumed tokens: 2422210560 | elapsed time per iteration (s): 9.43 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.130050E+00 | grad norm: 0.220 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.268 | TFLOPs: 8.15 | 31: iteration 2320/ 52452 | consumed samples: 1187840 | consumed tokens: 2432696320 | elapsed time per iteration (s): 1.81 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.134148E+00 | grad norm: 0.224 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.945 | TFLOPs: 42.47 | 31: iteration 2330/ 52452 | consumed samples: 1192960 | consumed tokens: 2443182080 | elapsed time per iteration (s): 1.78 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.126935E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.050 | TFLOPs: 43.23 | 31: iteration 2340/ 52452 | consumed samples: 1198080 | consumed tokens: 2453667840 | elapsed time per iteration (s): 1.77 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.124427E+00 | grad norm: 0.213 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.199 | TFLOPs: 43.41 | 31: iteration 2350/ 52452 | consumed samples: 1203200 | consumed tokens: 2464153600 | elapsed time per iteration (s): 1.83 | learning rate: 1.995E-04 | global batch size: 512 | lm loss: 2.119286E+00 | grad norm: 0.242 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.438 | TFLOPs: 42.09 | 31: iteration 2360/ 52452 | consumed samples: 1208320 | consumed tokens: 2474639360 | elapsed time per iteration (s): 1.73 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.118450E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.984 | TFLOPs: 44.43 | 31: iteration 2370/ 52452 | consumed samples: 1213440 | consumed tokens: 2485125120 | elapsed time per iteration (s): 1.74 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.123664E+00 | grad norm: 0.226 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.806 | TFLOPs: 44.10 | 31: iteration 2380/ 52452 | consumed samples: 1218560 | consumed tokens: 2495610880 | elapsed time per iteration (s): 1.81 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.120362E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.342 | TFLOPs: 42.53 | 31: iteration 2390/ 52452 | consumed samples: 1223680 | consumed tokens: 2506096640 | elapsed time per iteration (s): 1.77 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.125275E+00 | grad norm: 0.248 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.569 | TFLOPs: 43.31 | 31: iteration 2400/ 52452 | consumed samples: 1228800 | consumed tokens: 2516582400 | elapsed time per iteration (s): 1.79 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.107343E+00 | grad norm: 0.236 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.678 | TFLOPs: 43.03 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2400 | lm loss value: 4.109704E+00 | lm loss PPL: 6.092868E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2400 | lm loss value: 4.236693E+00 | lm loss PPL: 6.917873E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2410/ 52452 | consumed samples: 1233920 | consumed tokens: 2527068160 | elapsed time per iteration (s): 9.17 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.103328E+00 | grad norm: 0.219 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.848 | TFLOPs: 8.38 | 31: iteration 2420/ 52452 | consumed samples: 1239040 | consumed tokens: 2537553920 | elapsed time per iteration (s): 1.84 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.100473E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.922 | TFLOPs: 41.86 | 31: iteration 2430/ 52452 | consumed samples: 1244160 | consumed tokens: 2548039680 | elapsed time per iteration (s): 1.97 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.103883E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 259.720 | TFLOPs: 38.98 | 31: iteration 2440/ 52452 | consumed samples: 1249280 | consumed tokens: 2558525440 | elapsed time per iteration (s): 1.84 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.104460E+00 | grad norm: 0.214 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.687 | TFLOPs: 41.68 | 31: iteration 2450/ 52452 | consumed samples: 1254400 | consumed tokens: 2569011200 | elapsed time per iteration (s): 1.77 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.096441E+00 | grad norm: 0.217 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.524 | TFLOPs: 43.46 | 31: iteration 2460/ 52452 | consumed samples: 1259520 | consumed tokens: 2579496960 | elapsed time per iteration (s): 1.76 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.110228E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.964 | TFLOPs: 43.67 | 31: iteration 2470/ 52452 | consumed samples: 1264640 | consumed tokens: 2589982720 | elapsed time per iteration (s): 1.73 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.091040E+00 | grad norm: 0.233 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.823 | TFLOPs: 44.40 | 31: iteration 2480/ 52452 | consumed samples: 1269760 | consumed tokens: 2600468480 | elapsed time per iteration (s): 1.76 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.089377E+00 | grad norm: 0.227 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.767 | TFLOPs: 43.64 | 31: iteration 2490/ 52452 | consumed samples: 1274880 | consumed tokens: 2610954240 | elapsed time per iteration (s): 1.77 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.081820E+00 | grad norm: 0.252 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.972 | TFLOPs: 43.37 | 31: iteration 2500/ 52452 | consumed samples: 1280000 | consumed tokens: 2621440000 | elapsed time per iteration (s): 1.73 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.106739E+00 | grad norm: 0.224 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.565 | TFLOPs: 44.36 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2500 | lm loss value: 4.081419E+00 | lm loss PPL: 5.922948E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2500 | lm loss value: 4.199397E+00 | lm loss PPL: 6.664614E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2510/ 52452 | consumed samples: 1285120 | consumed tokens: 2631925760 | elapsed time per iteration (s): 9.22 | learning rate: 1.994E-04 | global batch size: 512 | lm loss: 2.112749E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.541 | TFLOPs: 8.34 | 31: iteration 2520/ 52452 | consumed samples: 1290240 | consumed tokens: 2642411520 | elapsed time per iteration (s): 1.77 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.081824E+00 | grad norm: 0.259 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.021 | TFLOPs: 43.53 | 31: iteration 2530/ 52452 | consumed samples: 1295360 | consumed tokens: 2652897280 | elapsed time per iteration (s): 1.76 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.082034E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.101 | TFLOPs: 43.69 | 31: iteration 2540/ 52452 | consumed samples: 1300480 | consumed tokens: 2663383040 | elapsed time per iteration (s): 1.76 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.084561E+00 | grad norm: 0.224 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.359 | TFLOPs: 43.58 | 31: iteration 2550/ 52452 | consumed samples: 1305600 | consumed tokens: 2673868800 | elapsed time per iteration (s): 1.84 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.080772E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.080 | TFLOPs: 41.74 | 31: iteration 2560/ 52452 | consumed samples: 1310720 | consumed tokens: 2684354560 | elapsed time per iteration (s): 1.82 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.081657E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.149 | TFLOPs: 42.20 | 31: iteration 2570/ 52452 | consumed samples: 1315840 | consumed tokens: 2694840320 | elapsed time per iteration (s): 1.74 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.082054E+00 | grad norm: 0.225 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.939 | TFLOPs: 44.12 | 31: iteration 2580/ 52452 | consumed samples: 1320960 | consumed tokens: 2705326080 | elapsed time per iteration (s): 1.74 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.079477E+00 | grad norm: 0.225 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.247 | TFLOPs: 44.16 | 31: iteration 2590/ 52452 | consumed samples: 1326080 | consumed tokens: 2715811840 | elapsed time per iteration (s): 1.72 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.086282E+00 | grad norm: 0.227 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.176 | TFLOPs: 44.60 | 31: iteration 2600/ 52452 | consumed samples: 1331200 | consumed tokens: 2726297600 | elapsed time per iteration (s): 2.15 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.073118E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 238.284 | TFLOPs: 35.77 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2600 | lm loss value: 4.158478E+00 | lm loss PPL: 6.397407E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2600 | lm loss value: 4.289813E+00 | lm loss PPL: 7.295283E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2610/ 52452 | consumed samples: 1336320 | consumed tokens: 2736783360 | elapsed time per iteration (s): 9.32 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.074055E+00 | grad norm: 0.325 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.944 | TFLOPs: 8.25 | 31: iteration 2620/ 52452 | consumed samples: 1341440 | consumed tokens: 2747269120 | elapsed time per iteration (s): 2.04 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.077488E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 250.806 | TFLOPs: 37.64 | 31: iteration 2630/ 52452 | consumed samples: 1346560 | consumed tokens: 2757754880 | elapsed time per iteration (s): 1.76 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.066844E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.102 | TFLOPs: 43.54 | 31: iteration 2640/ 52452 | consumed samples: 1351680 | consumed tokens: 2768240640 | elapsed time per iteration (s): 1.84 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.061147E+00 | grad norm: 0.213 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.624 | TFLOPs: 41.82 | 31: iteration 2650/ 52452 | consumed samples: 1356800 | consumed tokens: 2778726400 | elapsed time per iteration (s): 1.84 | learning rate: 1.993E-04 | global batch size: 512 | lm loss: 2.058057E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.701 | TFLOPs: 41.83 | 31: iteration 2660/ 52452 | consumed samples: 1361920 | consumed tokens: 2789212160 | elapsed time per iteration (s): 1.79 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.044144E+00 | grad norm: 0.224 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.630 | TFLOPs: 42.87 | 31: iteration 2670/ 52452 | consumed samples: 1367040 | consumed tokens: 2799697920 | elapsed time per iteration (s): 1.77 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.058587E+00 | grad norm: 0.216 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.892 | TFLOPs: 43.36 | 31: iteration 2680/ 52452 | consumed samples: 1372160 | consumed tokens: 2810183680 | elapsed time per iteration (s): 1.74 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.055681E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.327 | TFLOPs: 44.18 | 31: iteration 2690/ 52452 | consumed samples: 1377280 | consumed tokens: 2820669440 | elapsed time per iteration (s): 1.83 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.062053E+00 | grad norm: 0.217 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.647 | TFLOPs: 41.97 | 31: iteration 2700/ 52452 | consumed samples: 1382400 | consumed tokens: 2831155200 | elapsed time per iteration (s): 1.74 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.056883E+00 | grad norm: 0.210 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.295 | TFLOPs: 44.17 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2700 | lm loss value: 4.113917E+00 | lm loss PPL: 6.118594E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2700 | lm loss value: 4.245784E+00 | lm loss PPL: 6.981046E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2710/ 52452 | consumed samples: 1387520 | consumed tokens: 2841640960 | elapsed time per iteration (s): 9.15 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.038025E+00 | grad norm: 0.229 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.963 | TFLOPs: 8.40 | 31: iteration 2720/ 52452 | consumed samples: 1392640 | consumed tokens: 2852126720 | elapsed time per iteration (s): 1.74 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.043435E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.506 | TFLOPs: 44.20 | 31: iteration 2730/ 52452 | consumed samples: 1397760 | consumed tokens: 2862612480 | elapsed time per iteration (s): 1.74 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.040981E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.324 | TFLOPs: 44.18 | 31: iteration 2740/ 52452 | consumed samples: 1402880 | consumed tokens: 2873098240 | elapsed time per iteration (s): 2.33 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.048255E+00 | grad norm: 0.225 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 219.588 | TFLOPs: 32.96 | 31: iteration 2750/ 52452 | consumed samples: 1408000 | consumed tokens: 2883584000 | elapsed time per iteration (s): 1.77 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.042519E+00 | grad norm: 0.217 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.048 | TFLOPs: 43.53 | 31: iteration 2760/ 52452 | consumed samples: 1413120 | consumed tokens: 2894069760 | elapsed time per iteration (s): 1.78 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.042385E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.329 | TFLOPs: 43.28 | 31: iteration 2770/ 52452 | consumed samples: 1418240 | consumed tokens: 2904555520 | elapsed time per iteration (s): 2.07 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.036577E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 247.398 | TFLOPs: 37.13 | 31: iteration 2780/ 52452 | consumed samples: 1423360 | consumed tokens: 2915041280 | elapsed time per iteration (s): 1.75 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.029070E+00 | grad norm: 0.233 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.284 | TFLOPs: 43.87 | 31: iteration 2790/ 52452 | consumed samples: 1428480 | consumed tokens: 2925527040 | elapsed time per iteration (s): 1.78 | learning rate: 1.992E-04 | global batch size: 512 | lm loss: 2.043253E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.054 | TFLOPs: 43.24 | 31: iteration 2800/ 52452 | consumed samples: 1433600 | consumed tokens: 2936012800 | elapsed time per iteration (s): 1.84 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.033554E+00 | grad norm: 0.221 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.964 | TFLOPs: 41.72 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2800 | lm loss value: 4.161134E+00 | lm loss PPL: 6.414421E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2800 | lm loss value: 4.280032E+00 | lm loss PPL: 7.224276E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2810/ 52452 | consumed samples: 1438720 | consumed tokens: 2946498560 | elapsed time per iteration (s): 9.18 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.038633E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.783 | TFLOPs: 8.37 | 31: iteration 2820/ 52452 | consumed samples: 1443840 | consumed tokens: 2956984320 | elapsed time per iteration (s): 1.76 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.021413E+00 | grad norm: 0.234 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.333 | TFLOPs: 43.58 | 31: iteration 2830/ 52452 | consumed samples: 1448960 | consumed tokens: 2967470080 | elapsed time per iteration (s): 1.78 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.023367E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.852 | TFLOPs: 43.05 | 31: iteration 2840/ 52452 | consumed samples: 1454080 | consumed tokens: 2977955840 | elapsed time per iteration (s): 1.90 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.021554E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.976 | TFLOPs: 40.52 | 31: iteration 2850/ 52452 | consumed samples: 1459200 | consumed tokens: 2988441600 | elapsed time per iteration (s): 1.77 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.038783E+00 | grad norm: 0.217 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.925 | TFLOPs: 43.37 | 31: iteration 2860/ 52452 | consumed samples: 1464320 | consumed tokens: 2998927360 | elapsed time per iteration (s): 1.80 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.010540E+00 | grad norm: 0.229 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.727 | TFLOPs: 42.59 | 31: iteration 2870/ 52452 | consumed samples: 1469440 | consumed tokens: 3009413120 | elapsed time per iteration (s): 1.77 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.030637E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.595 | TFLOPs: 43.47 | 31: iteration 2880/ 52452 | consumed samples: 1474560 | consumed tokens: 3019898880 | elapsed time per iteration (s): 1.85 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.020925E+00 | grad norm: 0.291 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.437 | TFLOPs: 41.64 | 31: iteration 2890/ 52452 | consumed samples: 1479680 | consumed tokens: 3030384640 | elapsed time per iteration (s): 1.76 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.026197E+00 | grad norm: 0.210 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.097 | TFLOPs: 43.69 | 31: iteration 2900/ 52452 | consumed samples: 1484800 | consumed tokens: 3040870400 | elapsed time per iteration (s): 1.87 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.010757E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.208 | TFLOPs: 41.16 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 2900 | lm loss value: 4.256103E+00 | lm loss PPL: 7.053458E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 2900 | lm loss value: 4.338671E+00 | lm loss PPL: 7.660564E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 2910/ 52452 | consumed samples: 1489920 | consumed tokens: 3051356160 | elapsed time per iteration (s): 9.18 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.016177E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.765 | TFLOPs: 8.37 | 31: iteration 2920/ 52452 | consumed samples: 1495040 | consumed tokens: 3061841920 | elapsed time per iteration (s): 1.78 | learning rate: 1.991E-04 | global batch size: 512 | lm loss: 2.015213E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.626 | TFLOPs: 43.17 | 31: iteration 2930/ 52452 | consumed samples: 1500160 | consumed tokens: 3072327680 | elapsed time per iteration (s): 1.74 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 2.024556E+00 | grad norm: 0.244 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.516 | TFLOPs: 44.06 | 31: iteration 2940/ 52452 | consumed samples: 1505280 | consumed tokens: 3082813440 | elapsed time per iteration (s): 1.79 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 2.013317E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.280 | TFLOPs: 42.82 | 31: iteration 2950/ 52452 | consumed samples: 1510400 | consumed tokens: 3093299200 | elapsed time per iteration (s): 1.84 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 2.006723E+00 | grad norm: 0.218 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.068 | TFLOPs: 41.74 | 31: iteration 2960/ 52452 | consumed samples: 1515520 | consumed tokens: 3103784960 | elapsed time per iteration (s): 1.78 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 2.005670E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.372 | TFLOPs: 43.13 | 31: iteration 2970/ 52452 | consumed samples: 1520640 | consumed tokens: 3114270720 | elapsed time per iteration (s): 1.82 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 2.014482E+00 | grad norm: 0.216 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.859 | TFLOPs: 42.16 | 31: iteration 2980/ 52452 | consumed samples: 1525760 | consumed tokens: 3124756480 | elapsed time per iteration (s): 1.77 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 2.000734E+00 | grad norm: 0.211 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.834 | TFLOPs: 43.35 | 31: iteration 2990/ 52452 | consumed samples: 1530880 | consumed tokens: 3135242240 | elapsed time per iteration (s): 1.82 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 2.000909E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.061 | TFLOPs: 42.19 | 31: iteration 3000/ 52452 | consumed samples: 1536000 | consumed tokens: 3145728000 | elapsed time per iteration (s): 1.89 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 2.002061E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.180 | TFLOPs: 40.70 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3000 | lm loss value: 4.035001E+00 | lm loss PPL: 5.654299E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3000 | lm loss value: 4.144803E+00 | lm loss PPL: 6.310516E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3010/ 52452 | consumed samples: 1541120 | consumed tokens: 3156213760 | elapsed time per iteration (s): 9.22 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 1.998234E+00 | grad norm: 0.242 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.521 | TFLOPs: 8.33 | 31: iteration 3020/ 52452 | consumed samples: 1546240 | consumed tokens: 3166699520 | elapsed time per iteration (s): 1.86 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 2.008068E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.720 | TFLOPs: 41.38 | 31: iteration 3030/ 52452 | consumed samples: 1551360 | consumed tokens: 3177185280 | elapsed time per iteration (s): 1.84 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 1.993218E+00 | grad norm: 0.216 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.985 | TFLOPs: 41.87 | 31: iteration 3040/ 52452 | consumed samples: 1556480 | consumed tokens: 3187671040 | elapsed time per iteration (s): 1.79 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 1.993936E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.795 | TFLOPs: 43.05 | 31: iteration 3050/ 52452 | consumed samples: 1561600 | consumed tokens: 3198156800 | elapsed time per iteration (s): 1.86 | learning rate: 1.990E-04 | global batch size: 512 | lm loss: 1.998862E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.708 | TFLOPs: 41.23 | 31: iteration 3060/ 52452 | consumed samples: 1566720 | consumed tokens: 3208642560 | elapsed time per iteration (s): 1.81 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.989466E+00 | grad norm: 0.211 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.911 | TFLOPs: 42.46 | 31: iteration 3070/ 52452 | consumed samples: 1571840 | consumed tokens: 3219128320 | elapsed time per iteration (s): 1.82 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.986268E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.578 | TFLOPs: 42.26 | 31: iteration 3080/ 52452 | consumed samples: 1576960 | consumed tokens: 3229614080 | elapsed time per iteration (s): 1.85 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.994905E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.394 | TFLOPs: 41.64 | 31: iteration 3090/ 52452 | consumed samples: 1582080 | consumed tokens: 3240099840 | elapsed time per iteration (s): 1.78 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.986866E+00 | grad norm: 0.219 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.784 | TFLOPs: 43.19 | 31: iteration 3100/ 52452 | consumed samples: 1587200 | consumed tokens: 3250585600 | elapsed time per iteration (s): 1.80 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.985423E+00 | grad norm: 0.232 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.673 | TFLOPs: 42.73 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3100 | lm loss value: 4.070639E+00 | lm loss PPL: 5.859437E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3100 | lm loss value: 4.168194E+00 | lm loss PPL: 6.459867E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3110/ 52452 | consumed samples: 1592320 | consumed tokens: 3261071360 | elapsed time per iteration (s): 9.36 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.993820E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.686 | TFLOPs: 8.21 | 31: iteration 3120/ 52452 | consumed samples: 1597440 | consumed tokens: 3271557120 | elapsed time per iteration (s): 1.81 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.986446E+00 | grad norm: 0.223 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.987 | TFLOPs: 42.47 | 31: iteration 3130/ 52452 | consumed samples: 1602560 | consumed tokens: 3282042880 | elapsed time per iteration (s): 1.81 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.992658E+00 | grad norm: 0.230 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.384 | TFLOPs: 42.53 | 31: iteration 3140/ 52452 | consumed samples: 1607680 | consumed tokens: 3292528640 | elapsed time per iteration (s): 1.78 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.967000E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.958 | TFLOPs: 43.07 | 31: iteration 3150/ 52452 | consumed samples: 1612800 | consumed tokens: 3303014400 | elapsed time per iteration (s): 1.95 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.978978E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 262.567 | TFLOPs: 39.41 | 31: iteration 3160/ 52452 | consumed samples: 1617920 | consumed tokens: 3313500160 | elapsed time per iteration (s): 1.80 | learning rate: 1.989E-04 | global batch size: 512 | lm loss: 1.984805E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.961 | TFLOPs: 42.77 | 31: iteration 3170/ 52452 | consumed samples: 1623040 | consumed tokens: 3323985920 | elapsed time per iteration (s): 1.89 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.976173E+00 | grad norm: 0.211 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.345 | TFLOPs: 40.73 | 31: iteration 3180/ 52452 | consumed samples: 1628160 | consumed tokens: 3334471680 | elapsed time per iteration (s): 1.91 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.968585E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.210 | TFLOPs: 40.26 | 31: iteration 3190/ 52452 | consumed samples: 1633280 | consumed tokens: 3344957440 | elapsed time per iteration (s): 1.83 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.975159E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.331 | TFLOPs: 41.93 | 31: iteration 3200/ 52452 | consumed samples: 1638400 | consumed tokens: 3355443200 | elapsed time per iteration (s): 1.76 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.951547E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.508 | TFLOPs: 43.60 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3200 | lm loss value: 4.061587E+00 | lm loss PPL: 5.806641E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3200 | lm loss value: 4.173786E+00 | lm loss PPL: 6.496091E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3210/ 52452 | consumed samples: 1643520 | consumed tokens: 3365928960 | elapsed time per iteration (s): 9.27 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.979169E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.241 | TFLOPs: 8.29 | 31: iteration 3220/ 52452 | consumed samples: 1648640 | consumed tokens: 3376414720 | elapsed time per iteration (s): 1.82 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.976428E+00 | grad norm: 0.220 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.995 | TFLOPs: 42.33 | 31: iteration 3230/ 52452 | consumed samples: 1653760 | consumed tokens: 3386900480 | elapsed time per iteration (s): 1.85 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.957280E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.034 | TFLOPs: 41.43 | 31: iteration 3240/ 52452 | consumed samples: 1658880 | consumed tokens: 3397386240 | elapsed time per iteration (s): 1.83 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.962453E+00 | grad norm: 0.216 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.317 | TFLOPs: 42.07 | 31: iteration 3250/ 52452 | consumed samples: 1664000 | consumed tokens: 3407872000 | elapsed time per iteration (s): 1.74 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.965471E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.459 | TFLOPs: 44.05 | 31: iteration 3260/ 52452 | consumed samples: 1669120 | consumed tokens: 3418357760 | elapsed time per iteration (s): 1.80 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.974787E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.275 | TFLOPs: 42.67 | 31: iteration 3270/ 52452 | consumed samples: 1674240 | consumed tokens: 3428843520 | elapsed time per iteration (s): 1.81 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.959095E+00 | grad norm: 0.236 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.217 | TFLOPs: 42.36 | 31: iteration 3280/ 52452 | consumed samples: 1679360 | consumed tokens: 3439329280 | elapsed time per iteration (s): 1.81 | learning rate: 1.988E-04 | global batch size: 512 | lm loss: 1.937669E+00 | grad norm: 0.219 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.608 | TFLOPs: 42.57 | 31: iteration 3290/ 52452 | consumed samples: 1684480 | consumed tokens: 3449815040 | elapsed time per iteration (s): 1.78 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.965552E+00 | grad norm: 0.245 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.467 | TFLOPs: 43.15 | 31: iteration 3300/ 52452 | consumed samples: 1689600 | consumed tokens: 3460300800 | elapsed time per iteration (s): 1.84 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.964733E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.516 | TFLOPs: 41.80 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3300 | lm loss value: 4.197056E+00 | lm loss PPL: 6.649031E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3300 | lm loss value: 4.299114E+00 | lm loss PPL: 7.363451E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3310/ 52452 | consumed samples: 1694720 | consumed tokens: 3470786560 | elapsed time per iteration (s): 9.26 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.961557E+00 | grad norm: 0.237 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.283 | TFLOPs: 8.30 | 31: iteration 3320/ 52452 | consumed samples: 1699840 | consumed tokens: 3481272320 | elapsed time per iteration (s): 1.88 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.950970E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.856 | TFLOPs: 40.95 | 31: iteration 3330/ 52452 | consumed samples: 1704960 | consumed tokens: 3491758080 | elapsed time per iteration (s): 1.91 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.956446E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.117 | TFLOPs: 40.24 | 31: iteration 3340/ 52452 | consumed samples: 1710080 | consumed tokens: 3502243840 | elapsed time per iteration (s): 1.83 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.954545E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.491 | TFLOPs: 41.95 | 31: iteration 3350/ 52452 | consumed samples: 1715200 | consumed tokens: 3512729600 | elapsed time per iteration (s): 1.86 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.952597E+00 | grad norm: 0.211 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.780 | TFLOPs: 41.39 | 31: iteration 3360/ 52452 | consumed samples: 1720320 | consumed tokens: 3523215360 | elapsed time per iteration (s): 1.86 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.961350E+00 | grad norm: 0.224 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.855 | TFLOPs: 41.40 | 31: iteration 3370/ 52452 | consumed samples: 1725440 | consumed tokens: 3533701120 | elapsed time per iteration (s): 1.83 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.948205E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.576 | TFLOPs: 41.96 | 31: iteration 3380/ 52452 | consumed samples: 1730560 | consumed tokens: 3544186880 | elapsed time per iteration (s): 1.84 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.958710E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.559 | TFLOPs: 41.66 | 31: iteration 3390/ 52452 | consumed samples: 1735680 | consumed tokens: 3554672640 | elapsed time per iteration (s): 1.87 | learning rate: 1.987E-04 | global batch size: 512 | lm loss: 1.938107E+00 | grad norm: 0.292 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.786 | TFLOPs: 41.09 | 31: iteration 3400/ 52452 | consumed samples: 1740800 | consumed tokens: 3565158400 | elapsed time per iteration (s): 1.85 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.942775E+00 | grad norm: 0.247 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.022 | TFLOPs: 41.58 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3400 | lm loss value: 4.065009E+00 | lm loss PPL: 5.826541E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3400 | lm loss value: 4.179653E+00 | lm loss PPL: 6.534315E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3410/ 52452 | consumed samples: 1745920 | consumed tokens: 3575644160 | elapsed time per iteration (s): 9.28 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.958600E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.192 | TFLOPs: 8.28 | 31: iteration 3420/ 52452 | consumed samples: 1751040 | consumed tokens: 3586129920 | elapsed time per iteration (s): 1.82 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.948062E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.230 | TFLOPs: 42.21 | 31: iteration 3430/ 52452 | consumed samples: 1756160 | consumed tokens: 3596615680 | elapsed time per iteration (s): 1.94 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.950633E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 263.593 | TFLOPs: 39.56 | 31: iteration 3440/ 52452 | consumed samples: 1761280 | consumed tokens: 3607101440 | elapsed time per iteration (s): 1.88 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.965705E+00 | grad norm: 0.210 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.469 | TFLOPs: 40.90 | 31: iteration 3450/ 52452 | consumed samples: 1766400 | consumed tokens: 3617587200 | elapsed time per iteration (s): 1.85 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.946974E+00 | grad norm: 0.238 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.289 | TFLOPs: 41.47 | 31: iteration 3460/ 52452 | consumed samples: 1771520 | consumed tokens: 3628072960 | elapsed time per iteration (s): 1.86 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.953240E+00 | grad norm: 0.253 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.666 | TFLOPs: 41.23 | 31: iteration 3470/ 52452 | consumed samples: 1776640 | consumed tokens: 3638558720 | elapsed time per iteration (s): 1.84 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.945054E+00 | grad norm: 0.535 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.689 | TFLOPs: 41.83 | 31: iteration 3480/ 52452 | consumed samples: 1781760 | consumed tokens: 3649044480 | elapsed time per iteration (s): 1.80 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.928890E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.181 | TFLOPs: 42.80 | 31: iteration 3490/ 52452 | consumed samples: 1786880 | consumed tokens: 3659530240 | elapsed time per iteration (s): 1.81 | learning rate: 1.986E-04 | global batch size: 512 | lm loss: 1.928238E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.614 | TFLOPs: 42.42 | 31: iteration 3500/ 52452 | consumed samples: 1792000 | consumed tokens: 3670016000 | elapsed time per iteration (s): 1.87 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.930400E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.716 | TFLOPs: 41.08 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3500 | lm loss value: 4.167128E+00 | lm loss PPL: 6.452986E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3500 | lm loss value: 4.244282E+00 | lm loss PPL: 6.970568E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3510/ 52452 | consumed samples: 1797120 | consumed tokens: 3680501760 | elapsed time per iteration (s): 9.25 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.930941E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.360 | TFLOPs: 8.31 | 31: iteration 3520/ 52452 | consumed samples: 1802240 | consumed tokens: 3690987520 | elapsed time per iteration (s): 1.76 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.930929E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.568 | TFLOPs: 43.76 | 31: iteration 3530/ 52452 | consumed samples: 1807360 | consumed tokens: 3701473280 | elapsed time per iteration (s): 1.88 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.929388E+00 | grad norm: 0.231 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.993 | TFLOPs: 40.97 | 31: iteration 3540/ 52452 | consumed samples: 1812480 | consumed tokens: 3711959040 | elapsed time per iteration (s): 1.86 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.923143E+00 | grad norm: 0.216 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.562 | TFLOPs: 41.21 | 31: iteration 3550/ 52452 | consumed samples: 1817600 | consumed tokens: 3722444800 | elapsed time per iteration (s): 1.80 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.909567E+00 | grad norm: 0.221 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.936 | TFLOPs: 42.62 | 31: iteration 3560/ 52452 | consumed samples: 1822720 | consumed tokens: 3732930560 | elapsed time per iteration (s): 1.83 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.915282E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.989 | TFLOPs: 42.02 | 31: iteration 3570/ 52452 | consumed samples: 1827840 | consumed tokens: 3743416320 | elapsed time per iteration (s): 1.83 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.927698E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.986 | TFLOPs: 42.02 | 31: iteration 3580/ 52452 | consumed samples: 1832960 | consumed tokens: 3753902080 | elapsed time per iteration (s): 1.82 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.923824E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.990 | TFLOPs: 42.18 | 31: iteration 3590/ 52452 | consumed samples: 1838080 | consumed tokens: 3764387840 | elapsed time per iteration (s): 1.82 | learning rate: 1.985E-04 | global batch size: 512 | lm loss: 1.911394E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.175 | TFLOPs: 42.20 | 31: iteration 3600/ 52452 | consumed samples: 1843200 | consumed tokens: 3774873600 | elapsed time per iteration (s): 1.77 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.920936E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.964 | TFLOPs: 43.52 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3600 | lm loss value: 4.102712E+00 | lm loss PPL: 6.050416E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3600 | lm loss value: 4.205354E+00 | lm loss PPL: 6.704434E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3610/ 52452 | consumed samples: 1848320 | consumed tokens: 3785359360 | elapsed time per iteration (s): 9.14 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.915927E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.006 | TFLOPs: 8.41 | 31: iteration 3620/ 52452 | consumed samples: 1853440 | consumed tokens: 3795845120 | elapsed time per iteration (s): 1.82 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.915335E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.749 | TFLOPs: 42.29 | 31: iteration 3630/ 52452 | consumed samples: 1858560 | consumed tokens: 3806330880 | elapsed time per iteration (s): 1.76 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.914969E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.087 | TFLOPs: 43.54 | 31: iteration 3640/ 52452 | consumed samples: 1863680 | consumed tokens: 3816816640 | elapsed time per iteration (s): 1.78 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.912978E+00 | grad norm: 0.219 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.492 | TFLOPs: 43.15 | 31: iteration 3650/ 52452 | consumed samples: 1868800 | consumed tokens: 3827302400 | elapsed time per iteration (s): 1.84 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.927619E+00 | grad norm: 0.229 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.137 | TFLOPs: 41.75 | 31: iteration 3660/ 52452 | consumed samples: 1873920 | consumed tokens: 3837788160 | elapsed time per iteration (s): 1.82 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.925462E+00 | grad norm: 0.231 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.326 | TFLOPs: 42.23 | 31: iteration 3670/ 52452 | consumed samples: 1879040 | consumed tokens: 3848273920 | elapsed time per iteration (s): 1.79 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.912369E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.781 | TFLOPs: 42.89 | 31: iteration 3680/ 52452 | consumed samples: 1884160 | consumed tokens: 3858759680 | elapsed time per iteration (s): 1.78 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.902952E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.940 | TFLOPs: 43.22 | 31: iteration 3690/ 52452 | consumed samples: 1889280 | consumed tokens: 3869245440 | elapsed time per iteration (s): 1.74 | learning rate: 1.984E-04 | global batch size: 512 | lm loss: 1.915249E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.448 | TFLOPs: 44.20 | 31: iteration 3700/ 52452 | consumed samples: 1894400 | consumed tokens: 3879731200 | elapsed time per iteration (s): 1.82 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 1.912918E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.814 | TFLOPs: 42.30 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3700 | lm loss value: 4.062294E+00 | lm loss PPL: 5.810743E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3700 | lm loss value: 4.162211E+00 | lm loss PPL: 6.421337E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3710/ 52452 | consumed samples: 1899520 | consumed tokens: 3890216960 | elapsed time per iteration (s): 9.15 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 1.909584E+00 | grad norm: 0.220 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.949 | TFLOPs: 8.40 | 31: iteration 3720/ 52452 | consumed samples: 1904640 | consumed tokens: 3900702720 | elapsed time per iteration (s): 1.84 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 1.911074E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.277 | TFLOPs: 41.77 | 31: iteration 3730/ 52452 | consumed samples: 1909760 | consumed tokens: 3911188480 | elapsed time per iteration (s): 1.74 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 1.915722E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.469 | TFLOPs: 44.05 | 31: iteration 3740/ 52452 | consumed samples: 1914880 | consumed tokens: 3921674240 | elapsed time per iteration (s): 1.80 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 1.904198E+00 | grad norm: 0.230 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.512 | TFLOPs: 42.70 | 31: iteration 3750/ 52452 | consumed samples: 1920000 | consumed tokens: 3932160000 | elapsed time per iteration (s): 1.78 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 1.920166E+00 | grad norm: 0.303 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.148 | TFLOPs: 43.25 | 31: iteration 3760/ 52452 | consumed samples: 1925120 | consumed tokens: 3942645760 | elapsed time per iteration (s): 1.76 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 1.910409E+00 | grad norm: 0.231 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.332 | TFLOPs: 43.58 | 31: iteration 3770/ 52452 | consumed samples: 1930240 | consumed tokens: 3953131520 | elapsed time per iteration (s): 1.80 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 1.906092E+00 | grad norm: 0.213 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.716 | TFLOPs: 42.58 | 31: iteration 3780/ 52452 | consumed samples: 1935360 | consumed tokens: 3963617280 | elapsed time per iteration (s): 1.88 | learning rate: 1.983E-04 | global batch size: 512 | lm loss: 1.903673E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.839 | TFLOPs: 40.80 | 31: iteration 3790/ 52452 | consumed samples: 1940480 | consumed tokens: 3974103040 | elapsed time per iteration (s): 1.76 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.904032E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.576 | TFLOPs: 43.61 | 31: iteration 3800/ 52452 | consumed samples: 1945600 | consumed tokens: 3984588800 | elapsed time per iteration (s): 1.81 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.879421E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.267 | TFLOPs: 42.52 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3800 | lm loss value: 4.015930E+00 | lm loss PPL: 5.547485E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3800 | lm loss value: 4.122065E+00 | lm loss PPL: 6.168650E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3810/ 52452 | consumed samples: 1950720 | consumed tokens: 3995074560 | elapsed time per iteration (s): 9.15 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.899162E+00 | grad norm: 0.225 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.941 | TFLOPs: 8.40 | 31: iteration 3820/ 52452 | consumed samples: 1955840 | consumed tokens: 4005560320 | elapsed time per iteration (s): 1.85 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.889918E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.709 | TFLOPs: 41.53 | 31: iteration 3830/ 52452 | consumed samples: 1960960 | consumed tokens: 4016046080 | elapsed time per iteration (s): 1.76 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.891380E+00 | grad norm: 0.217 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.739 | TFLOPs: 43.64 | 31: iteration 3840/ 52452 | consumed samples: 1966080 | consumed tokens: 4026531840 | elapsed time per iteration (s): 1.78 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.905329E+00 | grad norm: 0.208 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.374 | TFLOPs: 43.28 | 31: iteration 3850/ 52452 | consumed samples: 1971200 | consumed tokens: 4037017600 | elapsed time per iteration (s): 1.79 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.890014E+00 | grad norm: 0.218 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.347 | TFLOPs: 42.83 | 31: iteration 3860/ 52452 | consumed samples: 1976320 | consumed tokens: 4047503360 | elapsed time per iteration (s): 1.75 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.893276E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.969 | TFLOPs: 43.82 | 31: iteration 3870/ 52452 | consumed samples: 1981440 | consumed tokens: 4057989120 | elapsed time per iteration (s): 1.80 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.904195E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.182 | TFLOPs: 42.80 | 31: iteration 3880/ 52452 | consumed samples: 1986560 | consumed tokens: 4068474880 | elapsed time per iteration (s): 1.77 | learning rate: 1.982E-04 | global batch size: 512 | lm loss: 1.890030E+00 | grad norm: 0.216 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.064 | TFLOPs: 43.54 | 31: iteration 3890/ 52452 | consumed samples: 1991680 | consumed tokens: 4078960640 | elapsed time per iteration (s): 1.78 | learning rate: 1.981E-04 | global batch size: 512 | lm loss: 1.897483E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.359 | TFLOPs: 43.13 | 31: iteration 3900/ 52452 | consumed samples: 1996800 | consumed tokens: 4089446400 | elapsed time per iteration (s): 1.74 | learning rate: 1.981E-04 | global batch size: 512 | lm loss: 1.890769E+00 | grad norm: 0.231 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.989 | TFLOPs: 44.13 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 3900 | lm loss value: 4.098772E+00 | lm loss PPL: 6.026624E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 3900 | lm loss value: 4.225174E+00 | lm loss PPL: 6.838640E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 3910/ 52452 | consumed samples: 2001920 | consumed tokens: 4099932160 | elapsed time per iteration (s): 9.21 | learning rate: 1.981E-04 | global batch size: 512 | lm loss: 1.897002E+00 | grad norm: 0.219 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.601 | TFLOPs: 8.35 | 31: iteration 3920/ 52452 | consumed samples: 2007040 | consumed tokens: 4110417920 | elapsed time per iteration (s): 1.74 | learning rate: 1.981E-04 | global batch size: 512 | lm loss: 1.892576E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.562 | TFLOPs: 44.21 | 31: iteration 3930/ 52452 | consumed samples: 2012160 | consumed tokens: 4120903680 | elapsed time per iteration (s): 1.80 | learning rate: 1.981E-04 | global batch size: 512 | lm loss: 1.892996E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.652 | TFLOPs: 42.72 | 31: iteration 3940/ 52452 | consumed samples: 2017280 | consumed tokens: 4131389440 | elapsed time per iteration (s): 1.80 | learning rate: 1.981E-04 | global batch size: 512 | lm loss: 1.866684E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.026 | TFLOPs: 42.78 | 31: iteration 3950/ 52452 | consumed samples: 2022400 | consumed tokens: 4141875200 | elapsed time per iteration (s): 1.78 | learning rate: 1.981E-04 | global batch size: 512 | lm loss: 1.886159E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.939 | TFLOPs: 43.07 | 31: iteration 3960/ 52452 | consumed samples: 2027520 | consumed tokens: 4152360960 | elapsed time per iteration (s): 1.78 | learning rate: 1.981E-04 | global batch size: 512 | lm loss: 1.881382E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.484 | TFLOPs: 43.15 | 31: iteration 3970/ 52452 | consumed samples: 2032640 | consumed tokens: 4162846720 | elapsed time per iteration (s): 1.90 | learning rate: 1.981E-04 | global batch size: 512 | lm loss: 1.894619E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.602 | TFLOPs: 40.47 | 31: iteration 3980/ 52452 | consumed samples: 2037760 | consumed tokens: 4173332480 | elapsed time per iteration (s): 1.89 | learning rate: 1.980E-04 | global batch size: 512 | lm loss: 1.882084E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.175 | TFLOPs: 40.70 | 31: iteration 3990/ 52452 | consumed samples: 2042880 | consumed tokens: 4183818240 | elapsed time per iteration (s): 1.74 | learning rate: 1.980E-04 | global batch size: 512 | lm loss: 1.870502E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.915 | TFLOPs: 44.12 | 0: [2022-12-25 11:50:47,013] [INFO] [logging.py:68:log_dist] [Rank 0] step=4000, skipped=0, lr=[0.0001980178117989457, 0.0001980178117989457, 0.0001980178117989457], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 4000/ 52452 | consumed samples: 2048000 | consumed tokens: 4194304000 | elapsed time per iteration (s): 1.76 | learning rate: 1.980E-04 | global batch size: 512 | lm loss: 1.894848E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.202 | TFLOPs: 43.71 | 0: steps: 4000 loss: 1.8666 iter time (s): 1.814 samples/sec: 282.218 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4000 | lm loss value: 4.097601E+00 | lm loss PPL: 6.019573E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4000 | lm loss value: 4.181209E+00 | lm loss PPL: 6.544493E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4010/ 52452 | consumed samples: 2053120 | consumed tokens: 4204789760 | elapsed time per iteration (s): 9.29 | learning rate: 1.980E-04 | global batch size: 512 | lm loss: 1.872091E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.114 | TFLOPs: 8.27 | 31: iteration 4020/ 52452 | consumed samples: 2058240 | consumed tokens: 4215275520 | elapsed time per iteration (s): 1.88 | learning rate: 1.980E-04 | global batch size: 512 | lm loss: 1.884733E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.953 | TFLOPs: 40.97 | 31: iteration 4030/ 52452 | consumed samples: 2063360 | consumed tokens: 4225761280 | elapsed time per iteration (s): 1.80 | learning rate: 1.980E-04 | global batch size: 512 | lm loss: 1.863545E+00 | grad norm: 0.218 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.140 | TFLOPs: 42.65 | 31: iteration 4040/ 52452 | consumed samples: 2068480 | consumed tokens: 4236247040 | elapsed time per iteration (s): 1.73 | learning rate: 1.980E-04 | global batch size: 512 | lm loss: 1.865030E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.956 | TFLOPs: 44.42 | 31: iteration 4050/ 52452 | consumed samples: 2073600 | consumed tokens: 4246732800 | elapsed time per iteration (s): 1.84 | learning rate: 1.980E-04 | global batch size: 512 | lm loss: 1.870149E+00 | grad norm: 0.211 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.478 | TFLOPs: 41.80 | 31: iteration 4060/ 52452 | consumed samples: 2078720 | consumed tokens: 4257218560 | elapsed time per iteration (s): 1.75 | learning rate: 1.979E-04 | global batch size: 512 | lm loss: 1.871990E+00 | grad norm: 0.213 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.148 | TFLOPs: 43.85 | 31: iteration 4070/ 52452 | consumed samples: 2083840 | consumed tokens: 4267704320 | elapsed time per iteration (s): 1.80 | learning rate: 1.979E-04 | global batch size: 512 | lm loss: 1.874771E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.366 | TFLOPs: 42.68 | 31: iteration 4080/ 52452 | consumed samples: 2088960 | consumed tokens: 4278190080 | elapsed time per iteration (s): 1.77 | learning rate: 1.979E-04 | global batch size: 512 | lm loss: 1.854905E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.076 | TFLOPs: 43.54 | 31: iteration 4090/ 52452 | consumed samples: 2094080 | consumed tokens: 4288675840 | elapsed time per iteration (s): 1.81 | learning rate: 1.979E-04 | global batch size: 512 | lm loss: 1.865551E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.826 | TFLOPs: 42.45 | 31: iteration 4100/ 52452 | consumed samples: 2099200 | consumed tokens: 4299161600 | elapsed time per iteration (s): 1.76 | learning rate: 1.979E-04 | global batch size: 512 | lm loss: 1.864177E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.817 | TFLOPs: 43.65 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4100 | lm loss value: 4.211604E+00 | lm loss PPL: 6.746467E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4100 | lm loss value: 4.310816E+00 | lm loss PPL: 7.450124E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4110/ 52452 | consumed samples: 2104320 | consumed tokens: 4309647360 | elapsed time per iteration (s): 9.26 | learning rate: 1.979E-04 | global batch size: 512 | lm loss: 1.861209E+00 | grad norm: 0.235 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.308 | TFLOPs: 8.30 | 31: iteration 4120/ 52452 | consumed samples: 2109440 | consumed tokens: 4320133120 | elapsed time per iteration (s): 1.83 | learning rate: 1.979E-04 | global batch size: 512 | lm loss: 1.873147E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.538 | TFLOPs: 42.11 | 31: iteration 4130/ 52452 | consumed samples: 2114560 | consumed tokens: 4330618880 | elapsed time per iteration (s): 1.76 | learning rate: 1.979E-04 | global batch size: 512 | lm loss: 1.857338E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.404 | TFLOPs: 43.74 | 31: iteration 4140/ 52452 | consumed samples: 2119680 | consumed tokens: 4341104640 | elapsed time per iteration (s): 1.76 | learning rate: 1.979E-04 | global batch size: 512 | lm loss: 1.852905E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.655 | TFLOPs: 43.78 | 31: iteration 4150/ 52452 | consumed samples: 2124800 | consumed tokens: 4351590400 | elapsed time per iteration (s): 1.77 | learning rate: 1.978E-04 | global batch size: 512 | lm loss: 1.866323E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.398 | TFLOPs: 43.44 | 31: iteration 4160/ 52452 | consumed samples: 2129920 | consumed tokens: 4362076160 | elapsed time per iteration (s): 1.79 | learning rate: 1.978E-04 | global batch size: 512 | lm loss: 1.856295E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.661 | TFLOPs: 43.03 | 31: iteration 4170/ 52452 | consumed samples: 2135040 | consumed tokens: 4372561920 | elapsed time per iteration (s): 1.83 | learning rate: 1.978E-04 | global batch size: 512 | lm loss: 1.854831E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.276 | TFLOPs: 42.07 | 31: iteration 4180/ 52452 | consumed samples: 2140160 | consumed tokens: 4383047680 | elapsed time per iteration (s): 1.77 | learning rate: 1.978E-04 | global batch size: 512 | lm loss: 1.861077E+00 | grad norm: 0.217 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.053 | TFLOPs: 43.39 | 31: iteration 4190/ 52452 | consumed samples: 2145280 | consumed tokens: 4393533440 | elapsed time per iteration (s): 1.74 | learning rate: 1.978E-04 | global batch size: 512 | lm loss: 1.866445E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.610 | TFLOPs: 44.07 | 31: iteration 4200/ 52452 | consumed samples: 2150400 | consumed tokens: 4404019200 | elapsed time per iteration (s): 1.75 | learning rate: 1.978E-04 | global batch size: 512 | lm loss: 1.855097E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.630 | TFLOPs: 43.92 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4200 | lm loss value: 4.172153E+00 | lm loss PPL: 6.485497E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4200 | lm loss value: 4.277151E+00 | lm loss PPL: 7.203493E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4210/ 52452 | consumed samples: 2155520 | consumed tokens: 4414504960 | elapsed time per iteration (s): 9.17 | learning rate: 1.978E-04 | global batch size: 512 | lm loss: 1.860548E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.810 | TFLOPs: 8.38 | 31: iteration 4220/ 52452 | consumed samples: 2160640 | consumed tokens: 4424990720 | elapsed time per iteration (s): 1.82 | learning rate: 1.978E-04 | global batch size: 512 | lm loss: 1.865935E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.042 | TFLOPs: 42.18 | 31: iteration 4230/ 52452 | consumed samples: 2165760 | consumed tokens: 4435476480 | elapsed time per iteration (s): 1.75 | learning rate: 1.977E-04 | global batch size: 512 | lm loss: 1.865929E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.164 | TFLOPs: 44.00 | 31: iteration 4240/ 52452 | consumed samples: 2170880 | consumed tokens: 4445962240 | elapsed time per iteration (s): 1.88 | learning rate: 1.977E-04 | global batch size: 512 | lm loss: 1.850823E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.526 | TFLOPs: 40.90 | 31: iteration 4250/ 52452 | consumed samples: 2176000 | consumed tokens: 4456448000 | elapsed time per iteration (s): 1.79 | learning rate: 1.977E-04 | global batch size: 512 | lm loss: 1.859574E+00 | grad norm: 0.228 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.615 | TFLOPs: 43.02 | 31: iteration 4260/ 52452 | consumed samples: 2181120 | consumed tokens: 4466933760 | elapsed time per iteration (s): 1.81 | learning rate: 1.977E-04 | global batch size: 512 | lm loss: 1.862332E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.480 | TFLOPs: 42.55 | 31: iteration 4270/ 52452 | consumed samples: 2186240 | consumed tokens: 4477419520 | elapsed time per iteration (s): 1.85 | learning rate: 1.977E-04 | global batch size: 512 | lm loss: 1.846376E+00 | grad norm: 0.214 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.776 | TFLOPs: 41.54 | 31: iteration 4280/ 52452 | consumed samples: 2191360 | consumed tokens: 4487905280 | elapsed time per iteration (s): 1.78 | learning rate: 1.977E-04 | global batch size: 512 | lm loss: 1.838125E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.390 | TFLOPs: 43.14 | 31: iteration 4290/ 52452 | consumed samples: 2196480 | consumed tokens: 4498391040 | elapsed time per iteration (s): 1.83 | learning rate: 1.977E-04 | global batch size: 512 | lm loss: 1.853327E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.658 | TFLOPs: 41.98 | 31: iteration 4300/ 52452 | consumed samples: 2201600 | consumed tokens: 4508876800 | elapsed time per iteration (s): 1.79 | learning rate: 1.977E-04 | global batch size: 512 | lm loss: 1.847776E+00 | grad norm: 0.234 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.505 | TFLOPs: 42.85 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4300 | lm loss value: 4.066783E+00 | lm loss PPL: 5.836891E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4300 | lm loss value: 4.160692E+00 | lm loss PPL: 6.411589E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4310/ 52452 | consumed samples: 2206720 | consumed tokens: 4519362560 | elapsed time per iteration (s): 9.26 | learning rate: 1.977E-04 | global batch size: 512 | lm loss: 1.851592E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.264 | TFLOPs: 8.29 | 31: iteration 4320/ 52452 | consumed samples: 2211840 | consumed tokens: 4529848320 | elapsed time per iteration (s): 1.84 | learning rate: 1.976E-04 | global batch size: 512 | lm loss: 1.850008E+00 | grad norm: 0.208 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.533 | TFLOPs: 41.81 | 31: iteration 4330/ 52452 | consumed samples: 2216960 | consumed tokens: 4540334080 | elapsed time per iteration (s): 1.79 | learning rate: 1.976E-04 | global batch size: 512 | lm loss: 1.847086E+00 | grad norm: 0.211 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.656 | TFLOPs: 42.88 | 31: iteration 4340/ 52452 | consumed samples: 2222080 | consumed tokens: 4550819840 | elapsed time per iteration (s): 1.83 | learning rate: 1.976E-04 | global batch size: 512 | lm loss: 1.842714E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.231 | TFLOPs: 41.91 | 31: iteration 4350/ 52452 | consumed samples: 2227200 | consumed tokens: 4561305600 | elapsed time per iteration (s): 1.82 | learning rate: 1.976E-04 | global batch size: 512 | lm loss: 1.831102E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.631 | TFLOPs: 42.12 | 31: iteration 4360/ 52452 | consumed samples: 2232320 | consumed tokens: 4571791360 | elapsed time per iteration (s): 1.80 | learning rate: 1.976E-04 | global batch size: 512 | lm loss: 1.834717E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.412 | TFLOPs: 42.69 | 31: iteration 4370/ 52452 | consumed samples: 2237440 | consumed tokens: 4582277120 | elapsed time per iteration (s): 1.76 | learning rate: 1.976E-04 | global batch size: 512 | lm loss: 1.839542E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.405 | TFLOPs: 43.74 | 31: iteration 4380/ 52452 | consumed samples: 2242560 | consumed tokens: 4592762880 | elapsed time per iteration (s): 1.75 | learning rate: 1.976E-04 | global batch size: 512 | lm loss: 1.826838E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.907 | TFLOPs: 43.96 | 31: iteration 4390/ 52452 | consumed samples: 2247680 | consumed tokens: 4603248640 | elapsed time per iteration (s): 1.79 | learning rate: 1.976E-04 | global batch size: 512 | lm loss: 1.851950E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.341 | TFLOPs: 42.98 | 31: iteration 4400/ 52452 | consumed samples: 2252800 | consumed tokens: 4613734400 | elapsed time per iteration (s): 1.86 | learning rate: 1.975E-04 | global batch size: 512 | lm loss: 1.847224E+00 | grad norm: 0.211 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.664 | TFLOPs: 41.38 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4400 | lm loss value: 4.042306E+00 | lm loss PPL: 5.695753E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4400 | lm loss value: 4.154374E+00 | lm loss PPL: 6.371205E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4410/ 52452 | consumed samples: 2257920 | consumed tokens: 4624220160 | elapsed time per iteration (s): 9.14 | learning rate: 1.975E-04 | global batch size: 512 | lm loss: 1.835991E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.035 | TFLOPs: 8.41 | 31: iteration 4420/ 52452 | consumed samples: 2263040 | consumed tokens: 4634705920 | elapsed time per iteration (s): 1.84 | learning rate: 1.975E-04 | global batch size: 512 | lm loss: 1.841931E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.236 | TFLOPs: 41.76 | 31: iteration 4430/ 52452 | consumed samples: 2268160 | consumed tokens: 4645191680 | elapsed time per iteration (s): 1.74 | learning rate: 1.975E-04 | global batch size: 512 | lm loss: 1.844015E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.437 | TFLOPs: 44.04 | 31: iteration 4440/ 52452 | consumed samples: 2273280 | consumed tokens: 4655677440 | elapsed time per iteration (s): 1.75 | learning rate: 1.975E-04 | global batch size: 512 | lm loss: 1.839391E+00 | grad norm: 0.220 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.507 | TFLOPs: 43.90 | 31: iteration 4450/ 52452 | consumed samples: 2278400 | consumed tokens: 4666163200 | elapsed time per iteration (s): 1.97 | learning rate: 1.975E-04 | global batch size: 512 | lm loss: 1.828897E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 259.887 | TFLOPs: 39.01 | 31: iteration 4460/ 52452 | consumed samples: 2283520 | consumed tokens: 4676648960 | elapsed time per iteration (s): 1.74 | learning rate: 1.975E-04 | global batch size: 512 | lm loss: 1.825209E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.045 | TFLOPs: 44.13 | 31: iteration 4470/ 52452 | consumed samples: 2288640 | consumed tokens: 4687134720 | elapsed time per iteration (s): 1.75 | learning rate: 1.974E-04 | global batch size: 512 | lm loss: 1.833611E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.273 | TFLOPs: 43.87 | 31: iteration 4480/ 52452 | consumed samples: 2293760 | consumed tokens: 4697620480 | elapsed time per iteration (s): 1.77 | learning rate: 1.974E-04 | global batch size: 512 | lm loss: 1.829862E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.881 | TFLOPs: 43.36 | 31: iteration 4490/ 52452 | consumed samples: 2298880 | consumed tokens: 4708106240 | elapsed time per iteration (s): 1.72 | learning rate: 1.974E-04 | global batch size: 512 | lm loss: 1.818175E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.286 | TFLOPs: 44.62 | 31: iteration 4500/ 52452 | consumed samples: 2304000 | consumed tokens: 4718592000 | elapsed time per iteration (s): 1.89 | learning rate: 1.974E-04 | global batch size: 512 | lm loss: 1.839289E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.325 | TFLOPs: 40.57 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4500 | lm loss value: 4.126932E+00 | lm loss PPL: 6.198746E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4500 | lm loss value: 4.212569E+00 | lm loss PPL: 6.752979E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4510/ 52452 | consumed samples: 2309120 | consumed tokens: 4729077760 | elapsed time per iteration (s): 9.43 | learning rate: 1.974E-04 | global batch size: 512 | lm loss: 1.827187E+00 | grad norm: 0.211 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.313 | TFLOPs: 8.15 | 31: iteration 4520/ 52452 | consumed samples: 2314240 | consumed tokens: 4739563520 | elapsed time per iteration (s): 1.80 | learning rate: 1.974E-04 | global batch size: 512 | lm loss: 1.819728E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.256 | TFLOPs: 42.67 | 31: iteration 4530/ 52452 | consumed samples: 2319360 | consumed tokens: 4750049280 | elapsed time per iteration (s): 1.75 | learning rate: 1.974E-04 | global batch size: 512 | lm loss: 1.838281E+00 | grad norm: 0.221 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.857 | TFLOPs: 43.81 | 31: iteration 4540/ 52452 | consumed samples: 2324480 | consumed tokens: 4760535040 | elapsed time per iteration (s): 1.85 | learning rate: 1.974E-04 | global batch size: 512 | lm loss: 1.826769E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.154 | TFLOPs: 41.45 | 31: iteration 4550/ 52452 | consumed samples: 2329600 | consumed tokens: 4771020800 | elapsed time per iteration (s): 1.80 | learning rate: 1.973E-04 | global batch size: 512 | lm loss: 1.827004E+00 | grad norm: 0.247 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.718 | TFLOPs: 42.73 | 31: iteration 4560/ 52452 | consumed samples: 2334720 | consumed tokens: 4781506560 | elapsed time per iteration (s): 1.78 | learning rate: 1.973E-04 | global batch size: 512 | lm loss: 1.825869E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.837 | TFLOPs: 43.05 | 31: iteration 4570/ 52452 | consumed samples: 2339840 | consumed tokens: 4791992320 | elapsed time per iteration (s): 1.90 | learning rate: 1.973E-04 | global batch size: 512 | lm loss: 1.831212E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.440 | TFLOPs: 40.44 | 31: iteration 4580/ 52452 | consumed samples: 2344960 | consumed tokens: 4802478080 | elapsed time per iteration (s): 1.75 | learning rate: 1.973E-04 | global batch size: 512 | lm loss: 1.822290E+00 | grad norm: 0.208 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.236 | TFLOPs: 44.01 | 31: iteration 4590/ 52452 | consumed samples: 2350080 | consumed tokens: 4812963840 | elapsed time per iteration (s): 1.75 | learning rate: 1.973E-04 | global batch size: 512 | lm loss: 1.809896E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.917 | TFLOPs: 43.82 | 31: iteration 4600/ 52452 | consumed samples: 2355200 | consumed tokens: 4823449600 | elapsed time per iteration (s): 1.74 | learning rate: 1.973E-04 | global batch size: 512 | lm loss: 1.819743E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.400 | TFLOPs: 44.19 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4600 | lm loss value: 4.199001E+00 | lm loss PPL: 6.661973E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4600 | lm loss value: 4.293099E+00 | lm loss PPL: 7.319294E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4610/ 52452 | consumed samples: 2360320 | consumed tokens: 4833935360 | elapsed time per iteration (s): 9.22 | learning rate: 1.973E-04 | global batch size: 512 | lm loss: 1.818485E+00 | grad norm: 0.213 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.554 | TFLOPs: 8.34 | 31: iteration 4620/ 52452 | consumed samples: 2365440 | consumed tokens: 4844421120 | elapsed time per iteration (s): 1.75 | learning rate: 1.973E-04 | global batch size: 512 | lm loss: 1.833760E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.622 | TFLOPs: 43.92 | 31: iteration 4630/ 52452 | consumed samples: 2370560 | consumed tokens: 4854906880 | elapsed time per iteration (s): 1.95 | learning rate: 1.972E-04 | global batch size: 512 | lm loss: 1.832458E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 262.651 | TFLOPs: 39.42 | 31: iteration 4640/ 52452 | consumed samples: 2375680 | consumed tokens: 4865392640 | elapsed time per iteration (s): 1.77 | learning rate: 1.972E-04 | global batch size: 512 | lm loss: 1.811479E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.657 | TFLOPs: 43.48 | 31: iteration 4650/ 52452 | consumed samples: 2380800 | consumed tokens: 4875878400 | elapsed time per iteration (s): 1.83 | learning rate: 1.972E-04 | global batch size: 512 | lm loss: 1.816470E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.454 | TFLOPs: 42.09 | 31: iteration 4660/ 52452 | consumed samples: 2385920 | consumed tokens: 4886364160 | elapsed time per iteration (s): 1.80 | learning rate: 1.972E-04 | global batch size: 512 | lm loss: 1.818226E+00 | grad norm: 0.210 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.136 | TFLOPs: 42.80 | 31: iteration 4670/ 52452 | consumed samples: 2391040 | consumed tokens: 4896849920 | elapsed time per iteration (s): 1.88 | learning rate: 1.972E-04 | global batch size: 512 | lm loss: 1.820567E+00 | grad norm: 0.226 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.626 | TFLOPs: 40.92 | 31: iteration 4680/ 52452 | consumed samples: 2396160 | consumed tokens: 4907335680 | elapsed time per iteration (s): 1.76 | learning rate: 1.972E-04 | global batch size: 512 | lm loss: 1.818048E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.161 | TFLOPs: 43.55 | 31: iteration 4690/ 52452 | consumed samples: 2401280 | consumed tokens: 4917821440 | elapsed time per iteration (s): 1.79 | learning rate: 1.972E-04 | global batch size: 512 | lm loss: 1.814508E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.087 | TFLOPs: 42.94 | 31: iteration 4700/ 52452 | consumed samples: 2406400 | consumed tokens: 4928307200 | elapsed time per iteration (s): 1.78 | learning rate: 1.971E-04 | global batch size: 512 | lm loss: 1.811942E+00 | grad norm: 0.219 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.820 | TFLOPs: 43.20 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4700 | lm loss value: 3.987652E+00 | lm loss PPL: 5.392811E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4700 | lm loss value: 4.099243E+00 | lm loss PPL: 6.029464E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4710/ 52452 | consumed samples: 2411520 | consumed tokens: 4938792960 | elapsed time per iteration (s): 9.24 | learning rate: 1.971E-04 | global batch size: 512 | lm loss: 1.832391E+00 | grad norm: 0.224 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.384 | TFLOPs: 8.31 | 31: iteration 4720/ 52452 | consumed samples: 2416640 | consumed tokens: 4949278720 | elapsed time per iteration (s): 1.75 | learning rate: 1.971E-04 | global batch size: 512 | lm loss: 1.804979E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.878 | TFLOPs: 43.81 | 31: iteration 4730/ 52452 | consumed samples: 2421760 | consumed tokens: 4959764480 | elapsed time per iteration (s): 1.83 | learning rate: 1.971E-04 | global batch size: 512 | lm loss: 1.810113E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.263 | TFLOPs: 41.92 | 31: iteration 4740/ 52452 | consumed samples: 2426880 | consumed tokens: 4970250240 | elapsed time per iteration (s): 1.85 | learning rate: 1.971E-04 | global batch size: 512 | lm loss: 1.817958E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.677 | TFLOPs: 41.53 | 31: iteration 4750/ 52452 | consumed samples: 2432000 | consumed tokens: 4980736000 | elapsed time per iteration (s): 1.89 | learning rate: 1.971E-04 | global batch size: 512 | lm loss: 1.814356E+00 | grad norm: 0.230 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.516 | TFLOPs: 40.60 | 31: iteration 4760/ 52452 | consumed samples: 2437120 | consumed tokens: 4991221760 | elapsed time per iteration (s): 1.87 | learning rate: 1.971E-04 | global batch size: 512 | lm loss: 1.806397E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.908 | TFLOPs: 41.11 | 31: iteration 4770/ 52452 | consumed samples: 2442240 | consumed tokens: 5001707520 | elapsed time per iteration (s): 1.84 | learning rate: 1.970E-04 | global batch size: 512 | lm loss: 1.818232E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.681 | TFLOPs: 41.83 | 31: iteration 4780/ 52452 | consumed samples: 2447360 | consumed tokens: 5012193280 | elapsed time per iteration (s): 1.78 | learning rate: 1.970E-04 | global batch size: 512 | lm loss: 1.817692E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.811 | TFLOPs: 43.20 | 31: iteration 4790/ 52452 | consumed samples: 2452480 | consumed tokens: 5022679040 | elapsed time per iteration (s): 1.78 | learning rate: 1.970E-04 | global batch size: 512 | lm loss: 1.821338E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.156 | TFLOPs: 43.10 | 31: iteration 4800/ 52452 | consumed samples: 2457600 | consumed tokens: 5033164800 | elapsed time per iteration (s): 1.77 | learning rate: 1.970E-04 | global batch size: 512 | lm loss: 1.801215E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.368 | TFLOPs: 43.43 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4800 | lm loss value: 4.032534E+00 | lm loss PPL: 5.640364E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4800 | lm loss value: 4.124902E+00 | lm loss PPL: 6.186173E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4810/ 52452 | consumed samples: 2462720 | consumed tokens: 5043650560 | elapsed time per iteration (s): 9.23 | learning rate: 1.970E-04 | global batch size: 512 | lm loss: 1.796061E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.453 | TFLOPs: 8.32 | 31: iteration 4820/ 52452 | consumed samples: 2467840 | consumed tokens: 5054136320 | elapsed time per iteration (s): 1.88 | learning rate: 1.970E-04 | global batch size: 512 | lm loss: 1.817316E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.156 | TFLOPs: 40.85 | 31: iteration 4830/ 52452 | consumed samples: 2472960 | consumed tokens: 5064622080 | elapsed time per iteration (s): 1.78 | learning rate: 1.970E-04 | global batch size: 512 | lm loss: 1.803329E+00 | grad norm: 0.214 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.910 | TFLOPs: 43.21 | 31: iteration 4840/ 52452 | consumed samples: 2478080 | consumed tokens: 5075107840 | elapsed time per iteration (s): 1.80 | learning rate: 1.969E-04 | global batch size: 512 | lm loss: 1.807592E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.416 | TFLOPs: 42.69 | 31: iteration 4850/ 52452 | consumed samples: 2483200 | consumed tokens: 5085593600 | elapsed time per iteration (s): 1.76 | learning rate: 1.969E-04 | global batch size: 512 | lm loss: 1.799028E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.300 | TFLOPs: 43.72 | 31: iteration 4860/ 52452 | consumed samples: 2488320 | consumed tokens: 5096079360 | elapsed time per iteration (s): 1.92 | learning rate: 1.969E-04 | global batch size: 512 | lm loss: 1.816901E+00 | grad norm: 0.242 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.026 | TFLOPs: 39.93 | 31: iteration 4870/ 52452 | consumed samples: 2493440 | consumed tokens: 5106565120 | elapsed time per iteration (s): 1.79 | learning rate: 1.969E-04 | global batch size: 512 | lm loss: 1.798674E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.115 | TFLOPs: 42.94 | 31: iteration 4880/ 52452 | consumed samples: 2498560 | consumed tokens: 5117050880 | elapsed time per iteration (s): 1.83 | learning rate: 1.969E-04 | global batch size: 512 | lm loss: 1.821031E+00 | grad norm: 0.212 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.034 | TFLOPs: 41.88 | 31: iteration 4890/ 52452 | consumed samples: 2503680 | consumed tokens: 5127536640 | elapsed time per iteration (s): 1.76 | learning rate: 1.969E-04 | global batch size: 512 | lm loss: 1.811116E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.101 | TFLOPs: 43.54 | 31: iteration 4900/ 52452 | consumed samples: 2508800 | consumed tokens: 5138022400 | elapsed time per iteration (s): 1.95 | learning rate: 1.969E-04 | global batch size: 512 | lm loss: 1.796342E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 262.517 | TFLOPs: 39.40 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 4900 | lm loss value: 4.233959E+00 | lm loss PPL: 6.898980E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 4900 | lm loss value: 4.295568E+00 | lm loss PPL: 7.337391E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 4910/ 52452 | consumed samples: 2513920 | consumed tokens: 5148508160 | elapsed time per iteration (s): 9.12 | learning rate: 1.969E-04 | global batch size: 512 | lm loss: 1.796881E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.128 | TFLOPs: 8.42 | 31: iteration 4920/ 52452 | consumed samples: 2519040 | consumed tokens: 5158993920 | elapsed time per iteration (s): 1.81 | learning rate: 1.968E-04 | global batch size: 512 | lm loss: 1.783429E+00 | grad norm: 0.212 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.488 | TFLOPs: 42.55 | 31: iteration 4930/ 52452 | consumed samples: 2524160 | consumed tokens: 5169479680 | elapsed time per iteration (s): 1.82 | learning rate: 1.968E-04 | global batch size: 512 | lm loss: 1.789419E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.083 | TFLOPs: 42.19 | 31: iteration 4940/ 52452 | consumed samples: 2529280 | consumed tokens: 5179965440 | elapsed time per iteration (s): 1.82 | learning rate: 1.968E-04 | global batch size: 512 | lm loss: 1.795941E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.904 | TFLOPs: 42.16 | 31: iteration 4950/ 52452 | consumed samples: 2534400 | consumed tokens: 5190451200 | elapsed time per iteration (s): 1.84 | learning rate: 1.968E-04 | global batch size: 512 | lm loss: 1.806590E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.818 | TFLOPs: 41.85 | 31: iteration 4960/ 52452 | consumed samples: 2539520 | consumed tokens: 5200936960 | elapsed time per iteration (s): 1.79 | learning rate: 1.968E-04 | global batch size: 512 | lm loss: 1.799386E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.979 | TFLOPs: 42.92 | 31: iteration 4970/ 52452 | consumed samples: 2544640 | consumed tokens: 5211422720 | elapsed time per iteration (s): 1.84 | learning rate: 1.968E-04 | global batch size: 512 | lm loss: 1.800295E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.566 | TFLOPs: 41.81 | 31: iteration 4980/ 52452 | consumed samples: 2549760 | consumed tokens: 5221908480 | elapsed time per iteration (s): 1.79 | learning rate: 1.968E-04 | global batch size: 512 | lm loss: 1.795156E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.803 | TFLOPs: 42.90 | 31: iteration 4990/ 52452 | consumed samples: 2554880 | consumed tokens: 5232394240 | elapsed time per iteration (s): 1.79 | learning rate: 1.967E-04 | global batch size: 512 | lm loss: 1.781137E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.841 | TFLOPs: 42.90 | 31: iteration 5000/ 52452 | consumed samples: 2560000 | consumed tokens: 5242880000 | elapsed time per iteration (s): 1.80 | learning rate: 1.967E-04 | global batch size: 512 | lm loss: 1.797902E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.855 | TFLOPs: 42.61 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5000 | lm loss value: 4.212153E+00 | lm loss PPL: 6.750174E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5000 | lm loss value: 4.286953E+00 | lm loss PPL: 7.274451E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5010/ 52452 | consumed samples: 2565120 | consumed tokens: 5253365760 | elapsed time per iteration (s): 9.20 | learning rate: 1.967E-04 | global batch size: 512 | lm loss: 1.795633E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.650 | TFLOPs: 8.35 | 31: iteration 5020/ 52452 | consumed samples: 2570240 | consumed tokens: 5263851520 | elapsed time per iteration (s): 1.82 | learning rate: 1.967E-04 | global batch size: 512 | lm loss: 1.787610E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.572 | TFLOPs: 42.26 | 31: iteration 5030/ 52452 | consumed samples: 2575360 | consumed tokens: 5274337280 | elapsed time per iteration (s): 1.81 | learning rate: 1.967E-04 | global batch size: 512 | lm loss: 1.793337E+00 | grad norm: 0.223 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.464 | TFLOPs: 42.55 | 31: iteration 5040/ 52452 | consumed samples: 2580480 | consumed tokens: 5284823040 | elapsed time per iteration (s): 1.82 | learning rate: 1.967E-04 | global batch size: 512 | lm loss: 1.799292E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.855 | TFLOPs: 42.15 | 31: iteration 5050/ 52452 | consumed samples: 2585600 | consumed tokens: 5295308800 | elapsed time per iteration (s): 1.75 | learning rate: 1.966E-04 | global batch size: 512 | lm loss: 1.781099E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.116 | TFLOPs: 44.00 | 31: iteration 5060/ 52452 | consumed samples: 2590720 | consumed tokens: 5305794560 | elapsed time per iteration (s): 1.78 | learning rate: 1.966E-04 | global batch size: 512 | lm loss: 1.779274E+00 | grad norm: 0.216 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.046 | TFLOPs: 43.08 | 31: iteration 5070/ 52452 | consumed samples: 2595840 | consumed tokens: 5316280320 | elapsed time per iteration (s): 1.78 | learning rate: 1.966E-04 | global batch size: 512 | lm loss: 1.783548E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.334 | TFLOPs: 43.13 | 31: iteration 5080/ 52452 | consumed samples: 2600960 | consumed tokens: 5326766080 | elapsed time per iteration (s): 1.79 | learning rate: 1.966E-04 | global batch size: 512 | lm loss: 1.784319E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.654 | TFLOPs: 42.88 | 31: iteration 5090/ 52452 | consumed samples: 2606080 | consumed tokens: 5337251840 | elapsed time per iteration (s): 1.74 | learning rate: 1.966E-04 | global batch size: 512 | lm loss: 1.789858E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.907 | TFLOPs: 44.26 | 31: iteration 5100/ 52452 | consumed samples: 2611200 | consumed tokens: 5347737600 | elapsed time per iteration (s): 1.79 | learning rate: 1.966E-04 | global batch size: 512 | lm loss: 1.777224E+00 | grad norm: 0.208 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.723 | TFLOPs: 42.89 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5100 | lm loss value: 4.089185E+00 | lm loss PPL: 5.969124E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5100 | lm loss value: 4.157495E+00 | lm loss PPL: 6.391120E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5110/ 52452 | consumed samples: 2616320 | consumed tokens: 5358223360 | elapsed time per iteration (s): 9.20 | learning rate: 1.966E-04 | global batch size: 512 | lm loss: 1.783536E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.652 | TFLOPs: 8.35 | 31: iteration 5120/ 52452 | consumed samples: 2621440 | consumed tokens: 5368709120 | elapsed time per iteration (s): 1.78 | learning rate: 1.965E-04 | global batch size: 512 | lm loss: 1.771990E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.375 | TFLOPs: 43.13 | 31: iteration 5130/ 52452 | consumed samples: 2626560 | consumed tokens: 5379194880 | elapsed time per iteration (s): 1.78 | learning rate: 1.965E-04 | global batch size: 512 | lm loss: 1.787763E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.995 | TFLOPs: 43.08 | 31: iteration 5140/ 52452 | consumed samples: 2631680 | consumed tokens: 5389680640 | elapsed time per iteration (s): 1.75 | learning rate: 1.965E-04 | global batch size: 512 | lm loss: 1.777251E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.142 | TFLOPs: 43.85 | 31: iteration 5150/ 52452 | consumed samples: 2636800 | consumed tokens: 5400166400 | elapsed time per iteration (s): 1.80 | learning rate: 1.965E-04 | global batch size: 512 | lm loss: 1.772589E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.841 | TFLOPs: 42.75 | 31: iteration 5160/ 52452 | consumed samples: 2641920 | consumed tokens: 5410652160 | elapsed time per iteration (s): 1.75 | learning rate: 1.965E-04 | global batch size: 512 | lm loss: 1.779869E+00 | grad norm: 0.302 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.280 | TFLOPs: 43.87 | 31: iteration 5170/ 52452 | consumed samples: 2647040 | consumed tokens: 5421137920 | elapsed time per iteration (s): 1.78 | learning rate: 1.965E-04 | global batch size: 512 | lm loss: 1.794336E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.972 | TFLOPs: 43.07 | 31: iteration 5180/ 52452 | consumed samples: 2652160 | consumed tokens: 5431623680 | elapsed time per iteration (s): 1.80 | learning rate: 1.965E-04 | global batch size: 512 | lm loss: 1.795200E+00 | grad norm: 0.357 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.621 | TFLOPs: 42.72 | 31: iteration 5190/ 52452 | consumed samples: 2657280 | consumed tokens: 5442109440 | elapsed time per iteration (s): 1.75 | learning rate: 1.964E-04 | global batch size: 512 | lm loss: 1.802697E+00 | grad norm: 0.310 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.992 | TFLOPs: 43.98 | 31: iteration 5200/ 52452 | consumed samples: 2662400 | consumed tokens: 5452595200 | elapsed time per iteration (s): 1.77 | learning rate: 1.964E-04 | global batch size: 512 | lm loss: 1.783903E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.120 | TFLOPs: 43.40 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5200 | lm loss value: 4.190376E+00 | lm loss PPL: 6.604761E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5200 | lm loss value: 4.254770E+00 | lm loss PPL: 7.044063E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5210/ 52452 | consumed samples: 2667520 | consumed tokens: 5463080960 | elapsed time per iteration (s): 9.14 | learning rate: 1.964E-04 | global batch size: 512 | lm loss: 1.777055E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.006 | TFLOPs: 8.41 | 31: iteration 5220/ 52452 | consumed samples: 2672640 | consumed tokens: 5473566720 | elapsed time per iteration (s): 1.75 | learning rate: 1.964E-04 | global batch size: 512 | lm loss: 1.777582E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.847 | TFLOPs: 43.95 | 31: iteration 5230/ 52452 | consumed samples: 2677760 | consumed tokens: 5484052480 | elapsed time per iteration (s): 1.81 | learning rate: 1.964E-04 | global batch size: 512 | lm loss: 1.788488E+00 | grad norm: 0.208 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.642 | TFLOPs: 42.42 | 31: iteration 5240/ 52452 | consumed samples: 2682880 | consumed tokens: 5494538240 | elapsed time per iteration (s): 1.75 | learning rate: 1.964E-04 | global batch size: 512 | lm loss: 1.763351E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.922 | TFLOPs: 43.97 | 31: iteration 5250/ 52452 | consumed samples: 2688000 | consumed tokens: 5505024000 | elapsed time per iteration (s): 1.76 | learning rate: 1.963E-04 | global batch size: 512 | lm loss: 1.777968E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.138 | TFLOPs: 43.55 | 31: iteration 5260/ 52452 | consumed samples: 2693120 | consumed tokens: 5515509760 | elapsed time per iteration (s): 1.73 | learning rate: 1.963E-04 | global batch size: 512 | lm loss: 1.770185E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.079 | TFLOPs: 44.44 | 31: iteration 5270/ 52452 | consumed samples: 2698240 | consumed tokens: 5525995520 | elapsed time per iteration (s): 1.90 | learning rate: 1.963E-04 | global batch size: 512 | lm loss: 1.774352E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.775 | TFLOPs: 40.49 | 31: iteration 5280/ 52452 | consumed samples: 2703360 | consumed tokens: 5536481280 | elapsed time per iteration (s): 1.78 | learning rate: 1.963E-04 | global batch size: 512 | lm loss: 1.769453E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.162 | TFLOPs: 43.10 | 31: iteration 5290/ 52452 | consumed samples: 2708480 | consumed tokens: 5546967040 | elapsed time per iteration (s): 1.78 | learning rate: 1.963E-04 | global batch size: 512 | lm loss: 1.765129E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.163 | TFLOPs: 43.25 | 31: iteration 5300/ 52452 | consumed samples: 2713600 | consumed tokens: 5557452800 | elapsed time per iteration (s): 1.81 | learning rate: 1.963E-04 | global batch size: 512 | lm loss: 1.771350E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.271 | TFLOPs: 42.37 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5300 | lm loss value: 4.168139E+00 | lm loss PPL: 6.459510E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5300 | lm loss value: 4.217377E+00 | lm loss PPL: 6.785525E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5310/ 52452 | consumed samples: 2718720 | consumed tokens: 5567938560 | elapsed time per iteration (s): 9.25 | learning rate: 1.963E-04 | global batch size: 512 | lm loss: 1.761796E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.359 | TFLOPs: 8.31 | 31: iteration 5320/ 52452 | consumed samples: 2723840 | consumed tokens: 5578424320 | elapsed time per iteration (s): 1.74 | learning rate: 1.962E-04 | global batch size: 512 | lm loss: 1.776941E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.491 | TFLOPs: 44.05 | 31: iteration 5330/ 52452 | consumed samples: 2728960 | consumed tokens: 5588910080 | elapsed time per iteration (s): 1.77 | learning rate: 1.962E-04 | global batch size: 512 | lm loss: 1.749258E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.905 | TFLOPs: 43.51 | 31: iteration 5340/ 52452 | consumed samples: 2734080 | consumed tokens: 5599395840 | elapsed time per iteration (s): 2.47 | learning rate: 1.962E-04 | global batch size: 512 | lm loss: 1.752201E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 207.057 | TFLOPs: 31.08 | 31: iteration 5350/ 52452 | consumed samples: 2739200 | consumed tokens: 5609881600 | elapsed time per iteration (s): 1.81 | learning rate: 1.962E-04 | global batch size: 512 | lm loss: 1.769483E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.795 | TFLOPs: 42.45 | 31: iteration 5360/ 52452 | consumed samples: 2744320 | consumed tokens: 5620367360 | elapsed time per iteration (s): 1.81 | learning rate: 1.962E-04 | global batch size: 512 | lm loss: 1.777452E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.480 | TFLOPs: 42.55 | 31: iteration 5370/ 52452 | consumed samples: 2749440 | consumed tokens: 5630853120 | elapsed time per iteration (s): 1.82 | learning rate: 1.962E-04 | global batch size: 512 | lm loss: 1.771465E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.593 | TFLOPs: 42.12 | 31: iteration 5380/ 52452 | consumed samples: 2754560 | consumed tokens: 5641338880 | elapsed time per iteration (s): 1.77 | learning rate: 1.961E-04 | global batch size: 512 | lm loss: 1.773374E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.069 | TFLOPs: 43.54 | 31: iteration 5390/ 52452 | consumed samples: 2759680 | consumed tokens: 5651824640 | elapsed time per iteration (s): 1.81 | learning rate: 1.961E-04 | global batch size: 512 | lm loss: 1.776755E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.060 | TFLOPs: 42.49 | 31: iteration 5400/ 52452 | consumed samples: 2764800 | consumed tokens: 5662310400 | elapsed time per iteration (s): 1.77 | learning rate: 1.961E-04 | global batch size: 512 | lm loss: 1.773161E+00 | grad norm: 0.238 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.105 | TFLOPs: 43.39 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5400 | lm loss value: 4.086251E+00 | lm loss PPL: 5.951636E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5400 | lm loss value: 4.141406E+00 | lm loss PPL: 6.289116E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5410/ 52452 | consumed samples: 2769920 | consumed tokens: 5672796160 | elapsed time per iteration (s): 9.23 | learning rate: 1.961E-04 | global batch size: 512 | lm loss: 1.752902E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.460 | TFLOPs: 8.32 | 31: iteration 5420/ 52452 | consumed samples: 2775040 | consumed tokens: 5683281920 | elapsed time per iteration (s): 1.74 | learning rate: 1.961E-04 | global batch size: 512 | lm loss: 1.764253E+00 | grad norm: 0.228 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.950 | TFLOPs: 44.27 | 31: iteration 5430/ 52452 | consumed samples: 2780160 | consumed tokens: 5693767680 | elapsed time per iteration (s): 1.79 | learning rate: 1.961E-04 | global batch size: 512 | lm loss: 1.759062E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.756 | TFLOPs: 42.89 | 31: iteration 5440/ 52452 | consumed samples: 2785280 | consumed tokens: 5704253440 | elapsed time per iteration (s): 1.74 | learning rate: 1.960E-04 | global batch size: 512 | lm loss: 1.767471E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.000 | TFLOPs: 44.28 | 31: iteration 5450/ 52452 | consumed samples: 2790400 | consumed tokens: 5714739200 | elapsed time per iteration (s): 1.83 | learning rate: 1.960E-04 | global batch size: 512 | lm loss: 1.771583E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.213 | TFLOPs: 41.91 | 31: iteration 5460/ 52452 | consumed samples: 2795520 | consumed tokens: 5725224960 | elapsed time per iteration (s): 1.78 | learning rate: 1.960E-04 | global batch size: 512 | lm loss: 1.750711E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.042 | TFLOPs: 43.08 | 31: iteration 5470/ 52452 | consumed samples: 2800640 | consumed tokens: 5735710720 | elapsed time per iteration (s): 1.77 | learning rate: 1.960E-04 | global batch size: 512 | lm loss: 1.754534E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.566 | TFLOPs: 43.46 | 31: iteration 5480/ 52452 | consumed samples: 2805760 | consumed tokens: 5746196480 | elapsed time per iteration (s): 1.76 | learning rate: 1.960E-04 | global batch size: 512 | lm loss: 1.768747E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.571 | TFLOPs: 43.61 | 31: iteration 5490/ 52452 | consumed samples: 2810880 | consumed tokens: 5756682240 | elapsed time per iteration (s): 1.84 | learning rate: 1.960E-04 | global batch size: 512 | lm loss: 1.761328E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.359 | TFLOPs: 41.78 | 31: iteration 5500/ 52452 | consumed samples: 2816000 | consumed tokens: 5767168000 | elapsed time per iteration (s): 1.83 | learning rate: 1.960E-04 | global batch size: 512 | lm loss: 1.757927E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.996 | TFLOPs: 42.03 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5500 | lm loss value: 4.087809E+00 | lm loss PPL: 5.960915E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5500 | lm loss value: 4.156377E+00 | lm loss PPL: 6.383983E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5510/ 52452 | consumed samples: 2821120 | consumed tokens: 5777653760 | elapsed time per iteration (s): 9.17 | learning rate: 1.959E-04 | global batch size: 512 | lm loss: 1.750837E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.822 | TFLOPs: 8.38 | 31: iteration 5520/ 52452 | consumed samples: 2826240 | consumed tokens: 5788139520 | elapsed time per iteration (s): 1.79 | learning rate: 1.959E-04 | global batch size: 512 | lm loss: 1.749162E+00 | grad norm: 0.222 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.677 | TFLOPs: 43.03 | 31: iteration 5530/ 52452 | consumed samples: 2831360 | consumed tokens: 5798625280 | elapsed time per iteration (s): 1.78 | learning rate: 1.959E-04 | global batch size: 512 | lm loss: 1.766164E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.031 | TFLOPs: 43.23 | 31: iteration 5540/ 52452 | consumed samples: 2836480 | consumed tokens: 5809111040 | elapsed time per iteration (s): 1.74 | learning rate: 1.959E-04 | global batch size: 512 | lm loss: 1.764718E+00 | grad norm: 0.228 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.691 | TFLOPs: 44.23 | 31: iteration 5550/ 52452 | consumed samples: 2841600 | consumed tokens: 5819596800 | elapsed time per iteration (s): 1.91 | learning rate: 1.959E-04 | global batch size: 512 | lm loss: 1.754669E+00 | grad norm: 0.230 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.421 | TFLOPs: 40.29 | 31: iteration 5560/ 52452 | consumed samples: 2846720 | consumed tokens: 5830082560 | elapsed time per iteration (s): 1.77 | learning rate: 1.959E-04 | global batch size: 512 | lm loss: 1.762126E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.134 | TFLOPs: 43.40 | 31: iteration 5570/ 52452 | consumed samples: 2851840 | consumed tokens: 5840568320 | elapsed time per iteration (s): 1.77 | learning rate: 1.958E-04 | global batch size: 512 | lm loss: 1.749986E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.620 | TFLOPs: 43.47 | 31: iteration 5580/ 52452 | consumed samples: 2856960 | consumed tokens: 5851054080 | elapsed time per iteration (s): 1.78 | learning rate: 1.958E-04 | global batch size: 512 | lm loss: 1.751142E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.494 | TFLOPs: 43.15 | 31: iteration 5590/ 52452 | consumed samples: 2862080 | consumed tokens: 5861539840 | elapsed time per iteration (s): 1.81 | learning rate: 1.958E-04 | global batch size: 512 | lm loss: 1.745444E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.534 | TFLOPs: 42.41 | 31: iteration 5600/ 52452 | consumed samples: 2867200 | consumed tokens: 5872025600 | elapsed time per iteration (s): 1.85 | learning rate: 1.958E-04 | global batch size: 512 | lm loss: 1.752795E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.311 | TFLOPs: 41.62 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5600 | lm loss value: 4.319757E+00 | lm loss PPL: 7.517039E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5600 | lm loss value: 4.364420E+00 | lm loss PPL: 7.860379E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5610/ 52452 | consumed samples: 2872320 | consumed tokens: 5882511360 | elapsed time per iteration (s): 9.16 | learning rate: 1.958E-04 | global batch size: 512 | lm loss: 1.760588E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.876 | TFLOPs: 8.39 | 31: iteration 5620/ 52452 | consumed samples: 2877440 | consumed tokens: 5892997120 | elapsed time per iteration (s): 1.73 | learning rate: 1.958E-04 | global batch size: 512 | lm loss: 1.758885E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.973 | TFLOPs: 44.42 | 31: iteration 5630/ 52452 | consumed samples: 2882560 | consumed tokens: 5903482880 | elapsed time per iteration (s): 1.76 | learning rate: 1.957E-04 | global batch size: 512 | lm loss: 1.752901E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.550 | TFLOPs: 43.76 | 31: iteration 5640/ 52452 | consumed samples: 2887680 | consumed tokens: 5913968640 | elapsed time per iteration (s): 1.76 | learning rate: 1.957E-04 | global batch size: 512 | lm loss: 1.737113E+00 | grad norm: 0.210 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.798 | TFLOPs: 43.65 | 31: iteration 5650/ 52452 | consumed samples: 2892800 | consumed tokens: 5924454400 | elapsed time per iteration (s): 1.80 | learning rate: 1.957E-04 | global batch size: 512 | lm loss: 1.757014E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.765 | TFLOPs: 42.74 | 31: iteration 5660/ 52452 | consumed samples: 2897920 | consumed tokens: 5934940160 | elapsed time per iteration (s): 1.84 | learning rate: 1.957E-04 | global batch size: 512 | lm loss: 1.761183E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.782 | TFLOPs: 41.84 | 31: iteration 5670/ 52452 | consumed samples: 2903040 | consumed tokens: 5945425920 | elapsed time per iteration (s): 1.79 | learning rate: 1.957E-04 | global batch size: 512 | lm loss: 1.741674E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.378 | TFLOPs: 42.98 | 31: iteration 5680/ 52452 | consumed samples: 2908160 | consumed tokens: 5955911680 | elapsed time per iteration (s): 1.96 | learning rate: 1.957E-04 | global batch size: 512 | lm loss: 1.744290E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 261.491 | TFLOPs: 39.25 | 31: iteration 5690/ 52452 | consumed samples: 2913280 | consumed tokens: 5966397440 | elapsed time per iteration (s): 1.92 | learning rate: 1.956E-04 | global batch size: 512 | lm loss: 1.750008E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.409 | TFLOPs: 39.99 | 31: iteration 5700/ 52452 | consumed samples: 2918400 | consumed tokens: 5976883200 | elapsed time per iteration (s): 1.77 | learning rate: 1.956E-04 | global batch size: 512 | lm loss: 1.749175E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.471 | TFLOPs: 43.45 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5700 | lm loss value: 4.254375E+00 | lm loss PPL: 7.041283E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5700 | lm loss value: 4.332050E+00 | lm loss PPL: 7.610016E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5710/ 52452 | consumed samples: 2923520 | consumed tokens: 5987368960 | elapsed time per iteration (s): 9.22 | learning rate: 1.956E-04 | global batch size: 512 | lm loss: 1.740106E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.516 | TFLOPs: 8.33 | 31: iteration 5720/ 52452 | consumed samples: 2928640 | consumed tokens: 5997854720 | elapsed time per iteration (s): 1.79 | learning rate: 1.956E-04 | global batch size: 512 | lm loss: 1.749295E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.509 | TFLOPs: 43.00 | 31: iteration 5730/ 52452 | consumed samples: 2933760 | consumed tokens: 6008340480 | elapsed time per iteration (s): 1.79 | learning rate: 1.956E-04 | global batch size: 512 | lm loss: 1.737527E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.453 | TFLOPs: 43.00 | 31: iteration 5740/ 52452 | consumed samples: 2938880 | consumed tokens: 6018826240 | elapsed time per iteration (s): 1.82 | learning rate: 1.956E-04 | global batch size: 512 | lm loss: 1.742616E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.050 | TFLOPs: 42.33 | 31: iteration 5750/ 52452 | consumed samples: 2944000 | consumed tokens: 6029312000 | elapsed time per iteration (s): 1.75 | learning rate: 1.955E-04 | global batch size: 512 | lm loss: 1.748113E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.705 | TFLOPs: 43.93 | 31: iteration 5760/ 52452 | consumed samples: 2949120 | consumed tokens: 6039797760 | elapsed time per iteration (s): 1.74 | learning rate: 1.955E-04 | global batch size: 512 | lm loss: 1.751062E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.897 | TFLOPs: 44.11 | 31: iteration 5770/ 52452 | consumed samples: 2954240 | consumed tokens: 6050283520 | elapsed time per iteration (s): 1.74 | learning rate: 1.955E-04 | global batch size: 512 | lm loss: 1.739829E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.423 | TFLOPs: 44.04 | 31: iteration 5780/ 52452 | consumed samples: 2959360 | consumed tokens: 6060769280 | elapsed time per iteration (s): 1.98 | learning rate: 1.955E-04 | global batch size: 512 | lm loss: 1.760376E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 258.762 | TFLOPs: 38.84 | 31: iteration 5790/ 52452 | consumed samples: 2964480 | consumed tokens: 6071255040 | elapsed time per iteration (s): 1.86 | learning rate: 1.955E-04 | global batch size: 512 | lm loss: 1.736863E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.679 | TFLOPs: 41.23 | 31: iteration 5800/ 52452 | consumed samples: 2969600 | consumed tokens: 6081740800 | elapsed time per iteration (s): 1.73 | learning rate: 1.955E-04 | global batch size: 512 | lm loss: 1.742528E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.061 | TFLOPs: 44.44 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5800 | lm loss value: 4.087048E+00 | lm loss PPL: 5.956380E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5800 | lm loss value: 4.131250E+00 | lm loss PPL: 6.225572E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5810/ 52452 | consumed samples: 2974720 | consumed tokens: 6092226560 | elapsed time per iteration (s): 9.30 | learning rate: 1.954E-04 | global batch size: 512 | lm loss: 1.734337E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.052 | TFLOPs: 8.26 | 31: iteration 5820/ 52452 | consumed samples: 2979840 | consumed tokens: 6102712320 | elapsed time per iteration (s): 1.84 | learning rate: 1.954E-04 | global batch size: 512 | lm loss: 1.733968E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.104 | TFLOPs: 41.74 | 31: iteration 5830/ 52452 | consumed samples: 2984960 | consumed tokens: 6113198080 | elapsed time per iteration (s): 1.81 | learning rate: 1.954E-04 | global batch size: 512 | lm loss: 1.756927E+00 | grad norm: 0.213 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.013 | TFLOPs: 42.48 | 31: iteration 5840/ 52452 | consumed samples: 2990080 | consumed tokens: 6123683840 | elapsed time per iteration (s): 1.84 | learning rate: 1.954E-04 | global batch size: 512 | lm loss: 1.743705E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.037 | TFLOPs: 41.73 | 31: iteration 5850/ 52452 | consumed samples: 2995200 | consumed tokens: 6134169600 | elapsed time per iteration (s): 1.80 | learning rate: 1.954E-04 | global batch size: 512 | lm loss: 1.718529E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.718 | TFLOPs: 42.73 | 31: iteration 5860/ 52452 | consumed samples: 3000320 | consumed tokens: 6144655360 | elapsed time per iteration (s): 1.80 | learning rate: 1.954E-04 | global batch size: 512 | lm loss: 1.737570E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.004 | TFLOPs: 42.63 | 31: iteration 5870/ 52452 | consumed samples: 3005440 | consumed tokens: 6155141120 | elapsed time per iteration (s): 1.78 | learning rate: 1.953E-04 | global batch size: 512 | lm loss: 1.736402E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.619 | TFLOPs: 43.17 | 31: iteration 5880/ 52452 | consumed samples: 3010560 | consumed tokens: 6165626880 | elapsed time per iteration (s): 1.83 | learning rate: 1.953E-04 | global batch size: 512 | lm loss: 1.739360E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.721 | TFLOPs: 41.98 | 31: iteration 5890/ 52452 | consumed samples: 3015680 | consumed tokens: 6176112640 | elapsed time per iteration (s): 1.82 | learning rate: 1.953E-04 | global batch size: 512 | lm loss: 1.732358E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.592 | TFLOPs: 42.27 | 31: iteration 5900/ 52452 | consumed samples: 3020800 | consumed tokens: 6186598400 | elapsed time per iteration (s): 1.77 | learning rate: 1.953E-04 | global batch size: 512 | lm loss: 1.741893E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.016 | TFLOPs: 43.38 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 5900 | lm loss value: 4.084776E+00 | lm loss PPL: 5.942862E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 5900 | lm loss value: 4.155840E+00 | lm loss PPL: 6.380553E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 5910/ 52452 | consumed samples: 3025920 | consumed tokens: 6197084160 | elapsed time per iteration (s): 9.22 | learning rate: 1.953E-04 | global batch size: 512 | lm loss: 1.741659E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.517 | TFLOPs: 8.33 | 31: iteration 5920/ 52452 | consumed samples: 3031040 | consumed tokens: 6207569920 | elapsed time per iteration (s): 1.79 | learning rate: 1.952E-04 | global batch size: 512 | lm loss: 1.741644E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.577 | TFLOPs: 43.01 | 31: iteration 5930/ 52452 | consumed samples: 3036160 | consumed tokens: 6218055680 | elapsed time per iteration (s): 1.79 | learning rate: 1.952E-04 | global batch size: 512 | lm loss: 1.736273E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.562 | TFLOPs: 42.86 | 31: iteration 5940/ 52452 | consumed samples: 3041280 | consumed tokens: 6228541440 | elapsed time per iteration (s): 1.76 | learning rate: 1.952E-04 | global batch size: 512 | lm loss: 1.751610E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.449 | TFLOPs: 43.74 | 31: iteration 5950/ 52452 | consumed samples: 3046400 | consumed tokens: 6239027200 | elapsed time per iteration (s): 1.93 | learning rate: 1.952E-04 | global batch size: 512 | lm loss: 1.728398E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.821 | TFLOPs: 39.90 | 31: iteration 5960/ 52452 | consumed samples: 3051520 | consumed tokens: 6249512960 | elapsed time per iteration (s): 1.90 | learning rate: 1.952E-04 | global batch size: 512 | lm loss: 1.732138E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.873 | TFLOPs: 40.51 | 31: iteration 5970/ 52452 | consumed samples: 3056640 | consumed tokens: 6259998720 | elapsed time per iteration (s): 1.76 | learning rate: 1.952E-04 | global batch size: 512 | lm loss: 1.727829E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.358 | TFLOPs: 43.73 | 31: iteration 5980/ 52452 | consumed samples: 3061760 | consumed tokens: 6270484480 | elapsed time per iteration (s): 1.80 | learning rate: 1.951E-04 | global batch size: 512 | lm loss: 1.717861E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.995 | TFLOPs: 42.78 | 31: iteration 5990/ 52452 | consumed samples: 3066880 | consumed tokens: 6280970240 | elapsed time per iteration (s): 1.76 | learning rate: 1.951E-04 | global batch size: 512 | lm loss: 1.731476E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.997 | TFLOPs: 43.68 | 0: [2022-12-25 13:15:37,819] [INFO] [logging.py:68:log_dist] [Rank 0] step=6000, skipped=0, lr=[0.000195106885569433, 0.000195106885569433, 0.000195106885569433], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 6000/ 52452 | consumed samples: 3072000 | consumed tokens: 6291456000 | elapsed time per iteration (s): 1.77 | learning rate: 1.951E-04 | global batch size: 512 | lm loss: 1.727363E+00 | grad norm: 0.225 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.013 | TFLOPs: 43.38 | 0: steps: 6000 loss: 1.7247 iter time (s): 1.805 samples/sec: 283.599 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6000 | lm loss value: 4.086633E+00 | lm loss PPL: 5.953907E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6000 | lm loss value: 4.145152E+00 | lm loss PPL: 6.312722E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6010/ 52452 | consumed samples: 3077120 | consumed tokens: 6301941760 | elapsed time per iteration (s): 9.25 | learning rate: 1.951E-04 | global batch size: 512 | lm loss: 1.733750E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.358 | TFLOPs: 8.31 | 31: iteration 6020/ 52452 | consumed samples: 3082240 | consumed tokens: 6312427520 | elapsed time per iteration (s): 1.75 | learning rate: 1.951E-04 | global batch size: 512 | lm loss: 1.723876E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.260 | TFLOPs: 43.87 | 31: iteration 6030/ 52452 | consumed samples: 3087360 | consumed tokens: 6322913280 | elapsed time per iteration (s): 1.82 | learning rate: 1.951E-04 | global batch size: 512 | lm loss: 1.735658E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.236 | TFLOPs: 42.21 | 31: iteration 6040/ 52452 | consumed samples: 3092480 | consumed tokens: 6333399040 | elapsed time per iteration (s): 1.74 | learning rate: 1.950E-04 | global batch size: 512 | lm loss: 1.728403E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.162 | TFLOPs: 44.15 | 31: iteration 6050/ 52452 | consumed samples: 3097600 | consumed tokens: 6343884800 | elapsed time per iteration (s): 1.82 | learning rate: 1.950E-04 | global batch size: 512 | lm loss: 1.735469E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.912 | TFLOPs: 42.16 | 31: iteration 6060/ 52452 | consumed samples: 3102720 | consumed tokens: 6354370560 | elapsed time per iteration (s): 1.82 | learning rate: 1.950E-04 | global batch size: 512 | lm loss: 1.720126E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.698 | TFLOPs: 42.28 | 31: iteration 6070/ 52452 | consumed samples: 3107840 | consumed tokens: 6364856320 | elapsed time per iteration (s): 1.80 | learning rate: 1.950E-04 | global batch size: 512 | lm loss: 1.732123E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.805 | TFLOPs: 42.75 | 31: iteration 6080/ 52452 | consumed samples: 3112960 | consumed tokens: 6375342080 | elapsed time per iteration (s): 1.80 | learning rate: 1.950E-04 | global batch size: 512 | lm loss: 1.716002E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.363 | TFLOPs: 42.68 | 31: iteration 6090/ 52452 | consumed samples: 3118080 | consumed tokens: 6385827840 | elapsed time per iteration (s): 1.86 | learning rate: 1.949E-04 | global batch size: 512 | lm loss: 1.708841E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.698 | TFLOPs: 41.38 | 31: iteration 6100/ 52452 | consumed samples: 3123200 | consumed tokens: 6396313600 | elapsed time per iteration (s): 1.79 | learning rate: 1.949E-04 | global batch size: 512 | lm loss: 1.727572E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.540 | TFLOPs: 42.86 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6100 | lm loss value: 4.142648E+00 | lm loss PPL: 6.296936E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6100 | lm loss value: 4.203969E+00 | lm loss PPL: 6.695150E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6110/ 52452 | consumed samples: 3128320 | consumed tokens: 6406799360 | elapsed time per iteration (s): 9.29 | learning rate: 1.949E-04 | global batch size: 512 | lm loss: 1.727623E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.132 | TFLOPs: 8.28 | 31: iteration 6120/ 52452 | consumed samples: 3133440 | consumed tokens: 6417285120 | elapsed time per iteration (s): 1.78 | learning rate: 1.949E-04 | global batch size: 512 | lm loss: 1.731976E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.929 | TFLOPs: 43.07 | 31: iteration 6130/ 52452 | consumed samples: 3138560 | consumed tokens: 6427770880 | elapsed time per iteration (s): 1.78 | learning rate: 1.949E-04 | global batch size: 512 | lm loss: 1.714852E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.129 | TFLOPs: 43.25 | 31: iteration 6140/ 52452 | consumed samples: 3143680 | consumed tokens: 6438256640 | elapsed time per iteration (s): 1.83 | learning rate: 1.949E-04 | global batch size: 512 | lm loss: 1.722575E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.831 | TFLOPs: 42.00 | 31: iteration 6150/ 52452 | consumed samples: 3148800 | consumed tokens: 6448742400 | elapsed time per iteration (s): 1.81 | learning rate: 1.948E-04 | global batch size: 512 | lm loss: 1.725401E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.537 | TFLOPs: 42.56 | 31: iteration 6160/ 52452 | consumed samples: 3153920 | consumed tokens: 6459228160 | elapsed time per iteration (s): 1.79 | learning rate: 1.948E-04 | global batch size: 512 | lm loss: 1.726546E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.801 | TFLOPs: 43.05 | 31: iteration 6170/ 52452 | consumed samples: 3159040 | consumed tokens: 6469713920 | elapsed time per iteration (s): 1.85 | learning rate: 1.948E-04 | global batch size: 512 | lm loss: 1.711362E+00 | grad norm: 0.239 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.813 | TFLOPs: 41.55 | 31: iteration 6180/ 52452 | consumed samples: 3164160 | consumed tokens: 6480199680 | elapsed time per iteration (s): 1.74 | learning rate: 1.948E-04 | global batch size: 512 | lm loss: 1.728615E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.603 | TFLOPs: 44.07 | 31: iteration 6190/ 52452 | consumed samples: 3169280 | consumed tokens: 6490685440 | elapsed time per iteration (s): 1.75 | learning rate: 1.948E-04 | global batch size: 512 | lm loss: 1.712034E+00 | grad norm: 0.212 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.353 | TFLOPs: 43.88 | 31: iteration 6200/ 52452 | consumed samples: 3174400 | consumed tokens: 6501171200 | elapsed time per iteration (s): 1.85 | learning rate: 1.947E-04 | global batch size: 512 | lm loss: 1.713877E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.464 | TFLOPs: 41.65 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6200 | lm loss value: 4.438554E+00 | lm loss PPL: 8.465247E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6200 | lm loss value: 4.486610E+00 | lm loss PPL: 8.881987E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6210/ 52452 | consumed samples: 3179520 | consumed tokens: 6511656960 | elapsed time per iteration (s): 9.34 | learning rate: 1.947E-04 | global batch size: 512 | lm loss: 1.714922E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.820 | TFLOPs: 8.23 | 31: iteration 6220/ 52452 | consumed samples: 3184640 | consumed tokens: 6522142720 | elapsed time per iteration (s): 1.78 | learning rate: 1.947E-04 | global batch size: 512 | lm loss: 1.705479E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.852 | TFLOPs: 43.05 | 31: iteration 6230/ 52452 | consumed samples: 3189760 | consumed tokens: 6532628480 | elapsed time per iteration (s): 1.79 | learning rate: 1.947E-04 | global batch size: 512 | lm loss: 1.720108E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.585 | TFLOPs: 43.01 | 31: iteration 6240/ 52452 | consumed samples: 3194880 | consumed tokens: 6543114240 | elapsed time per iteration (s): 1.79 | learning rate: 1.947E-04 | global batch size: 512 | lm loss: 1.718890E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.915 | TFLOPs: 42.91 | 31: iteration 6250/ 52452 | consumed samples: 3200000 | consumed tokens: 6553600000 | elapsed time per iteration (s): 1.83 | learning rate: 1.947E-04 | global batch size: 512 | lm loss: 1.725613E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.329 | TFLOPs: 41.93 | 31: iteration 6260/ 52452 | consumed samples: 3205120 | consumed tokens: 6564085760 | elapsed time per iteration (s): 1.80 | learning rate: 1.946E-04 | global batch size: 512 | lm loss: 1.719813E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.920 | TFLOPs: 42.61 | 31: iteration 6270/ 52452 | consumed samples: 3210240 | consumed tokens: 6574571520 | elapsed time per iteration (s): 1.80 | learning rate: 1.946E-04 | global batch size: 512 | lm loss: 1.714752E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.349 | TFLOPs: 42.68 | 31: iteration 6280/ 52452 | consumed samples: 3215360 | consumed tokens: 6585057280 | elapsed time per iteration (s): 1.74 | learning rate: 1.946E-04 | global batch size: 512 | lm loss: 1.713790E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.509 | TFLOPs: 44.05 | 31: iteration 6290/ 52452 | consumed samples: 3220480 | consumed tokens: 6595543040 | elapsed time per iteration (s): 1.81 | learning rate: 1.946E-04 | global batch size: 512 | lm loss: 1.713500E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.920 | TFLOPs: 42.46 | 31: iteration 6300/ 52452 | consumed samples: 3225600 | consumed tokens: 6606028800 | elapsed time per iteration (s): 1.77 | learning rate: 1.946E-04 | global batch size: 512 | lm loss: 1.697563E+00 | grad norm: 0.223 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.878 | TFLOPs: 43.36 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6300 | lm loss value: 4.296205E+00 | lm loss PPL: 7.342060E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6300 | lm loss value: 4.324998E+00 | lm loss PPL: 7.556535E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6310/ 52452 | consumed samples: 3230720 | consumed tokens: 6616514560 | elapsed time per iteration (s): 9.15 | learning rate: 1.945E-04 | global batch size: 512 | lm loss: 1.721288E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.946 | TFLOPs: 8.40 | 31: iteration 6320/ 52452 | consumed samples: 3235840 | consumed tokens: 6627000320 | elapsed time per iteration (s): 1.74 | learning rate: 1.945E-04 | global batch size: 512 | lm loss: 1.711460E+00 | grad norm: 0.213 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.461 | TFLOPs: 44.05 | 31: iteration 6330/ 52452 | consumed samples: 3240960 | consumed tokens: 6637486080 | elapsed time per iteration (s): 1.78 | learning rate: 1.945E-04 | global batch size: 512 | lm loss: 1.702496E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.950 | TFLOPs: 43.22 | 31: iteration 6340/ 52452 | consumed samples: 3246080 | consumed tokens: 6647971840 | elapsed time per iteration (s): 1.79 | learning rate: 1.945E-04 | global batch size: 512 | lm loss: 1.708143E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.912 | TFLOPs: 42.91 | 31: iteration 6350/ 52452 | consumed samples: 3251200 | consumed tokens: 6658457600 | elapsed time per iteration (s): 1.84 | learning rate: 1.945E-04 | global batch size: 512 | lm loss: 1.701448E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.158 | TFLOPs: 41.75 | 31: iteration 6360/ 52452 | consumed samples: 3256320 | consumed tokens: 6668943360 | elapsed time per iteration (s): 1.84 | learning rate: 1.944E-04 | global batch size: 512 | lm loss: 1.703732E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.846 | TFLOPs: 41.85 | 31: iteration 6370/ 52452 | consumed samples: 3261440 | consumed tokens: 6679429120 | elapsed time per iteration (s): 1.74 | learning rate: 1.944E-04 | global batch size: 512 | lm loss: 1.709851E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.657 | TFLOPs: 44.23 | 31: iteration 6380/ 52452 | consumed samples: 3266560 | consumed tokens: 6689914880 | elapsed time per iteration (s): 1.78 | learning rate: 1.944E-04 | global batch size: 512 | lm loss: 1.715620E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.897 | TFLOPs: 43.06 | 31: iteration 6390/ 52452 | consumed samples: 3271680 | consumed tokens: 6700400640 | elapsed time per iteration (s): 1.82 | learning rate: 1.944E-04 | global batch size: 512 | lm loss: 1.709280E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.899 | TFLOPs: 42.16 | 31: iteration 6400/ 52452 | consumed samples: 3276800 | consumed tokens: 6710886400 | elapsed time per iteration (s): 1.78 | learning rate: 1.944E-04 | global batch size: 512 | lm loss: 1.716645E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.827 | TFLOPs: 43.20 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6400 | lm loss value: 4.069637E+00 | lm loss PPL: 5.853573E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6400 | lm loss value: 4.149037E+00 | lm loss PPL: 6.337297E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6410/ 52452 | consumed samples: 3281920 | consumed tokens: 6721372160 | elapsed time per iteration (s): 9.19 | learning rate: 1.944E-04 | global batch size: 512 | lm loss: 1.693120E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.691 | TFLOPs: 8.36 | 31: iteration 6420/ 52452 | consumed samples: 3287040 | consumed tokens: 6731857920 | elapsed time per iteration (s): 1.80 | learning rate: 1.943E-04 | global batch size: 512 | lm loss: 1.706153E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.230 | TFLOPs: 42.81 | 31: iteration 6430/ 52452 | consumed samples: 3292160 | consumed tokens: 6742343680 | elapsed time per iteration (s): 1.74 | learning rate: 1.943E-04 | global batch size: 512 | lm loss: 1.708321E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.615 | TFLOPs: 44.07 | 31: iteration 6440/ 52452 | consumed samples: 3297280 | consumed tokens: 6752829440 | elapsed time per iteration (s): 1.75 | learning rate: 1.943E-04 | global batch size: 512 | lm loss: 1.698071E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.281 | TFLOPs: 43.87 | 31: iteration 6450/ 52452 | consumed samples: 3302400 | consumed tokens: 6763315200 | elapsed time per iteration (s): 1.74 | learning rate: 1.943E-04 | global batch size: 512 | lm loss: 1.709063E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.008 | TFLOPs: 44.28 | 31: iteration 6460/ 52452 | consumed samples: 3307520 | consumed tokens: 6773800960 | elapsed time per iteration (s): 1.86 | learning rate: 1.943E-04 | global batch size: 512 | lm loss: 1.703451E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.802 | TFLOPs: 41.40 | 31: iteration 6470/ 52452 | consumed samples: 3312640 | consumed tokens: 6784286720 | elapsed time per iteration (s): 1.74 | learning rate: 1.942E-04 | global batch size: 512 | lm loss: 1.703416E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.469 | TFLOPs: 44.05 | 31: iteration 6480/ 52452 | consumed samples: 3317760 | consumed tokens: 6794772480 | elapsed time per iteration (s): 1.83 | learning rate: 1.942E-04 | global batch size: 512 | lm loss: 1.714088E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.763 | TFLOPs: 41.99 | 31: iteration 6490/ 52452 | consumed samples: 3322880 | consumed tokens: 6805258240 | elapsed time per iteration (s): 1.81 | learning rate: 1.942E-04 | global batch size: 512 | lm loss: 1.716636E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.149 | TFLOPs: 42.35 | 31: iteration 6500/ 52452 | consumed samples: 3328000 | consumed tokens: 6815744000 | elapsed time per iteration (s): 1.83 | learning rate: 1.942E-04 | global batch size: 512 | lm loss: 1.709841E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.038 | TFLOPs: 41.88 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6500 | lm loss value: 4.091559E+00 | lm loss PPL: 5.983309E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6500 | lm loss value: 4.120061E+00 | lm loss PPL: 6.156302E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6510/ 52452 | consumed samples: 3333120 | consumed tokens: 6826229760 | elapsed time per iteration (s): 9.20 | learning rate: 1.942E-04 | global batch size: 512 | lm loss: 1.698198E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.669 | TFLOPs: 8.36 | 31: iteration 6520/ 52452 | consumed samples: 3338240 | consumed tokens: 6836715520 | elapsed time per iteration (s): 1.75 | learning rate: 1.941E-04 | global batch size: 512 | lm loss: 1.692694E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.132 | TFLOPs: 43.85 | 31: iteration 6530/ 52452 | consumed samples: 3343360 | consumed tokens: 6847201280 | elapsed time per iteration (s): 1.77 | learning rate: 1.941E-04 | global batch size: 512 | lm loss: 1.698029E+00 | grad norm: 0.212 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.926 | TFLOPs: 43.52 | 31: iteration 6540/ 52452 | consumed samples: 3348480 | consumed tokens: 6857687040 | elapsed time per iteration (s): 1.75 | learning rate: 1.941E-04 | global batch size: 512 | lm loss: 1.698558E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.750 | TFLOPs: 43.79 | 31: iteration 6550/ 52452 | consumed samples: 3353600 | consumed tokens: 6868172800 | elapsed time per iteration (s): 1.77 | learning rate: 1.941E-04 | global batch size: 512 | lm loss: 1.710655E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.309 | TFLOPs: 43.42 | 31: iteration 6560/ 52452 | consumed samples: 3358720 | consumed tokens: 6878658560 | elapsed time per iteration (s): 1.75 | learning rate: 1.941E-04 | global batch size: 512 | lm loss: 1.699417E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.167 | TFLOPs: 44.00 | 31: iteration 6570/ 52452 | consumed samples: 3363840 | consumed tokens: 6889144320 | elapsed time per iteration (s): 1.76 | learning rate: 1.940E-04 | global batch size: 512 | lm loss: 1.704461E+00 | grad norm: 0.275 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.286 | TFLOPs: 43.72 | 31: iteration 6580/ 52452 | consumed samples: 3368960 | consumed tokens: 6899630080 | elapsed time per iteration (s): 1.74 | learning rate: 1.940E-04 | global batch size: 512 | lm loss: 1.710196E+00 | grad norm: 0.223 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.964 | TFLOPs: 44.12 | 31: iteration 6590/ 52452 | consumed samples: 3374080 | consumed tokens: 6910115840 | elapsed time per iteration (s): 1.78 | learning rate: 1.940E-04 | global batch size: 512 | lm loss: 1.714545E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.158 | TFLOPs: 43.10 | 31: iteration 6600/ 52452 | consumed samples: 3379200 | consumed tokens: 6920601600 | elapsed time per iteration (s): 1.79 | learning rate: 1.940E-04 | global batch size: 512 | lm loss: 1.699647E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.426 | TFLOPs: 42.99 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6600 | lm loss value: 4.181418E+00 | lm loss PPL: 6.545860E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6600 | lm loss value: 4.207065E+00 | lm loss PPL: 6.715911E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6610/ 52452 | consumed samples: 3384320 | consumed tokens: 6931087360 | elapsed time per iteration (s): 9.17 | learning rate: 1.940E-04 | global batch size: 512 | lm loss: 1.701470E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.864 | TFLOPs: 8.38 | 31: iteration 6620/ 52452 | consumed samples: 3389440 | consumed tokens: 6941573120 | elapsed time per iteration (s): 1.77 | learning rate: 1.939E-04 | global batch size: 512 | lm loss: 1.690097E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.064 | TFLOPs: 43.54 | 31: iteration 6630/ 52452 | consumed samples: 3394560 | consumed tokens: 6952058880 | elapsed time per iteration (s): 1.75 | learning rate: 1.939E-04 | global batch size: 512 | lm loss: 1.697278E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.260 | TFLOPs: 44.02 | 31: iteration 6640/ 52452 | consumed samples: 3399680 | consumed tokens: 6962544640 | elapsed time per iteration (s): 1.74 | learning rate: 1.939E-04 | global batch size: 512 | lm loss: 1.690398E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.759 | TFLOPs: 44.09 | 31: iteration 6650/ 52452 | consumed samples: 3404800 | consumed tokens: 6973030400 | elapsed time per iteration (s): 1.78 | learning rate: 1.939E-04 | global batch size: 512 | lm loss: 1.689339E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.042 | TFLOPs: 43.23 | 31: iteration 6660/ 52452 | consumed samples: 3409920 | consumed tokens: 6983516160 | elapsed time per iteration (s): 1.77 | learning rate: 1.939E-04 | global batch size: 512 | lm loss: 1.697933E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.566 | TFLOPs: 43.46 | 31: iteration 6670/ 52452 | consumed samples: 3415040 | consumed tokens: 6994001920 | elapsed time per iteration (s): 1.76 | learning rate: 1.939E-04 | global batch size: 512 | lm loss: 1.686075E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.448 | TFLOPs: 43.74 | 31: iteration 6680/ 52452 | consumed samples: 3420160 | consumed tokens: 7004487680 | elapsed time per iteration (s): 1.79 | learning rate: 1.938E-04 | global batch size: 512 | lm loss: 1.692984E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.544 | TFLOPs: 43.01 | 31: iteration 6690/ 52452 | consumed samples: 3425280 | consumed tokens: 7014973440 | elapsed time per iteration (s): 1.80 | learning rate: 1.938E-04 | global batch size: 512 | lm loss: 1.703739E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.414 | TFLOPs: 42.69 | 31: iteration 6700/ 52452 | consumed samples: 3430400 | consumed tokens: 7025459200 | elapsed time per iteration (s): 1.78 | learning rate: 1.938E-04 | global batch size: 512 | lm loss: 1.697896E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.270 | TFLOPs: 43.27 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6700 | lm loss value: 4.130878E+00 | lm loss PPL: 6.223254E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6700 | lm loss value: 4.163995E+00 | lm loss PPL: 6.432802E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6710/ 52452 | consumed samples: 3435520 | consumed tokens: 7035944960 | elapsed time per iteration (s): 9.24 | learning rate: 1.938E-04 | global batch size: 512 | lm loss: 1.691840E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.383 | TFLOPs: 8.31 | 31: iteration 6720/ 52452 | consumed samples: 3440640 | consumed tokens: 7046430720 | elapsed time per iteration (s): 1.78 | learning rate: 1.938E-04 | global batch size: 512 | lm loss: 1.690234E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.273 | TFLOPs: 43.27 | 31: iteration 6730/ 52452 | consumed samples: 3445760 | consumed tokens: 7056916480 | elapsed time per iteration (s): 1.76 | learning rate: 1.937E-04 | global batch size: 512 | lm loss: 1.691611E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.397 | TFLOPs: 43.74 | 31: iteration 6740/ 52452 | consumed samples: 3450880 | consumed tokens: 7067402240 | elapsed time per iteration (s): 1.80 | learning rate: 1.937E-04 | global batch size: 512 | lm loss: 1.680877E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.740 | TFLOPs: 42.74 | 31: iteration 6750/ 52452 | consumed samples: 3456000 | consumed tokens: 7077888000 | elapsed time per iteration (s): 1.84 | learning rate: 1.937E-04 | global batch size: 512 | lm loss: 1.702428E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.048 | TFLOPs: 41.73 | 31: iteration 6760/ 52452 | consumed samples: 3461120 | consumed tokens: 7088373760 | elapsed time per iteration (s): 1.83 | learning rate: 1.937E-04 | global batch size: 512 | lm loss: 1.692501E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.389 | TFLOPs: 42.08 | 31: iteration 6770/ 52452 | consumed samples: 3466240 | consumed tokens: 7098859520 | elapsed time per iteration (s): 1.82 | learning rate: 1.937E-04 | global batch size: 512 | lm loss: 1.673825E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.983 | TFLOPs: 42.17 | 31: iteration 6780/ 52452 | consumed samples: 3471360 | consumed tokens: 7109345280 | elapsed time per iteration (s): 1.77 | learning rate: 1.936E-04 | global batch size: 512 | lm loss: 1.686149E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.959 | TFLOPs: 43.52 | 31: iteration 6790/ 52452 | consumed samples: 3476480 | consumed tokens: 7119831040 | elapsed time per iteration (s): 1.77 | learning rate: 1.936E-04 | global batch size: 512 | lm loss: 1.700142E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.845 | TFLOPs: 43.35 | 31: iteration 6800/ 52452 | consumed samples: 3481600 | consumed tokens: 7130316800 | elapsed time per iteration (s): 1.76 | learning rate: 1.936E-04 | global batch size: 512 | lm loss: 1.681212E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.812 | TFLOPs: 43.65 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6800 | lm loss value: 4.259569E+00 | lm loss PPL: 7.077948E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6800 | lm loss value: 4.289814E+00 | lm loss PPL: 7.295286E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6810/ 52452 | consumed samples: 3486720 | consumed tokens: 7140802560 | elapsed time per iteration (s): 9.17 | learning rate: 1.936E-04 | global batch size: 512 | lm loss: 1.678055E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.828 | TFLOPs: 8.38 | 31: iteration 6820/ 52452 | consumed samples: 3491840 | consumed tokens: 7151288320 | elapsed time per iteration (s): 1.75 | learning rate: 1.936E-04 | global batch size: 512 | lm loss: 1.692618E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.910 | TFLOPs: 43.81 | 31: iteration 6830/ 52452 | consumed samples: 3496960 | consumed tokens: 7161774080 | elapsed time per iteration (s): 1.77 | learning rate: 1.935E-04 | global batch size: 512 | lm loss: 1.692060E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.203 | TFLOPs: 43.41 | 31: iteration 6840/ 52452 | consumed samples: 3502080 | consumed tokens: 7172259840 | elapsed time per iteration (s): 1.76 | learning rate: 1.935E-04 | global batch size: 512 | lm loss: 1.686487E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.575 | TFLOPs: 43.76 | 31: iteration 6850/ 52452 | consumed samples: 3507200 | consumed tokens: 7182745600 | elapsed time per iteration (s): 1.78 | learning rate: 1.935E-04 | global batch size: 512 | lm loss: 1.694878E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.347 | TFLOPs: 43.13 | 31: iteration 6860/ 52452 | consumed samples: 3512320 | consumed tokens: 7193231360 | elapsed time per iteration (s): 1.79 | learning rate: 1.935E-04 | global batch size: 512 | lm loss: 1.693233E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.376 | TFLOPs: 42.98 | 31: iteration 6870/ 52452 | consumed samples: 3517440 | consumed tokens: 7203717120 | elapsed time per iteration (s): 1.83 | learning rate: 1.934E-04 | global batch size: 512 | lm loss: 1.678095E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.774 | TFLOPs: 41.99 | 31: iteration 6880/ 52452 | consumed samples: 3522560 | consumed tokens: 7214202880 | elapsed time per iteration (s): 1.90 | learning rate: 1.934E-04 | global batch size: 512 | lm loss: 1.672890E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.705 | TFLOPs: 40.48 | 31: iteration 6890/ 52452 | consumed samples: 3527680 | consumed tokens: 7224688640 | elapsed time per iteration (s): 1.79 | learning rate: 1.934E-04 | global batch size: 512 | lm loss: 1.696881E+00 | grad norm: 0.220 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.148 | TFLOPs: 42.95 | 31: iteration 6900/ 52452 | consumed samples: 3532800 | consumed tokens: 7235174400 | elapsed time per iteration (s): 1.79 | learning rate: 1.934E-04 | global batch size: 512 | lm loss: 1.666672E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.409 | TFLOPs: 42.84 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 6900 | lm loss value: 4.081850E+00 | lm loss PPL: 5.925499E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 6900 | lm loss value: 4.108860E+00 | lm loss PPL: 6.087731E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 6910/ 52452 | consumed samples: 3537920 | consumed tokens: 7245660160 | elapsed time per iteration (s): 9.12 | learning rate: 1.934E-04 | global batch size: 512 | lm loss: 1.678489E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.116 | TFLOPs: 8.42 | 31: iteration 6920/ 52452 | consumed samples: 3543040 | consumed tokens: 7256145920 | elapsed time per iteration (s): 1.76 | learning rate: 1.933E-04 | global batch size: 512 | lm loss: 1.692985E+00 | grad norm: 0.212 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.676 | TFLOPs: 43.78 | 31: iteration 6930/ 52452 | consumed samples: 3548160 | consumed tokens: 7266631680 | elapsed time per iteration (s): 1.87 | learning rate: 1.933E-04 | global batch size: 512 | lm loss: 1.670456E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.260 | TFLOPs: 41.16 | 31: iteration 6940/ 52452 | consumed samples: 3553280 | consumed tokens: 7277117440 | elapsed time per iteration (s): 1.79 | learning rate: 1.933E-04 | global batch size: 512 | lm loss: 1.688150E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.905 | TFLOPs: 42.91 | 31: iteration 6950/ 52452 | consumed samples: 3558400 | consumed tokens: 7287603200 | elapsed time per iteration (s): 1.80 | learning rate: 1.933E-04 | global batch size: 512 | lm loss: 1.683516E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.836 | TFLOPs: 42.75 | 31: iteration 6960/ 52452 | consumed samples: 3563520 | consumed tokens: 7298088960 | elapsed time per iteration (s): 1.85 | learning rate: 1.933E-04 | global batch size: 512 | lm loss: 1.682314E+00 | grad norm: 0.230 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.223 | TFLOPs: 41.46 | 31: iteration 6970/ 52452 | consumed samples: 3568640 | consumed tokens: 7308574720 | elapsed time per iteration (s): 1.80 | learning rate: 1.932E-04 | global batch size: 512 | lm loss: 1.676823E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.989 | TFLOPs: 42.63 | 31: iteration 6980/ 52452 | consumed samples: 3573760 | consumed tokens: 7319060480 | elapsed time per iteration (s): 1.75 | learning rate: 1.932E-04 | global batch size: 512 | lm loss: 1.677736E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.219 | TFLOPs: 44.01 | 31: iteration 6990/ 52452 | consumed samples: 3578880 | consumed tokens: 7329546240 | elapsed time per iteration (s): 1.75 | learning rate: 1.932E-04 | global batch size: 512 | lm loss: 1.693020E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.254 | TFLOPs: 44.02 | 31: iteration 7000/ 52452 | consumed samples: 3584000 | consumed tokens: 7340032000 | elapsed time per iteration (s): 1.80 | learning rate: 1.932E-04 | global batch size: 512 | lm loss: 1.675546E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.863 | TFLOPs: 42.76 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7000 | lm loss value: 4.093357E+00 | lm loss PPL: 5.994075E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7000 | lm loss value: 4.135385E+00 | lm loss PPL: 6.251363E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7010/ 52452 | consumed samples: 3589120 | consumed tokens: 7350517760 | elapsed time per iteration (s): 9.18 | learning rate: 1.932E-04 | global batch size: 512 | lm loss: 1.677515E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.775 | TFLOPs: 8.37 | 31: iteration 7020/ 52452 | consumed samples: 3594240 | consumed tokens: 7361003520 | elapsed time per iteration (s): 1.79 | learning rate: 1.931E-04 | global batch size: 512 | lm loss: 1.674245E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.596 | TFLOPs: 43.02 | 31: iteration 7030/ 52452 | consumed samples: 3599360 | consumed tokens: 7371489280 | elapsed time per iteration (s): 1.76 | learning rate: 1.931E-04 | global batch size: 512 | lm loss: 1.675100E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.210 | TFLOPs: 43.71 | 31: iteration 7040/ 52452 | consumed samples: 3604480 | consumed tokens: 7381975040 | elapsed time per iteration (s): 1.75 | learning rate: 1.931E-04 | global batch size: 512 | lm loss: 1.669754E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.409 | TFLOPs: 43.89 | 31: iteration 7050/ 52452 | consumed samples: 3609600 | consumed tokens: 7392460800 | elapsed time per iteration (s): 1.73 | learning rate: 1.931E-04 | global batch size: 512 | lm loss: 1.688449E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.550 | TFLOPs: 44.36 | 31: iteration 7060/ 52452 | consumed samples: 3614720 | consumed tokens: 7402946560 | elapsed time per iteration (s): 1.83 | learning rate: 1.931E-04 | global batch size: 512 | lm loss: 1.672569E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.386 | TFLOPs: 42.08 | 31: iteration 7070/ 52452 | consumed samples: 3619840 | consumed tokens: 7413432320 | elapsed time per iteration (s): 1.75 | learning rate: 1.930E-04 | global batch size: 512 | lm loss: 1.671175E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.197 | TFLOPs: 44.01 | 31: iteration 7080/ 52452 | consumed samples: 3624960 | consumed tokens: 7423918080 | elapsed time per iteration (s): 1.78 | learning rate: 1.930E-04 | global batch size: 512 | lm loss: 1.662998E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.148 | TFLOPs: 43.10 | 31: iteration 7090/ 52452 | consumed samples: 3630080 | consumed tokens: 7434403840 | elapsed time per iteration (s): 1.80 | learning rate: 1.930E-04 | global batch size: 512 | lm loss: 1.671479E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.427 | TFLOPs: 42.69 | 31: iteration 7100/ 52452 | consumed samples: 3635200 | consumed tokens: 7444889600 | elapsed time per iteration (s): 1.77 | learning rate: 1.930E-04 | global batch size: 512 | lm loss: 1.676223E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.310 | TFLOPs: 43.42 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7100 | lm loss value: 4.216200E+00 | lm loss PPL: 6.777544E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7100 | lm loss value: 4.262491E+00 | lm loss PPL: 7.098661E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7110/ 52452 | consumed samples: 3640320 | consumed tokens: 7455375360 | elapsed time per iteration (s): 9.12 | learning rate: 1.930E-04 | global batch size: 512 | lm loss: 1.679552E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.110 | TFLOPs: 8.42 | 31: iteration 7120/ 52452 | consumed samples: 3645440 | consumed tokens: 7465861120 | elapsed time per iteration (s): 1.75 | learning rate: 1.929E-04 | global batch size: 512 | lm loss: 1.679638E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.108 | TFLOPs: 43.84 | 31: iteration 7130/ 52452 | consumed samples: 3650560 | consumed tokens: 7476346880 | elapsed time per iteration (s): 1.77 | learning rate: 1.929E-04 | global batch size: 512 | lm loss: 1.691660E+00 | grad norm: 0.247 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.019 | TFLOPs: 43.53 | 31: iteration 7140/ 52452 | consumed samples: 3655680 | consumed tokens: 7486832640 | elapsed time per iteration (s): 1.75 | learning rate: 1.929E-04 | global batch size: 512 | lm loss: 1.663680E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.279 | TFLOPs: 44.02 | 31: iteration 7150/ 52452 | consumed samples: 3660800 | consumed tokens: 7497318400 | elapsed time per iteration (s): 1.75 | learning rate: 1.929E-04 | global batch size: 512 | lm loss: 1.663391E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.747 | TFLOPs: 43.94 | 31: iteration 7160/ 52452 | consumed samples: 3665920 | consumed tokens: 7507804160 | elapsed time per iteration (s): 1.81 | learning rate: 1.928E-04 | global batch size: 512 | lm loss: 1.675633E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.839 | TFLOPs: 42.45 | 31: iteration 7170/ 52452 | consumed samples: 3671040 | consumed tokens: 7518289920 | elapsed time per iteration (s): 1.77 | learning rate: 1.928E-04 | global batch size: 512 | lm loss: 1.677675E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.909 | TFLOPs: 43.36 | 31: iteration 7180/ 52452 | consumed samples: 3676160 | consumed tokens: 7528775680 | elapsed time per iteration (s): 1.77 | learning rate: 1.928E-04 | global batch size: 512 | lm loss: 1.677652E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.329 | TFLOPs: 43.43 | 31: iteration 7190/ 52452 | consumed samples: 3681280 | consumed tokens: 7539261440 | elapsed time per iteration (s): 1.76 | learning rate: 1.928E-04 | global batch size: 512 | lm loss: 1.667995E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.711 | TFLOPs: 43.63 | 31: iteration 7200/ 52452 | consumed samples: 3686400 | consumed tokens: 7549747200 | elapsed time per iteration (s): 1.79 | learning rate: 1.928E-04 | global batch size: 512 | lm loss: 1.654589E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.299 | TFLOPs: 42.82 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7200 | lm loss value: 4.104564E+00 | lm loss PPL: 6.061629E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7200 | lm loss value: 4.143370E+00 | lm loss PPL: 6.301483E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7210/ 52452 | consumed samples: 3691520 | consumed tokens: 7560232960 | elapsed time per iteration (s): 9.15 | learning rate: 1.927E-04 | global batch size: 512 | lm loss: 1.666253E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.929 | TFLOPs: 8.39 | 31: iteration 7220/ 52452 | consumed samples: 3696640 | consumed tokens: 7570718720 | elapsed time per iteration (s): 1.75 | learning rate: 1.927E-04 | global batch size: 512 | lm loss: 1.664756E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.204 | TFLOPs: 44.01 | 31: iteration 7230/ 52452 | consumed samples: 3701760 | consumed tokens: 7581204480 | elapsed time per iteration (s): 1.86 | learning rate: 1.927E-04 | global batch size: 512 | lm loss: 1.650040E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.526 | TFLOPs: 41.36 | 31: iteration 7240/ 52452 | consumed samples: 3706880 | consumed tokens: 7591690240 | elapsed time per iteration (s): 1.80 | learning rate: 1.927E-04 | global batch size: 512 | lm loss: 1.676620E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.651 | TFLOPs: 42.72 | 31: iteration 7250/ 52452 | consumed samples: 3712000 | consumed tokens: 7602176000 | elapsed time per iteration (s): 1.78 | learning rate: 1.927E-04 | global batch size: 512 | lm loss: 1.666855E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.905 | TFLOPs: 43.21 | 31: iteration 7260/ 52452 | consumed samples: 3717120 | consumed tokens: 7612661760 | elapsed time per iteration (s): 1.77 | learning rate: 1.926E-04 | global batch size: 512 | lm loss: 1.668147E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.664 | TFLOPs: 43.48 | 31: iteration 7270/ 52452 | consumed samples: 3722240 | consumed tokens: 7623147520 | elapsed time per iteration (s): 1.76 | learning rate: 1.926E-04 | global batch size: 512 | lm loss: 1.664265E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.740 | TFLOPs: 43.64 | 31: iteration 7280/ 52452 | consumed samples: 3727360 | consumed tokens: 7633633280 | elapsed time per iteration (s): 1.80 | learning rate: 1.926E-04 | global batch size: 512 | lm loss: 1.652024E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.863 | TFLOPs: 42.61 | 31: iteration 7290/ 52452 | consumed samples: 3732480 | consumed tokens: 7644119040 | elapsed time per iteration (s): 1.74 | learning rate: 1.926E-04 | global batch size: 512 | lm loss: 1.662718E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.410 | TFLOPs: 44.19 | 31: iteration 7300/ 52452 | consumed samples: 3737600 | consumed tokens: 7654604800 | elapsed time per iteration (s): 1.75 | learning rate: 1.925E-04 | global batch size: 512 | lm loss: 1.662701E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.008 | TFLOPs: 43.83 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7300 | lm loss value: 4.094306E+00 | lm loss PPL: 5.999766E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7300 | lm loss value: 4.125338E+00 | lm loss PPL: 6.188873E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7310/ 52452 | consumed samples: 3742720 | consumed tokens: 7665090560 | elapsed time per iteration (s): 9.25 | learning rate: 1.925E-04 | global batch size: 512 | lm loss: 1.664372E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.326 | TFLOPs: 8.30 | 31: iteration 7320/ 52452 | consumed samples: 3747840 | consumed tokens: 7675576320 | elapsed time per iteration (s): 1.73 | learning rate: 1.925E-04 | global batch size: 512 | lm loss: 1.675221E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.207 | TFLOPs: 44.46 | 31: iteration 7330/ 52452 | consumed samples: 3752960 | consumed tokens: 7686062080 | elapsed time per iteration (s): 1.74 | learning rate: 1.925E-04 | global batch size: 512 | lm loss: 1.658455E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.667 | TFLOPs: 44.08 | 31: iteration 7340/ 52452 | consumed samples: 3758080 | consumed tokens: 7696547840 | elapsed time per iteration (s): 1.74 | learning rate: 1.925E-04 | global batch size: 512 | lm loss: 1.651130E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.584 | TFLOPs: 44.22 | 31: iteration 7350/ 52452 | consumed samples: 3763200 | consumed tokens: 7707033600 | elapsed time per iteration (s): 1.80 | learning rate: 1.924E-04 | global batch size: 512 | lm loss: 1.654750E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.031 | TFLOPs: 42.63 | 31: iteration 7360/ 52452 | consumed samples: 3768320 | consumed tokens: 7717519360 | elapsed time per iteration (s): 1.79 | learning rate: 1.924E-04 | global batch size: 512 | lm loss: 1.673178E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.254 | TFLOPs: 42.97 | 31: iteration 7370/ 52452 | consumed samples: 3773440 | consumed tokens: 7728005120 | elapsed time per iteration (s): 1.75 | learning rate: 1.924E-04 | global batch size: 512 | lm loss: 1.663799E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.085 | TFLOPs: 43.84 | 31: iteration 7380/ 52452 | consumed samples: 3778560 | consumed tokens: 7738490880 | elapsed time per iteration (s): 1.74 | learning rate: 1.924E-04 | global batch size: 512 | lm loss: 1.664367E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.280 | TFLOPs: 44.17 | 31: iteration 7390/ 52452 | consumed samples: 3783680 | consumed tokens: 7748976640 | elapsed time per iteration (s): 1.73 | learning rate: 1.923E-04 | global batch size: 512 | lm loss: 1.676177E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.846 | TFLOPs: 44.40 | 31: iteration 7400/ 52452 | consumed samples: 3788800 | consumed tokens: 7759462400 | elapsed time per iteration (s): 1.73 | learning rate: 1.923E-04 | global batch size: 512 | lm loss: 1.657583E+00 | grad norm: 0.212 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.277 | TFLOPs: 44.47 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7400 | lm loss value: 4.208998E+00 | lm loss PPL: 6.728910E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7400 | lm loss value: 4.218347E+00 | lm loss PPL: 6.792109E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7410/ 52452 | consumed samples: 3793920 | consumed tokens: 7769948160 | elapsed time per iteration (s): 9.18 | learning rate: 1.923E-04 | global batch size: 512 | lm loss: 1.658568E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.795 | TFLOPs: 8.37 | 31: iteration 7420/ 52452 | consumed samples: 3799040 | consumed tokens: 7780433920 | elapsed time per iteration (s): 1.72 | learning rate: 1.923E-04 | global batch size: 512 | lm loss: 1.652930E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.166 | TFLOPs: 44.60 | 31: iteration 7430/ 52452 | consumed samples: 3804160 | consumed tokens: 7790919680 | elapsed time per iteration (s): 1.75 | learning rate: 1.923E-04 | global batch size: 512 | lm loss: 1.658025E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.250 | TFLOPs: 43.87 | 31: iteration 7440/ 52452 | consumed samples: 3809280 | consumed tokens: 7801405440 | elapsed time per iteration (s): 1.79 | learning rate: 1.922E-04 | global batch size: 512 | lm loss: 1.679154E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.945 | TFLOPs: 42.92 | 31: iteration 7450/ 52452 | consumed samples: 3814400 | consumed tokens: 7811891200 | elapsed time per iteration (s): 1.77 | learning rate: 1.922E-04 | global batch size: 512 | lm loss: 1.663658E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.989 | TFLOPs: 43.53 | 31: iteration 7460/ 52452 | consumed samples: 3819520 | consumed tokens: 7822376960 | elapsed time per iteration (s): 1.79 | learning rate: 1.922E-04 | global batch size: 512 | lm loss: 1.650078E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.764 | TFLOPs: 43.04 | 31: iteration 7470/ 52452 | consumed samples: 3824640 | consumed tokens: 7832862720 | elapsed time per iteration (s): 1.79 | learning rate: 1.922E-04 | global batch size: 512 | lm loss: 1.654316E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.334 | TFLOPs: 42.83 | 31: iteration 7480/ 52452 | consumed samples: 3829760 | consumed tokens: 7843348480 | elapsed time per iteration (s): 1.76 | learning rate: 1.921E-04 | global batch size: 512 | lm loss: 1.654691E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.414 | TFLOPs: 43.59 | 31: iteration 7490/ 52452 | consumed samples: 3834880 | consumed tokens: 7853834240 | elapsed time per iteration (s): 1.74 | learning rate: 1.921E-04 | global batch size: 512 | lm loss: 1.653184E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.613 | TFLOPs: 44.07 | 31: iteration 7500/ 52452 | consumed samples: 3840000 | consumed tokens: 7864320000 | elapsed time per iteration (s): 1.87 | learning rate: 1.921E-04 | global batch size: 512 | lm loss: 1.656249E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.403 | TFLOPs: 41.04 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7500 | lm loss value: 4.239764E+00 | lm loss PPL: 6.939149E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7500 | lm loss value: 4.228307E+00 | lm loss PPL: 6.860101E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7510/ 52452 | consumed samples: 3845120 | consumed tokens: 7874805760 | elapsed time per iteration (s): 9.24 | learning rate: 1.921E-04 | global batch size: 512 | lm loss: 1.653563E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.410 | TFLOPs: 8.32 | 31: iteration 7520/ 52452 | consumed samples: 3850240 | consumed tokens: 7885291520 | elapsed time per iteration (s): 1.81 | learning rate: 1.921E-04 | global batch size: 512 | lm loss: 1.660740E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.565 | TFLOPs: 42.56 | 31: iteration 7530/ 52452 | consumed samples: 3855360 | consumed tokens: 7895777280 | elapsed time per iteration (s): 1.85 | learning rate: 1.920E-04 | global batch size: 512 | lm loss: 1.655659E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.077 | TFLOPs: 41.59 | 31: iteration 7540/ 52452 | consumed samples: 3860480 | consumed tokens: 7906263040 | elapsed time per iteration (s): 1.79 | learning rate: 1.920E-04 | global batch size: 512 | lm loss: 1.657999E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.756 | TFLOPs: 43.04 | 31: iteration 7550/ 52452 | consumed samples: 3865600 | consumed tokens: 7916748800 | elapsed time per iteration (s): 1.83 | learning rate: 1.920E-04 | global batch size: 512 | lm loss: 1.664825E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.313 | TFLOPs: 41.92 | 31: iteration 7560/ 52452 | consumed samples: 3870720 | consumed tokens: 7927234560 | elapsed time per iteration (s): 1.79 | learning rate: 1.920E-04 | global batch size: 512 | lm loss: 1.669394E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.328 | TFLOPs: 42.98 | 31: iteration 7570/ 52452 | consumed samples: 3875840 | consumed tokens: 7937720320 | elapsed time per iteration (s): 1.83 | learning rate: 1.919E-04 | global batch size: 512 | lm loss: 1.660150E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.245 | TFLOPs: 42.06 | 31: iteration 7580/ 52452 | consumed samples: 3880960 | consumed tokens: 7948206080 | elapsed time per iteration (s): 1.79 | learning rate: 1.919E-04 | global batch size: 512 | lm loss: 1.646326E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.551 | TFLOPs: 42.86 | 31: iteration 7590/ 52452 | consumed samples: 3886080 | consumed tokens: 7958691840 | elapsed time per iteration (s): 1.88 | learning rate: 1.919E-04 | global batch size: 512 | lm loss: 1.647635E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.217 | TFLOPs: 40.86 | 31: iteration 7600/ 52452 | consumed samples: 3891200 | consumed tokens: 7969177600 | elapsed time per iteration (s): 1.75 | learning rate: 1.919E-04 | global batch size: 512 | lm loss: 1.648435E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.219 | TFLOPs: 43.86 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7600 | lm loss value: 4.045122E+00 | lm loss PPL: 5.711816E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7600 | lm loss value: 4.099087E+00 | lm loss PPL: 6.028521E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7610/ 52452 | consumed samples: 3896320 | consumed tokens: 7979663360 | elapsed time per iteration (s): 9.23 | learning rate: 1.919E-04 | global batch size: 512 | lm loss: 1.655228E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.496 | TFLOPs: 8.33 | 31: iteration 7620/ 52452 | consumed samples: 3901440 | consumed tokens: 7990149120 | elapsed time per iteration (s): 1.81 | learning rate: 1.918E-04 | global batch size: 512 | lm loss: 1.654367E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.915 | TFLOPs: 42.46 | 31: iteration 7630/ 52452 | consumed samples: 3906560 | consumed tokens: 8000634880 | elapsed time per iteration (s): 1.83 | learning rate: 1.918E-04 | global batch size: 512 | lm loss: 1.646465E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.526 | TFLOPs: 41.96 | 31: iteration 7640/ 52452 | consumed samples: 3911680 | consumed tokens: 8011120640 | elapsed time per iteration (s): 1.80 | learning rate: 1.918E-04 | global batch size: 512 | lm loss: 1.641393E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.587 | TFLOPs: 42.71 | 31: iteration 7650/ 52452 | consumed samples: 3916800 | consumed tokens: 8021606400 | elapsed time per iteration (s): 1.80 | learning rate: 1.918E-04 | global batch size: 512 | lm loss: 1.659057E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.064 | TFLOPs: 42.79 | 31: iteration 7660/ 52452 | consumed samples: 3921920 | consumed tokens: 8032092160 | elapsed time per iteration (s): 1.77 | learning rate: 1.917E-04 | global batch size: 512 | lm loss: 1.648947E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.302 | TFLOPs: 43.42 | 31: iteration 7670/ 52452 | consumed samples: 3927040 | consumed tokens: 8042577920 | elapsed time per iteration (s): 1.79 | learning rate: 1.917E-04 | global batch size: 512 | lm loss: 1.654130E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.414 | TFLOPs: 42.99 | 31: iteration 7680/ 52452 | consumed samples: 3932160 | consumed tokens: 8053063680 | elapsed time per iteration (s): 1.85 | learning rate: 1.917E-04 | global batch size: 512 | lm loss: 1.646745E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.176 | TFLOPs: 41.60 | 31: iteration 7690/ 52452 | consumed samples: 3937280 | consumed tokens: 8063549440 | elapsed time per iteration (s): 1.73 | learning rate: 1.917E-04 | global batch size: 512 | lm loss: 1.653488E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.226 | TFLOPs: 44.31 | 31: iteration 7700/ 52452 | consumed samples: 3942400 | consumed tokens: 8074035200 | elapsed time per iteration (s): 1.81 | learning rate: 1.917E-04 | global batch size: 512 | lm loss: 1.649833E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.372 | TFLOPs: 42.53 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7700 | lm loss value: 4.252024E+00 | lm loss PPL: 7.024746E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7700 | lm loss value: 4.291328E+00 | lm loss PPL: 7.306346E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7710/ 52452 | consumed samples: 3947520 | consumed tokens: 8084520960 | elapsed time per iteration (s): 9.23 | learning rate: 1.916E-04 | global batch size: 512 | lm loss: 1.642135E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.460 | TFLOPs: 8.32 | 31: iteration 7720/ 52452 | consumed samples: 3952640 | consumed tokens: 8095006720 | elapsed time per iteration (s): 1.78 | learning rate: 1.916E-04 | global batch size: 512 | lm loss: 1.651525E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.987 | TFLOPs: 43.23 | 31: iteration 7730/ 52452 | consumed samples: 3957760 | consumed tokens: 8105492480 | elapsed time per iteration (s): 1.75 | learning rate: 1.916E-04 | global batch size: 512 | lm loss: 1.644395E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.816 | TFLOPs: 43.80 | 31: iteration 7740/ 52452 | consumed samples: 3962880 | consumed tokens: 8115978240 | elapsed time per iteration (s): 1.77 | learning rate: 1.916E-04 | global batch size: 512 | lm loss: 1.660359E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.499 | TFLOPs: 43.45 | 31: iteration 7750/ 52452 | consumed samples: 3968000 | consumed tokens: 8126464000 | elapsed time per iteration (s): 1.80 | learning rate: 1.915E-04 | global batch size: 512 | lm loss: 1.648820E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.072 | TFLOPs: 42.79 | 31: iteration 7760/ 52452 | consumed samples: 3973120 | consumed tokens: 8136949760 | elapsed time per iteration (s): 1.74 | learning rate: 1.915E-04 | global batch size: 512 | lm loss: 1.641648E+00 | grad norm: 0.211 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.040 | TFLOPs: 44.13 | 31: iteration 7770/ 52452 | consumed samples: 3978240 | consumed tokens: 8147435520 | elapsed time per iteration (s): 1.75 | learning rate: 1.915E-04 | global batch size: 512 | lm loss: 1.650168E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.132 | TFLOPs: 44.00 | 31: iteration 7780/ 52452 | consumed samples: 3983360 | consumed tokens: 8157921280 | elapsed time per iteration (s): 1.79 | learning rate: 1.915E-04 | global batch size: 512 | lm loss: 1.637251E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.707 | TFLOPs: 42.88 | 31: iteration 7790/ 52452 | consumed samples: 3988480 | consumed tokens: 8168407040 | elapsed time per iteration (s): 1.79 | learning rate: 1.914E-04 | global batch size: 512 | lm loss: 1.642734E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.023 | TFLOPs: 42.93 | 31: iteration 7800/ 52452 | consumed samples: 3993600 | consumed tokens: 8178892800 | elapsed time per iteration (s): 1.76 | learning rate: 1.914E-04 | global batch size: 512 | lm loss: 1.655322E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.043 | TFLOPs: 43.68 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7800 | lm loss value: 4.100435E+00 | lm loss PPL: 6.036656E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7800 | lm loss value: 4.118963E+00 | lm loss PPL: 6.149542E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7810/ 52452 | consumed samples: 3998720 | consumed tokens: 8189378560 | elapsed time per iteration (s): 9.17 | learning rate: 1.914E-04 | global batch size: 512 | lm loss: 1.636534E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.808 | TFLOPs: 8.38 | 31: iteration 7820/ 52452 | consumed samples: 4003840 | consumed tokens: 8199864320 | elapsed time per iteration (s): 1.79 | learning rate: 1.914E-04 | global batch size: 512 | lm loss: 1.636956E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.823 | TFLOPs: 43.05 | 31: iteration 7830/ 52452 | consumed samples: 4008960 | consumed tokens: 8210350080 | elapsed time per iteration (s): 1.78 | learning rate: 1.914E-04 | global batch size: 512 | lm loss: 1.654199E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.359 | TFLOPs: 43.13 | 31: iteration 7840/ 52452 | consumed samples: 4014080 | consumed tokens: 8220835840 | elapsed time per iteration (s): 1.83 | learning rate: 1.913E-04 | global batch size: 512 | lm loss: 1.647637E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.402 | TFLOPs: 42.09 | 31: iteration 7850/ 52452 | consumed samples: 4019200 | consumed tokens: 8231321600 | elapsed time per iteration (s): 1.74 | learning rate: 1.913E-04 | global batch size: 512 | lm loss: 1.634762E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.294 | TFLOPs: 44.17 | 31: iteration 7860/ 52452 | consumed samples: 4024320 | consumed tokens: 8241807360 | elapsed time per iteration (s): 1.77 | learning rate: 1.913E-04 | global batch size: 512 | lm loss: 1.638903E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.576 | TFLOPs: 43.31 | 31: iteration 7870/ 52452 | consumed samples: 4029440 | consumed tokens: 8252293120 | elapsed time per iteration (s): 1.79 | learning rate: 1.913E-04 | global batch size: 512 | lm loss: 1.640257E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.035 | TFLOPs: 42.93 | 31: iteration 7880/ 52452 | consumed samples: 4034560 | consumed tokens: 8262778880 | elapsed time per iteration (s): 1.86 | learning rate: 1.912E-04 | global batch size: 512 | lm loss: 1.653254E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.764 | TFLOPs: 41.39 | 31: iteration 7890/ 52452 | consumed samples: 4039680 | consumed tokens: 8273264640 | elapsed time per iteration (s): 1.72 | learning rate: 1.912E-04 | global batch size: 512 | lm loss: 1.639653E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.476 | TFLOPs: 44.65 | 31: iteration 7900/ 52452 | consumed samples: 4044800 | consumed tokens: 8283750400 | elapsed time per iteration (s): 1.74 | learning rate: 1.912E-04 | global batch size: 512 | lm loss: 1.636415E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.721 | TFLOPs: 44.09 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 7900 | lm loss value: 4.265071E+00 | lm loss PPL: 7.116997E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 7900 | lm loss value: 4.262154E+00 | lm loss PPL: 7.096265E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 7910/ 52452 | consumed samples: 4049920 | consumed tokens: 8294236160 | elapsed time per iteration (s): 9.19 | learning rate: 1.912E-04 | global batch size: 512 | lm loss: 1.645745E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.726 | TFLOPs: 8.36 | 31: iteration 7920/ 52452 | consumed samples: 4055040 | consumed tokens: 8304721920 | elapsed time per iteration (s): 1.82 | learning rate: 1.911E-04 | global batch size: 512 | lm loss: 1.647821E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.381 | TFLOPs: 42.23 | 31: iteration 7930/ 52452 | consumed samples: 4060160 | consumed tokens: 8315207680 | elapsed time per iteration (s): 1.83 | learning rate: 1.911E-04 | global batch size: 512 | lm loss: 1.647274E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.853 | TFLOPs: 42.00 | 31: iteration 7940/ 52452 | consumed samples: 4065280 | consumed tokens: 8325693440 | elapsed time per iteration (s): 1.74 | learning rate: 1.911E-04 | global batch size: 512 | lm loss: 1.649469E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.791 | TFLOPs: 44.10 | 31: iteration 7950/ 52452 | consumed samples: 4070400 | consumed tokens: 8336179200 | elapsed time per iteration (s): 1.82 | learning rate: 1.911E-04 | global batch size: 512 | lm loss: 1.638795E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.071 | TFLOPs: 42.19 | 31: iteration 7960/ 52452 | consumed samples: 4075520 | consumed tokens: 8346664960 | elapsed time per iteration (s): 1.84 | learning rate: 1.910E-04 | global batch size: 512 | lm loss: 1.635453E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.754 | TFLOPs: 41.84 | 31: iteration 7970/ 52452 | consumed samples: 4080640 | consumed tokens: 8357150720 | elapsed time per iteration (s): 1.81 | learning rate: 1.910E-04 | global batch size: 512 | lm loss: 1.634435E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.168 | TFLOPs: 42.35 | 31: iteration 7980/ 52452 | consumed samples: 4085760 | consumed tokens: 8367636480 | elapsed time per iteration (s): 1.81 | learning rate: 1.910E-04 | global batch size: 512 | lm loss: 1.629629E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.540 | TFLOPs: 42.56 | 31: iteration 7990/ 52452 | consumed samples: 4090880 | consumed tokens: 8378122240 | elapsed time per iteration (s): 1.79 | learning rate: 1.910E-04 | global batch size: 512 | lm loss: 1.647565E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.277 | TFLOPs: 42.82 | 0: [2022-12-25 14:39:50,738] [INFO] [logging.py:68:log_dist] [Rank 0] step=8000, skipped=0, lr=[0.0001909514488443798, 0.0001909514488443798, 0.0001909514488443798], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 8000/ 52452 | consumed samples: 4096000 | consumed tokens: 8388608000 | elapsed time per iteration (s): 1.73 | learning rate: 1.910E-04 | global batch size: 512 | lm loss: 1.630145E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.633 | TFLOPs: 44.37 | 0: steps: 8000 loss: 1.6771 iter time (s): 1.786 samples/sec: 286.708 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8000 | lm loss value: 4.101540E+00 | lm loss PPL: 6.043326E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8000 | lm loss value: 4.113337E+00 | lm loss PPL: 6.115041E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8010/ 52452 | consumed samples: 4101120 | consumed tokens: 8399093760 | elapsed time per iteration (s): 9.19 | learning rate: 1.909E-04 | global batch size: 512 | lm loss: 1.630816E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.720 | TFLOPs: 8.36 | 31: iteration 8020/ 52452 | consumed samples: 4106240 | consumed tokens: 8409579520 | elapsed time per iteration (s): 1.83 | learning rate: 1.909E-04 | global batch size: 512 | lm loss: 1.636725E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.254 | TFLOPs: 42.06 | 31: iteration 8030/ 52452 | consumed samples: 4111360 | consumed tokens: 8420065280 | elapsed time per iteration (s): 1.80 | learning rate: 1.909E-04 | global batch size: 512 | lm loss: 1.628544E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.994 | TFLOPs: 42.78 | 31: iteration 8040/ 52452 | consumed samples: 4116480 | consumed tokens: 8430551040 | elapsed time per iteration (s): 1.76 | learning rate: 1.909E-04 | global batch size: 512 | lm loss: 1.638343E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.097 | TFLOPs: 43.54 | 31: iteration 8050/ 52452 | consumed samples: 4121600 | consumed tokens: 8441036800 | elapsed time per iteration (s): 1.77 | learning rate: 1.908E-04 | global batch size: 512 | lm loss: 1.623750E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.200 | TFLOPs: 43.41 | 31: iteration 8060/ 52452 | consumed samples: 4126720 | consumed tokens: 8451522560 | elapsed time per iteration (s): 1.76 | learning rate: 1.908E-04 | global batch size: 512 | lm loss: 1.639362E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.628 | TFLOPs: 43.62 | 31: iteration 8070/ 52452 | consumed samples: 4131840 | consumed tokens: 8462008320 | elapsed time per iteration (s): 1.76 | learning rate: 1.908E-04 | global batch size: 512 | lm loss: 1.630131E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.500 | TFLOPs: 43.60 | 31: iteration 8080/ 52452 | consumed samples: 4136960 | consumed tokens: 8472494080 | elapsed time per iteration (s): 1.84 | learning rate: 1.908E-04 | global batch size: 512 | lm loss: 1.622945E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.106 | TFLOPs: 41.74 | 31: iteration 8090/ 52452 | consumed samples: 4142080 | consumed tokens: 8482979840 | elapsed time per iteration (s): 1.79 | learning rate: 1.907E-04 | global batch size: 512 | lm loss: 1.633612E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.274 | TFLOPs: 42.82 | 31: iteration 8100/ 52452 | consumed samples: 4147200 | consumed tokens: 8493465600 | elapsed time per iteration (s): 1.77 | learning rate: 1.907E-04 | global batch size: 512 | lm loss: 1.634469E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.391 | TFLOPs: 43.44 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8100 | lm loss value: 4.031866E+00 | lm loss PPL: 5.636597E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8100 | lm loss value: 4.072258E+00 | lm loss PPL: 5.868933E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8110/ 52452 | consumed samples: 4152320 | consumed tokens: 8503951360 | elapsed time per iteration (s): 9.31 | learning rate: 1.907E-04 | global batch size: 512 | lm loss: 1.631428E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.985 | TFLOPs: 8.25 | 31: iteration 8120/ 52452 | consumed samples: 4157440 | consumed tokens: 8514437120 | elapsed time per iteration (s): 1.84 | learning rate: 1.907E-04 | global batch size: 512 | lm loss: 1.617247E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.468 | TFLOPs: 41.80 | 31: iteration 8130/ 52452 | consumed samples: 4162560 | consumed tokens: 8524922880 | elapsed time per iteration (s): 1.78 | learning rate: 1.906E-04 | global batch size: 512 | lm loss: 1.641891E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.499 | TFLOPs: 43.15 | 31: iteration 8140/ 52452 | consumed samples: 4167680 | consumed tokens: 8535408640 | elapsed time per iteration (s): 1.76 | learning rate: 1.906E-04 | global batch size: 512 | lm loss: 1.621877E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.214 | TFLOPs: 43.71 | 31: iteration 8150/ 52452 | consumed samples: 4172800 | consumed tokens: 8545894400 | elapsed time per iteration (s): 1.75 | learning rate: 1.906E-04 | global batch size: 512 | lm loss: 1.630243E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.395 | TFLOPs: 44.04 | 31: iteration 8160/ 52452 | consumed samples: 4177920 | consumed tokens: 8556380160 | elapsed time per iteration (s): 1.75 | learning rate: 1.906E-04 | global batch size: 512 | lm loss: 1.634845E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.072 | TFLOPs: 43.84 | 31: iteration 8170/ 52452 | consumed samples: 4183040 | consumed tokens: 8566865920 | elapsed time per iteration (s): 1.80 | learning rate: 1.905E-04 | global batch size: 512 | lm loss: 1.631623E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.192 | TFLOPs: 42.66 | 31: iteration 8180/ 52452 | consumed samples: 4188160 | consumed tokens: 8577351680 | elapsed time per iteration (s): 1.80 | learning rate: 1.905E-04 | global batch size: 512 | lm loss: 1.613588E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.116 | TFLOPs: 42.79 | 31: iteration 8190/ 52452 | consumed samples: 4193280 | consumed tokens: 8587837440 | elapsed time per iteration (s): 1.79 | learning rate: 1.905E-04 | global batch size: 512 | lm loss: 1.644658E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.950 | TFLOPs: 42.92 | 31: iteration 8200/ 52452 | consumed samples: 4198400 | consumed tokens: 8598323200 | elapsed time per iteration (s): 1.77 | learning rate: 1.905E-04 | global batch size: 512 | lm loss: 1.637288E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.954 | TFLOPs: 43.37 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8200 | lm loss value: 4.164558E+00 | lm loss PPL: 6.436422E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8200 | lm loss value: 4.165333E+00 | lm loss PPL: 6.441412E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8210/ 52452 | consumed samples: 4203520 | consumed tokens: 8608808960 | elapsed time per iteration (s): 9.19 | learning rate: 1.904E-04 | global batch size: 512 | lm loss: 1.627486E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.703 | TFLOPs: 8.36 | 31: iteration 8220/ 52452 | consumed samples: 4208640 | consumed tokens: 8619294720 | elapsed time per iteration (s): 1.75 | learning rate: 1.904E-04 | global batch size: 512 | lm loss: 1.629922E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.507 | TFLOPs: 43.90 | 31: iteration 8230/ 52452 | consumed samples: 4213760 | consumed tokens: 8629780480 | elapsed time per iteration (s): 1.74 | learning rate: 1.904E-04 | global batch size: 512 | lm loss: 1.628862E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.002 | TFLOPs: 44.13 | 31: iteration 8240/ 52452 | consumed samples: 4218880 | consumed tokens: 8640266240 | elapsed time per iteration (s): 1.78 | learning rate: 1.904E-04 | global batch size: 512 | lm loss: 1.609497E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.299 | TFLOPs: 43.27 | 31: iteration 8250/ 52452 | consumed samples: 4224000 | consumed tokens: 8650752000 | elapsed time per iteration (s): 1.77 | learning rate: 1.903E-04 | global batch size: 512 | lm loss: 1.618245E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.241 | TFLOPs: 43.41 | 31: iteration 8260/ 52452 | consumed samples: 4229120 | consumed tokens: 8661237760 | elapsed time per iteration (s): 1.83 | learning rate: 1.903E-04 | global batch size: 512 | lm loss: 1.612348E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.402 | TFLOPs: 41.94 | 31: iteration 8270/ 52452 | consumed samples: 4234240 | consumed tokens: 8671723520 | elapsed time per iteration (s): 1.77 | learning rate: 1.903E-04 | global batch size: 512 | lm loss: 1.619145E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.649 | TFLOPs: 43.47 | 31: iteration 8280/ 52452 | consumed samples: 4239360 | consumed tokens: 8682209280 | elapsed time per iteration (s): 1.76 | learning rate: 1.903E-04 | global batch size: 512 | lm loss: 1.615876E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.715 | TFLOPs: 43.78 | 31: iteration 8290/ 52452 | consumed samples: 4244480 | consumed tokens: 8692695040 | elapsed time per iteration (s): 1.83 | learning rate: 1.902E-04 | global batch size: 512 | lm loss: 1.640060E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.503 | TFLOPs: 41.95 | 31: iteration 8300/ 52452 | consumed samples: 4249600 | consumed tokens: 8703180800 | elapsed time per iteration (s): 1.78 | learning rate: 1.902E-04 | global batch size: 512 | lm loss: 1.618817E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.103 | TFLOPs: 43.24 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8300 | lm loss value: 4.161705E+00 | lm loss PPL: 6.418089E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8300 | lm loss value: 4.190990E+00 | lm loss PPL: 6.608818E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8310/ 52452 | consumed samples: 4254720 | consumed tokens: 8713666560 | elapsed time per iteration (s): 9.13 | learning rate: 1.902E-04 | global batch size: 512 | lm loss: 1.612859E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.054 | TFLOPs: 8.41 | 31: iteration 8320/ 52452 | consumed samples: 4259840 | consumed tokens: 8724152320 | elapsed time per iteration (s): 1.77 | learning rate: 1.902E-04 | global batch size: 512 | lm loss: 1.628766E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.399 | TFLOPs: 43.44 | 31: iteration 8330/ 52452 | consumed samples: 4264960 | consumed tokens: 8734638080 | elapsed time per iteration (s): 1.82 | learning rate: 1.902E-04 | global batch size: 512 | lm loss: 1.628252E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.727 | TFLOPs: 42.14 | 31: iteration 8340/ 52452 | consumed samples: 4270080 | consumed tokens: 8745123840 | elapsed time per iteration (s): 1.79 | learning rate: 1.901E-04 | global batch size: 512 | lm loss: 1.625692E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.434 | TFLOPs: 42.99 | 31: iteration 8350/ 52452 | consumed samples: 4275200 | consumed tokens: 8755609600 | elapsed time per iteration (s): 1.79 | learning rate: 1.901E-04 | global batch size: 512 | lm loss: 1.624124E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.524 | TFLOPs: 43.01 | 31: iteration 8360/ 52452 | consumed samples: 4280320 | consumed tokens: 8766095360 | elapsed time per iteration (s): 1.91 | learning rate: 1.901E-04 | global batch size: 512 | lm loss: 1.609443E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.356 | TFLOPs: 40.28 | 31: iteration 8370/ 52452 | consumed samples: 4285440 | consumed tokens: 8776581120 | elapsed time per iteration (s): 1.77 | learning rate: 1.901E-04 | global batch size: 512 | lm loss: 1.628031E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.862 | TFLOPs: 43.36 | 31: iteration 8380/ 52452 | consumed samples: 4290560 | consumed tokens: 8787066880 | elapsed time per iteration (s): 1.77 | learning rate: 1.900E-04 | global batch size: 512 | lm loss: 1.628746E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.642 | TFLOPs: 43.32 | 31: iteration 8390/ 52452 | consumed samples: 4295680 | consumed tokens: 8797552640 | elapsed time per iteration (s): 1.80 | learning rate: 1.900E-04 | global batch size: 512 | lm loss: 1.635685E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.201 | TFLOPs: 42.66 | 31: iteration 8400/ 52452 | consumed samples: 4300800 | consumed tokens: 8808038400 | elapsed time per iteration (s): 1.82 | learning rate: 1.900E-04 | global batch size: 512 | lm loss: 1.616667E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.880 | TFLOPs: 42.16 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8400 | lm loss value: 4.124664E+00 | lm loss PPL: 6.184704E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8400 | lm loss value: 4.116138E+00 | lm loss PPL: 6.132193E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8410/ 52452 | consumed samples: 4305920 | consumed tokens: 8818524160 | elapsed time per iteration (s): 9.20 | learning rate: 1.900E-04 | global batch size: 512 | lm loss: 1.624491E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.644 | TFLOPs: 8.35 | 31: iteration 8420/ 52452 | consumed samples: 4311040 | consumed tokens: 8829009920 | elapsed time per iteration (s): 1.75 | learning rate: 1.899E-04 | global batch size: 512 | lm loss: 1.608883E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.948 | TFLOPs: 43.82 | 31: iteration 8430/ 52452 | consumed samples: 4316160 | consumed tokens: 8839495680 | elapsed time per iteration (s): 1.79 | learning rate: 1.899E-04 | global batch size: 512 | lm loss: 1.609587E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.322 | TFLOPs: 42.98 | 31: iteration 8440/ 52452 | consumed samples: 4321280 | consumed tokens: 8849981440 | elapsed time per iteration (s): 1.88 | learning rate: 1.899E-04 | global batch size: 512 | lm loss: 1.625753E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.813 | TFLOPs: 40.95 | 31: iteration 8450/ 52452 | consumed samples: 4326400 | consumed tokens: 8860467200 | elapsed time per iteration (s): 1.85 | learning rate: 1.899E-04 | global batch size: 512 | lm loss: 1.622435E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.880 | TFLOPs: 41.56 | 31: iteration 8460/ 52452 | consumed samples: 4331520 | consumed tokens: 8870952960 | elapsed time per iteration (s): 1.77 | learning rate: 1.898E-04 | global batch size: 512 | lm loss: 1.620984E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.345 | TFLOPs: 43.43 | 31: iteration 8470/ 52452 | consumed samples: 4336640 | consumed tokens: 8881438720 | elapsed time per iteration (s): 1.82 | learning rate: 1.898E-04 | global batch size: 512 | lm loss: 1.600032E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.254 | TFLOPs: 42.21 | 31: iteration 8480/ 52452 | consumed samples: 4341760 | consumed tokens: 8891924480 | elapsed time per iteration (s): 1.80 | learning rate: 1.898E-04 | global batch size: 512 | lm loss: 1.628203E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.792 | TFLOPs: 42.60 | 31: iteration 8490/ 52452 | consumed samples: 4346880 | consumed tokens: 8902410240 | elapsed time per iteration (s): 1.83 | learning rate: 1.898E-04 | global batch size: 512 | lm loss: 1.624627E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.377 | TFLOPs: 41.93 | 31: iteration 8500/ 52452 | consumed samples: 4352000 | consumed tokens: 8912896000 | elapsed time per iteration (s): 1.76 | learning rate: 1.897E-04 | global batch size: 512 | lm loss: 1.617713E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.302 | TFLOPs: 43.72 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8500 | lm loss value: 4.158537E+00 | lm loss PPL: 6.397785E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8500 | lm loss value: 4.157943E+00 | lm loss PPL: 6.393988E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8510/ 52452 | consumed samples: 4357120 | consumed tokens: 8923381760 | elapsed time per iteration (s): 9.17 | learning rate: 1.897E-04 | global batch size: 512 | lm loss: 1.611556E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.856 | TFLOPs: 8.38 | 31: iteration 8520/ 52452 | consumed samples: 4362240 | consumed tokens: 8933867520 | elapsed time per iteration (s): 2.35 | learning rate: 1.897E-04 | global batch size: 512 | lm loss: 1.624492E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 217.440 | TFLOPs: 32.64 | 31: iteration 8530/ 52452 | consumed samples: 4367360 | consumed tokens: 8944353280 | elapsed time per iteration (s): 1.84 | learning rate: 1.896E-04 | global batch size: 512 | lm loss: 1.618559E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.285 | TFLOPs: 41.77 | 31: iteration 8540/ 52452 | consumed samples: 4372480 | consumed tokens: 8954839040 | elapsed time per iteration (s): 1.82 | learning rate: 1.896E-04 | global batch size: 512 | lm loss: 1.610342E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.199 | TFLOPs: 42.21 | 31: iteration 8550/ 52452 | consumed samples: 4377600 | consumed tokens: 8965324800 | elapsed time per iteration (s): 1.77 | learning rate: 1.896E-04 | global batch size: 512 | lm loss: 1.611927E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.132 | TFLOPs: 43.40 | 31: iteration 8560/ 52452 | consumed samples: 4382720 | consumed tokens: 8975810560 | elapsed time per iteration (s): 1.81 | learning rate: 1.896E-04 | global batch size: 512 | lm loss: 1.615833E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.345 | TFLOPs: 42.53 | 31: iteration 8570/ 52452 | consumed samples: 4387840 | consumed tokens: 8986296320 | elapsed time per iteration (s): 1.76 | learning rate: 1.895E-04 | global batch size: 512 | lm loss: 1.610117E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.142 | TFLOPs: 43.70 | 31: iteration 8580/ 52452 | consumed samples: 4392960 | consumed tokens: 8996782080 | elapsed time per iteration (s): 1.81 | learning rate: 1.895E-04 | global batch size: 512 | lm loss: 1.612796E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.198 | TFLOPs: 42.51 | 31: iteration 8590/ 52452 | consumed samples: 4398080 | consumed tokens: 9007267840 | elapsed time per iteration (s): 1.88 | learning rate: 1.895E-04 | global batch size: 512 | lm loss: 1.618205E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.809 | TFLOPs: 40.80 | 31: iteration 8600/ 52452 | consumed samples: 4403200 | consumed tokens: 9017753600 | elapsed time per iteration (s): 1.84 | learning rate: 1.895E-04 | global batch size: 512 | lm loss: 1.625578E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.860 | TFLOPs: 41.86 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8600 | lm loss value: 4.242100E+00 | lm loss PPL: 6.955378E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8600 | lm loss value: 4.239196E+00 | lm loss PPL: 6.935206E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8610/ 52452 | consumed samples: 4408320 | consumed tokens: 9028239360 | elapsed time per iteration (s): 9.22 | learning rate: 1.894E-04 | global batch size: 512 | lm loss: 1.603963E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.522 | TFLOPs: 8.33 | 31: iteration 8620/ 52452 | consumed samples: 4413440 | consumed tokens: 9038725120 | elapsed time per iteration (s): 1.78 | learning rate: 1.894E-04 | global batch size: 512 | lm loss: 1.615768E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.186 | TFLOPs: 43.26 | 31: iteration 8630/ 52452 | consumed samples: 4418560 | consumed tokens: 9049210880 | elapsed time per iteration (s): 1.88 | learning rate: 1.894E-04 | global batch size: 512 | lm loss: 1.611568E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.858 | TFLOPs: 40.80 | 31: iteration 8640/ 52452 | consumed samples: 4423680 | consumed tokens: 9059696640 | elapsed time per iteration (s): 1.83 | learning rate: 1.894E-04 | global batch size: 512 | lm loss: 1.610118E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.658 | TFLOPs: 41.98 | 31: iteration 8650/ 52452 | consumed samples: 4428800 | consumed tokens: 9070182400 | elapsed time per iteration (s): 1.80 | learning rate: 1.893E-04 | global batch size: 512 | lm loss: 1.607404E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.720 | TFLOPs: 42.73 | 31: iteration 8660/ 52452 | consumed samples: 4433920 | consumed tokens: 9080668160 | elapsed time per iteration (s): 1.80 | learning rate: 1.893E-04 | global batch size: 512 | lm loss: 1.609016E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.334 | TFLOPs: 42.68 | 31: iteration 8670/ 52452 | consumed samples: 4439040 | consumed tokens: 9091153920 | elapsed time per iteration (s): 1.94 | learning rate: 1.893E-04 | global batch size: 512 | lm loss: 1.616780E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 263.830 | TFLOPs: 39.60 | 31: iteration 8680/ 52452 | consumed samples: 4444160 | consumed tokens: 9101639680 | elapsed time per iteration (s): 1.79 | learning rate: 1.893E-04 | global batch size: 512 | lm loss: 1.621531E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.101 | TFLOPs: 42.94 | 31: iteration 8690/ 52452 | consumed samples: 4449280 | consumed tokens: 9112125440 | elapsed time per iteration (s): 1.81 | learning rate: 1.892E-04 | global batch size: 512 | lm loss: 1.603119E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.153 | TFLOPs: 42.35 | 31: iteration 8700/ 52452 | consumed samples: 4454400 | consumed tokens: 9122611200 | elapsed time per iteration (s): 1.73 | learning rate: 1.892E-04 | global batch size: 512 | lm loss: 1.605603E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.088 | TFLOPs: 44.44 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8700 | lm loss value: 4.034922E+00 | lm loss PPL: 5.653849E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8700 | lm loss value: 4.069071E+00 | lm loss PPL: 5.850261E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8710/ 52452 | consumed samples: 4459520 | consumed tokens: 9133096960 | elapsed time per iteration (s): 9.17 | learning rate: 1.892E-04 | global batch size: 512 | lm loss: 1.619792E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.833 | TFLOPs: 8.38 | 31: iteration 8720/ 52452 | consumed samples: 4464640 | consumed tokens: 9143582720 | elapsed time per iteration (s): 1.77 | learning rate: 1.892E-04 | global batch size: 512 | lm loss: 1.618806E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.168 | TFLOPs: 43.40 | 31: iteration 8730/ 52452 | consumed samples: 4469760 | consumed tokens: 9154068480 | elapsed time per iteration (s): 1.76 | learning rate: 1.891E-04 | global batch size: 512 | lm loss: 1.620889E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.315 | TFLOPs: 43.72 | 31: iteration 8740/ 52452 | consumed samples: 4474880 | consumed tokens: 9164554240 | elapsed time per iteration (s): 1.75 | learning rate: 1.891E-04 | global batch size: 512 | lm loss: 1.618266E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.653 | TFLOPs: 43.93 | 31: iteration 8750/ 52452 | consumed samples: 4480000 | consumed tokens: 9175040000 | elapsed time per iteration (s): 1.76 | learning rate: 1.891E-04 | global batch size: 512 | lm loss: 1.602591E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.434 | TFLOPs: 43.74 | 31: iteration 8760/ 52452 | consumed samples: 4485120 | consumed tokens: 9185525760 | elapsed time per iteration (s): 1.84 | learning rate: 1.891E-04 | global batch size: 512 | lm loss: 1.615123E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.213 | TFLOPs: 41.76 | 31: iteration 8770/ 52452 | consumed samples: 4490240 | consumed tokens: 9196011520 | elapsed time per iteration (s): 1.75 | learning rate: 1.890E-04 | global batch size: 512 | lm loss: 1.608274E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.400 | TFLOPs: 43.89 | 31: iteration 8780/ 52452 | consumed samples: 4495360 | consumed tokens: 9206497280 | elapsed time per iteration (s): 1.81 | learning rate: 1.890E-04 | global batch size: 512 | lm loss: 1.597750E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.460 | TFLOPs: 42.40 | 31: iteration 8790/ 52452 | consumed samples: 4500480 | consumed tokens: 9216983040 | elapsed time per iteration (s): 1.83 | learning rate: 1.890E-04 | global batch size: 512 | lm loss: 1.594024E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.266 | TFLOPs: 42.07 | 31: iteration 8800/ 52452 | consumed samples: 4505600 | consumed tokens: 9227468800 | elapsed time per iteration (s): 1.86 | learning rate: 1.890E-04 | global batch size: 512 | lm loss: 1.606091E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.869 | TFLOPs: 41.41 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8800 | lm loss value: 4.133986E+00 | lm loss PPL: 6.242626E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8800 | lm loss value: 4.159809E+00 | lm loss PPL: 6.405926E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8810/ 52452 | consumed samples: 4510720 | consumed tokens: 9237954560 | elapsed time per iteration (s): 9.18 | learning rate: 1.889E-04 | global batch size: 512 | lm loss: 1.608607E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.776 | TFLOPs: 8.37 | 31: iteration 8820/ 52452 | consumed samples: 4515840 | consumed tokens: 9248440320 | elapsed time per iteration (s): 1.79 | learning rate: 1.889E-04 | global batch size: 512 | lm loss: 1.612810E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.874 | TFLOPs: 42.91 | 31: iteration 8830/ 52452 | consumed samples: 4520960 | consumed tokens: 9258926080 | elapsed time per iteration (s): 1.90 | learning rate: 1.889E-04 | global batch size: 512 | lm loss: 1.604686E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.090 | TFLOPs: 40.54 | 31: iteration 8840/ 52452 | consumed samples: 4526080 | consumed tokens: 9269411840 | elapsed time per iteration (s): 1.80 | learning rate: 1.888E-04 | global batch size: 512 | lm loss: 1.610367E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.133 | TFLOPs: 42.65 | 31: iteration 8850/ 52452 | consumed samples: 4531200 | consumed tokens: 9279897600 | elapsed time per iteration (s): 1.79 | learning rate: 1.888E-04 | global batch size: 512 | lm loss: 1.615791E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.063 | TFLOPs: 42.94 | 31: iteration 8860/ 52452 | consumed samples: 4536320 | consumed tokens: 9290383360 | elapsed time per iteration (s): 1.83 | learning rate: 1.888E-04 | global batch size: 512 | lm loss: 1.611020E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.895 | TFLOPs: 42.01 | 31: iteration 8870/ 52452 | consumed samples: 4541440 | consumed tokens: 9300869120 | elapsed time per iteration (s): 1.76 | learning rate: 1.888E-04 | global batch size: 512 | lm loss: 1.602143E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.957 | TFLOPs: 43.67 | 31: iteration 8880/ 52452 | consumed samples: 4546560 | consumed tokens: 9311354880 | elapsed time per iteration (s): 1.84 | learning rate: 1.887E-04 | global batch size: 512 | lm loss: 1.610285E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.456 | TFLOPs: 41.79 | 31: iteration 8890/ 52452 | consumed samples: 4551680 | consumed tokens: 9321840640 | elapsed time per iteration (s): 1.81 | learning rate: 1.887E-04 | global batch size: 512 | lm loss: 1.599707E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.488 | TFLOPs: 42.40 | 31: iteration 8900/ 52452 | consumed samples: 4556800 | consumed tokens: 9332326400 | elapsed time per iteration (s): 1.77 | learning rate: 1.887E-04 | global batch size: 512 | lm loss: 1.610845E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.353 | TFLOPs: 43.43 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 8900 | lm loss value: 4.094787E+00 | lm loss PPL: 6.002656E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 8900 | lm loss value: 4.120077E+00 | lm loss PPL: 6.156396E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 8910/ 52452 | consumed samples: 4561920 | consumed tokens: 9342812160 | elapsed time per iteration (s): 9.19 | learning rate: 1.887E-04 | global batch size: 512 | lm loss: 1.605096E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.693 | TFLOPs: 8.36 | 31: iteration 8920/ 52452 | consumed samples: 4567040 | consumed tokens: 9353297920 | elapsed time per iteration (s): 1.78 | learning rate: 1.886E-04 | global batch size: 512 | lm loss: 1.613046E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.183 | TFLOPs: 43.10 | 31: iteration 8930/ 52452 | consumed samples: 4572160 | consumed tokens: 9363783680 | elapsed time per iteration (s): 1.78 | learning rate: 1.886E-04 | global batch size: 512 | lm loss: 1.594904E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.146 | TFLOPs: 43.25 | 31: iteration 8940/ 52452 | consumed samples: 4577280 | consumed tokens: 9374269440 | elapsed time per iteration (s): 1.79 | learning rate: 1.886E-04 | global batch size: 512 | lm loss: 1.609064E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.368 | TFLOPs: 42.98 | 31: iteration 8950/ 52452 | consumed samples: 4582400 | consumed tokens: 9384755200 | elapsed time per iteration (s): 1.76 | learning rate: 1.886E-04 | global batch size: 512 | lm loss: 1.615924E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.774 | TFLOPs: 43.64 | 31: iteration 8960/ 52452 | consumed samples: 4587520 | consumed tokens: 9395240960 | elapsed time per iteration (s): 2.15 | learning rate: 1.885E-04 | global batch size: 512 | lm loss: 1.605116E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 237.964 | TFLOPs: 35.72 | 31: iteration 8970/ 52452 | consumed samples: 4592640 | consumed tokens: 9405726720 | elapsed time per iteration (s): 1.87 | learning rate: 1.885E-04 | global batch size: 512 | lm loss: 1.603624E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.053 | TFLOPs: 41.13 | 31: iteration 8980/ 52452 | consumed samples: 4597760 | consumed tokens: 9416212480 | elapsed time per iteration (s): 1.79 | learning rate: 1.885E-04 | global batch size: 512 | lm loss: 1.603541E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.297 | TFLOPs: 42.82 | 31: iteration 8990/ 52452 | consumed samples: 4602880 | consumed tokens: 9426698240 | elapsed time per iteration (s): 1.77 | learning rate: 1.885E-04 | global batch size: 512 | lm loss: 1.602142E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.534 | TFLOPs: 43.31 | 31: iteration 9000/ 52452 | consumed samples: 4608000 | consumed tokens: 9437184000 | elapsed time per iteration (s): 1.77 | learning rate: 1.884E-04 | global batch size: 512 | lm loss: 1.600265E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.647 | TFLOPs: 43.32 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9000 | lm loss value: 4.003921E+00 | lm loss PPL: 5.481265E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9000 | lm loss value: 4.032845E+00 | lm loss PPL: 5.642123E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9010/ 52452 | consumed samples: 4613120 | consumed tokens: 9447669760 | elapsed time per iteration (s): 9.16 | learning rate: 1.884E-04 | global batch size: 512 | lm loss: 1.609662E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.917 | TFLOPs: 8.39 | 31: iteration 9020/ 52452 | consumed samples: 4618240 | consumed tokens: 9458155520 | elapsed time per iteration (s): 1.80 | learning rate: 1.884E-04 | global batch size: 512 | lm loss: 1.601808E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.189 | TFLOPs: 42.66 | 31: iteration 9030/ 52452 | consumed samples: 4623360 | consumed tokens: 9468641280 | elapsed time per iteration (s): 1.76 | learning rate: 1.883E-04 | global batch size: 512 | lm loss: 1.604726E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.526 | TFLOPs: 43.76 | 31: iteration 9040/ 52452 | consumed samples: 4628480 | consumed tokens: 9479127040 | elapsed time per iteration (s): 1.76 | learning rate: 1.883E-04 | global batch size: 512 | lm loss: 1.599351E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.220 | TFLOPs: 43.71 | 31: iteration 9050/ 52452 | consumed samples: 4633600 | consumed tokens: 9489612800 | elapsed time per iteration (s): 1.79 | learning rate: 1.883E-04 | global batch size: 512 | lm loss: 1.600260E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.368 | TFLOPs: 42.83 | 31: iteration 9060/ 52452 | consumed samples: 4638720 | consumed tokens: 9500098560 | elapsed time per iteration (s): 1.78 | learning rate: 1.883E-04 | global batch size: 512 | lm loss: 1.608232E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.755 | TFLOPs: 43.19 | 31: iteration 9070/ 52452 | consumed samples: 4643840 | consumed tokens: 9510584320 | elapsed time per iteration (s): 1.74 | learning rate: 1.882E-04 | global batch size: 512 | lm loss: 1.602258E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.041 | TFLOPs: 44.13 | 31: iteration 9080/ 52452 | consumed samples: 4648960 | consumed tokens: 9521070080 | elapsed time per iteration (s): 1.74 | learning rate: 1.882E-04 | global batch size: 512 | lm loss: 1.592140E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.049 | TFLOPs: 44.14 | 31: iteration 9090/ 52452 | consumed samples: 4654080 | consumed tokens: 9531555840 | elapsed time per iteration (s): 2.02 | learning rate: 1.882E-04 | global batch size: 512 | lm loss: 1.603278E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 252.945 | TFLOPs: 37.97 | 31: iteration 9100/ 52452 | consumed samples: 4659200 | consumed tokens: 9542041600 | elapsed time per iteration (s): 1.84 | learning rate: 1.882E-04 | global batch size: 512 | lm loss: 1.594409E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.932 | TFLOPs: 41.87 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9100 | lm loss value: 4.133159E+00 | lm loss PPL: 6.237466E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9100 | lm loss value: 4.150027E+00 | lm loss PPL: 6.343573E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9110/ 52452 | consumed samples: 4664320 | consumed tokens: 9552527360 | elapsed time per iteration (s): 9.18 | learning rate: 1.881E-04 | global batch size: 512 | lm loss: 1.605100E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.800 | TFLOPs: 8.38 | 31: iteration 9120/ 52452 | consumed samples: 4669440 | consumed tokens: 9563013120 | elapsed time per iteration (s): 1.75 | learning rate: 1.881E-04 | global batch size: 512 | lm loss: 1.592406E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.901 | TFLOPs: 43.81 | 31: iteration 9130/ 52452 | consumed samples: 4674560 | consumed tokens: 9573498880 | elapsed time per iteration (s): 1.79 | learning rate: 1.881E-04 | global batch size: 512 | lm loss: 1.602109E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.656 | TFLOPs: 43.03 | 31: iteration 9140/ 52452 | consumed samples: 4679680 | consumed tokens: 9583984640 | elapsed time per iteration (s): 1.77 | learning rate: 1.880E-04 | global batch size: 512 | lm loss: 1.602066E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.770 | TFLOPs: 43.34 | 31: iteration 9150/ 52452 | consumed samples: 4684800 | consumed tokens: 9594470400 | elapsed time per iteration (s): 1.78 | learning rate: 1.880E-04 | global batch size: 512 | lm loss: 1.599457E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.600 | TFLOPs: 43.17 | 31: iteration 9160/ 52452 | consumed samples: 4689920 | consumed tokens: 9604956160 | elapsed time per iteration (s): 1.76 | learning rate: 1.880E-04 | global batch size: 512 | lm loss: 1.584523E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.469 | TFLOPs: 43.60 | 31: iteration 9170/ 52452 | consumed samples: 4695040 | consumed tokens: 9615441920 | elapsed time per iteration (s): 1.79 | learning rate: 1.880E-04 | global batch size: 512 | lm loss: 1.598309E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.816 | TFLOPs: 43.05 | 31: iteration 9180/ 52452 | consumed samples: 4700160 | consumed tokens: 9625927680 | elapsed time per iteration (s): 1.79 | learning rate: 1.879E-04 | global batch size: 512 | lm loss: 1.603224E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.614 | TFLOPs: 43.02 | 31: iteration 9190/ 52452 | consumed samples: 4705280 | consumed tokens: 9636413440 | elapsed time per iteration (s): 1.76 | learning rate: 1.879E-04 | global batch size: 512 | lm loss: 1.608533E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.042 | TFLOPs: 43.68 | 31: iteration 9200/ 52452 | consumed samples: 4710400 | consumed tokens: 9646899200 | elapsed time per iteration (s): 1.79 | learning rate: 1.879E-04 | global batch size: 512 | lm loss: 1.592882E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.029 | TFLOPs: 42.93 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9200 | lm loss value: 4.112364E+00 | lm loss PPL: 6.109098E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9200 | lm loss value: 4.113488E+00 | lm loss PPL: 6.115968E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9210/ 52452 | consumed samples: 4715520 | consumed tokens: 9657384960 | elapsed time per iteration (s): 9.18 | learning rate: 1.879E-04 | global batch size: 512 | lm loss: 1.591390E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.776 | TFLOPs: 8.37 | 31: iteration 9220/ 52452 | consumed samples: 4720640 | consumed tokens: 9667870720 | elapsed time per iteration (s): 1.81 | learning rate: 1.878E-04 | global batch size: 512 | lm loss: 1.581251E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.370 | TFLOPs: 42.38 | 31: iteration 9230/ 52452 | consumed samples: 4725760 | consumed tokens: 9678356480 | elapsed time per iteration (s): 1.90 | learning rate: 1.878E-04 | global batch size: 512 | lm loss: 1.589979E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.150 | TFLOPs: 40.55 | 31: iteration 9240/ 52452 | consumed samples: 4730880 | consumed tokens: 9688842240 | elapsed time per iteration (s): 1.81 | learning rate: 1.878E-04 | global batch size: 512 | lm loss: 1.592381E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.305 | TFLOPs: 42.52 | 31: iteration 9250/ 52452 | consumed samples: 4736000 | consumed tokens: 9699328000 | elapsed time per iteration (s): 1.75 | learning rate: 1.877E-04 | global batch size: 512 | lm loss: 1.597541E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.831 | TFLOPs: 43.80 | 31: iteration 9260/ 52452 | consumed samples: 4741120 | consumed tokens: 9709813760 | elapsed time per iteration (s): 1.74 | learning rate: 1.877E-04 | global batch size: 512 | lm loss: 1.594475E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.603 | TFLOPs: 44.22 | 31: iteration 9270/ 52452 | consumed samples: 4746240 | consumed tokens: 9720299520 | elapsed time per iteration (s): 1.74 | learning rate: 1.877E-04 | global batch size: 512 | lm loss: 1.597566E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.394 | TFLOPs: 44.19 | 31: iteration 9280/ 52452 | consumed samples: 4751360 | consumed tokens: 9730785280 | elapsed time per iteration (s): 1.82 | learning rate: 1.877E-04 | global batch size: 512 | lm loss: 1.597070E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.676 | TFLOPs: 42.28 | 31: iteration 9290/ 52452 | consumed samples: 4756480 | consumed tokens: 9741271040 | elapsed time per iteration (s): 1.77 | learning rate: 1.876E-04 | global batch size: 512 | lm loss: 1.586897E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.736 | TFLOPs: 43.34 | 31: iteration 9300/ 52452 | consumed samples: 4761600 | consumed tokens: 9751756800 | elapsed time per iteration (s): 1.82 | learning rate: 1.876E-04 | global batch size: 512 | lm loss: 1.591919E+00 | grad norm: 0.228 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.340 | TFLOPs: 42.23 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9300 | lm loss value: 4.044748E+00 | lm loss PPL: 5.709681E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9300 | lm loss value: 4.064533E+00 | lm loss PPL: 5.823769E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9310/ 52452 | consumed samples: 4766720 | consumed tokens: 9762242560 | elapsed time per iteration (s): 9.22 | learning rate: 1.876E-04 | global batch size: 512 | lm loss: 1.596844E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.512 | TFLOPs: 8.33 | 31: iteration 9320/ 52452 | consumed samples: 4771840 | consumed tokens: 9772728320 | elapsed time per iteration (s): 1.76 | learning rate: 1.876E-04 | global batch size: 512 | lm loss: 1.606998E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.353 | TFLOPs: 43.58 | 31: iteration 9330/ 52452 | consumed samples: 4776960 | consumed tokens: 9783214080 | elapsed time per iteration (s): 1.73 | learning rate: 1.875E-04 | global batch size: 512 | lm loss: 1.593936E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.278 | TFLOPs: 44.32 | 31: iteration 9340/ 52452 | consumed samples: 4782080 | consumed tokens: 9793699840 | elapsed time per iteration (s): 1.77 | learning rate: 1.875E-04 | global batch size: 512 | lm loss: 1.582413E+00 | grad norm: 0.210 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.566 | TFLOPs: 43.31 | 31: iteration 9350/ 52452 | consumed samples: 4787200 | consumed tokens: 9804185600 | elapsed time per iteration (s): 1.78 | learning rate: 1.875E-04 | global batch size: 512 | lm loss: 1.591731E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.420 | TFLOPs: 43.29 | 31: iteration 9360/ 52452 | consumed samples: 4792320 | consumed tokens: 9814671360 | elapsed time per iteration (s): 1.73 | learning rate: 1.874E-04 | global batch size: 512 | lm loss: 1.605535E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.179 | TFLOPs: 44.45 | 31: iteration 9370/ 52452 | consumed samples: 4797440 | consumed tokens: 9825157120 | elapsed time per iteration (s): 1.88 | learning rate: 1.874E-04 | global batch size: 512 | lm loss: 1.593922E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.846 | TFLOPs: 40.95 | 31: iteration 9380/ 52452 | consumed samples: 4802560 | consumed tokens: 9835642880 | elapsed time per iteration (s): 1.87 | learning rate: 1.874E-04 | global batch size: 512 | lm loss: 1.596674E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.958 | TFLOPs: 41.12 | 31: iteration 9390/ 52452 | consumed samples: 4807680 | consumed tokens: 9846128640 | elapsed time per iteration (s): 1.78 | learning rate: 1.874E-04 | global batch size: 512 | lm loss: 1.582127E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.019 | TFLOPs: 43.23 | 31: iteration 9400/ 52452 | consumed samples: 4812800 | consumed tokens: 9856614400 | elapsed time per iteration (s): 1.75 | learning rate: 1.873E-04 | global batch size: 512 | lm loss: 1.587295E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.921 | TFLOPs: 43.97 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9400 | lm loss value: 4.142095E+00 | lm loss PPL: 6.293451E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9400 | lm loss value: 4.124533E+00 | lm loss PPL: 6.183891E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9410/ 52452 | consumed samples: 4817920 | consumed tokens: 9867100160 | elapsed time per iteration (s): 9.22 | learning rate: 1.873E-04 | global batch size: 512 | lm loss: 1.587929E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.513 | TFLOPs: 8.33 | 31: iteration 9420/ 52452 | consumed samples: 4823040 | consumed tokens: 9877585920 | elapsed time per iteration (s): 1.80 | learning rate: 1.873E-04 | global batch size: 512 | lm loss: 1.586477E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.384 | TFLOPs: 42.68 | 31: iteration 9430/ 52452 | consumed samples: 4828160 | consumed tokens: 9888071680 | elapsed time per iteration (s): 1.74 | learning rate: 1.873E-04 | global batch size: 512 | lm loss: 1.581303E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.693 | TFLOPs: 44.08 | 31: iteration 9440/ 52452 | consumed samples: 4833280 | consumed tokens: 9898557440 | elapsed time per iteration (s): 1.73 | learning rate: 1.872E-04 | global batch size: 512 | lm loss: 1.593256E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.200 | TFLOPs: 44.31 | 31: iteration 9450/ 52452 | consumed samples: 4838400 | consumed tokens: 9909043200 | elapsed time per iteration (s): 1.81 | learning rate: 1.872E-04 | global batch size: 512 | lm loss: 1.596580E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.190 | TFLOPs: 42.36 | 31: iteration 9460/ 52452 | consumed samples: 4843520 | consumed tokens: 9919528960 | elapsed time per iteration (s): 1.76 | learning rate: 1.872E-04 | global batch size: 512 | lm loss: 1.594779E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.301 | TFLOPs: 43.57 | 31: iteration 9470/ 52452 | consumed samples: 4848640 | consumed tokens: 9930014720 | elapsed time per iteration (s): 1.78 | learning rate: 1.871E-04 | global batch size: 512 | lm loss: 1.574331E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.773 | TFLOPs: 43.19 | 31: iteration 9480/ 52452 | consumed samples: 4853760 | consumed tokens: 9940500480 | elapsed time per iteration (s): 1.77 | learning rate: 1.871E-04 | global batch size: 512 | lm loss: 1.593882E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.005 | TFLOPs: 43.38 | 31: iteration 9490/ 52452 | consumed samples: 4858880 | consumed tokens: 9950986240 | elapsed time per iteration (s): 1.83 | learning rate: 1.871E-04 | global batch size: 512 | lm loss: 1.584959E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.411 | TFLOPs: 41.94 | 31: iteration 9500/ 52452 | consumed samples: 4864000 | consumed tokens: 9961472000 | elapsed time per iteration (s): 1.84 | learning rate: 1.871E-04 | global batch size: 512 | lm loss: 1.585542E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.508 | TFLOPs: 41.80 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9500 | lm loss value: 4.092063E+00 | lm loss PPL: 5.986329E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9500 | lm loss value: 4.085678E+00 | lm loss PPL: 5.948223E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9510/ 52452 | consumed samples: 4869120 | consumed tokens: 9971957760 | elapsed time per iteration (s): 9.27 | learning rate: 1.870E-04 | global batch size: 512 | lm loss: 1.582574E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.246 | TFLOPs: 8.29 | 31: iteration 9520/ 52452 | consumed samples: 4874240 | consumed tokens: 9982443520 | elapsed time per iteration (s): 1.83 | learning rate: 1.870E-04 | global batch size: 512 | lm loss: 1.578868E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.490 | TFLOPs: 42.10 | 31: iteration 9530/ 52452 | consumed samples: 4879360 | consumed tokens: 9992929280 | elapsed time per iteration (s): 1.84 | learning rate: 1.870E-04 | global batch size: 512 | lm loss: 1.571363E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.588 | TFLOPs: 41.81 | 31: iteration 9540/ 52452 | consumed samples: 4884480 | consumed tokens: 10003415040 | elapsed time per iteration (s): 1.77 | learning rate: 1.869E-04 | global batch size: 512 | lm loss: 1.589663E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.859 | TFLOPs: 43.51 | 31: iteration 9550/ 52452 | consumed samples: 4889600 | consumed tokens: 10013900800 | elapsed time per iteration (s): 1.76 | learning rate: 1.869E-04 | global batch size: 512 | lm loss: 1.582337E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.520 | TFLOPs: 43.76 | 31: iteration 9560/ 52452 | consumed samples: 4894720 | consumed tokens: 10024386560 | elapsed time per iteration (s): 1.77 | learning rate: 1.869E-04 | global batch size: 512 | lm loss: 1.577630E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.061 | TFLOPs: 43.54 | 31: iteration 9570/ 52452 | consumed samples: 4899840 | consumed tokens: 10034872320 | elapsed time per iteration (s): 1.74 | learning rate: 1.869E-04 | global batch size: 512 | lm loss: 1.580655E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.605 | TFLOPs: 44.07 | 31: iteration 9580/ 52452 | consumed samples: 4904960 | consumed tokens: 10045358080 | elapsed time per iteration (s): 1.78 | learning rate: 1.868E-04 | global batch size: 512 | lm loss: 1.595565E+00 | grad norm: 0.204 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.161 | TFLOPs: 43.10 | 31: iteration 9590/ 52452 | consumed samples: 4910080 | consumed tokens: 10055843840 | elapsed time per iteration (s): 1.75 | learning rate: 1.868E-04 | global batch size: 512 | lm loss: 1.573888E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.220 | TFLOPs: 44.01 | 31: iteration 9600/ 52452 | consumed samples: 4915200 | consumed tokens: 10066329600 | elapsed time per iteration (s): 1.88 | learning rate: 1.868E-04 | global batch size: 512 | lm loss: 1.578916E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.124 | TFLOPs: 40.84 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9600 | lm loss value: 4.096422E+00 | lm loss PPL: 6.012476E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9600 | lm loss value: 4.104602E+00 | lm loss PPL: 6.061861E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9610/ 52452 | consumed samples: 4920320 | consumed tokens: 10076815360 | elapsed time per iteration (s): 9.13 | learning rate: 1.867E-04 | global batch size: 512 | lm loss: 1.578698E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.081 | TFLOPs: 8.42 | 31: iteration 9620/ 52452 | consumed samples: 4925440 | consumed tokens: 10087301120 | elapsed time per iteration (s): 1.80 | learning rate: 1.867E-04 | global batch size: 512 | lm loss: 1.568465E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.848 | TFLOPs: 42.75 | 31: iteration 9630/ 52452 | consumed samples: 4930560 | consumed tokens: 10097786880 | elapsed time per iteration (s): 1.75 | learning rate: 1.867E-04 | global batch size: 512 | lm loss: 1.577890E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.002 | TFLOPs: 43.83 | 31: iteration 9640/ 52452 | consumed samples: 4935680 | consumed tokens: 10108272640 | elapsed time per iteration (s): 1.78 | learning rate: 1.867E-04 | global batch size: 512 | lm loss: 1.572866E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.403 | TFLOPs: 43.14 | 31: iteration 9650/ 52452 | consumed samples: 4940800 | consumed tokens: 10118758400 | elapsed time per iteration (s): 1.75 | learning rate: 1.866E-04 | global batch size: 512 | lm loss: 1.575772E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.926 | TFLOPs: 43.82 | 31: iteration 9660/ 52452 | consumed samples: 4945920 | consumed tokens: 10129244160 | elapsed time per iteration (s): 1.77 | learning rate: 1.866E-04 | global batch size: 512 | lm loss: 1.574361E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.498 | TFLOPs: 43.45 | 31: iteration 9670/ 52452 | consumed samples: 4951040 | consumed tokens: 10139729920 | elapsed time per iteration (s): 1.91 | learning rate: 1.866E-04 | global batch size: 512 | lm loss: 1.586896E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.182 | TFLOPs: 40.25 | 31: iteration 9680/ 52452 | consumed samples: 4956160 | consumed tokens: 10150215680 | elapsed time per iteration (s): 1.74 | learning rate: 1.865E-04 | global batch size: 512 | lm loss: 1.597312E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.416 | TFLOPs: 44.04 | 31: iteration 9690/ 52452 | consumed samples: 4961280 | consumed tokens: 10160701440 | elapsed time per iteration (s): 1.73 | learning rate: 1.865E-04 | global batch size: 512 | lm loss: 1.585368E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.512 | TFLOPs: 44.35 | 31: iteration 9700/ 52452 | consumed samples: 4966400 | consumed tokens: 10171187200 | elapsed time per iteration (s): 1.81 | learning rate: 1.865E-04 | global batch size: 512 | lm loss: 1.582585E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.437 | TFLOPs: 42.54 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9700 | lm loss value: 4.088974E+00 | lm loss PPL: 5.967863E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9700 | lm loss value: 4.101309E+00 | lm loss PPL: 6.041934E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9710/ 52452 | consumed samples: 4971520 | consumed tokens: 10181672960 | elapsed time per iteration (s): 9.19 | learning rate: 1.865E-04 | global batch size: 512 | lm loss: 1.591906E+00 | grad norm: 0.202 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.702 | TFLOPs: 8.36 | 31: iteration 9720/ 52452 | consumed samples: 4976640 | consumed tokens: 10192158720 | elapsed time per iteration (s): 1.72 | learning rate: 1.864E-04 | global batch size: 512 | lm loss: 1.586195E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.032 | TFLOPs: 44.58 | 31: iteration 9730/ 52452 | consumed samples: 4981760 | consumed tokens: 10202644480 | elapsed time per iteration (s): 1.75 | learning rate: 1.864E-04 | global batch size: 512 | lm loss: 1.575092E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.830 | TFLOPs: 43.95 | 31: iteration 9740/ 52452 | consumed samples: 4986880 | consumed tokens: 10213130240 | elapsed time per iteration (s): 1.75 | learning rate: 1.864E-04 | global batch size: 512 | lm loss: 1.592104E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.406 | TFLOPs: 44.04 | 31: iteration 9750/ 52452 | consumed samples: 4992000 | consumed tokens: 10223616000 | elapsed time per iteration (s): 1.76 | learning rate: 1.863E-04 | global batch size: 512 | lm loss: 1.597518E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.507 | TFLOPs: 43.60 | 31: iteration 9760/ 52452 | consumed samples: 4997120 | consumed tokens: 10234101760 | elapsed time per iteration (s): 1.74 | learning rate: 1.863E-04 | global batch size: 512 | lm loss: 1.582038E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.455 | TFLOPs: 44.20 | 31: iteration 9770/ 52452 | consumed samples: 5002240 | consumed tokens: 10244587520 | elapsed time per iteration (s): 1.74 | learning rate: 1.863E-04 | global batch size: 512 | lm loss: 1.577059E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.499 | TFLOPs: 44.20 | 31: iteration 9780/ 52452 | consumed samples: 5007360 | consumed tokens: 10255073280 | elapsed time per iteration (s): 1.92 | learning rate: 1.863E-04 | global batch size: 512 | lm loss: 1.583936E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.089 | TFLOPs: 40.09 | 31: iteration 9790/ 52452 | consumed samples: 5012480 | consumed tokens: 10265559040 | elapsed time per iteration (s): 1.84 | learning rate: 1.862E-04 | global batch size: 512 | lm loss: 1.581484E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.226 | TFLOPs: 41.76 | 31: iteration 9800/ 52452 | consumed samples: 5017600 | consumed tokens: 10276044800 | elapsed time per iteration (s): 1.76 | learning rate: 1.862E-04 | global batch size: 512 | lm loss: 1.573798E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.194 | TFLOPs: 43.71 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9800 | lm loss value: 4.014111E+00 | lm loss PPL: 5.537405E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9800 | lm loss value: 4.049327E+00 | lm loss PPL: 5.735886E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9810/ 52452 | consumed samples: 5022720 | consumed tokens: 10286530560 | elapsed time per iteration (s): 9.12 | learning rate: 1.862E-04 | global batch size: 512 | lm loss: 1.590923E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.162 | TFLOPs: 8.43 | 31: iteration 9820/ 52452 | consumed samples: 5027840 | consumed tokens: 10297016320 | elapsed time per iteration (s): 1.74 | learning rate: 1.861E-04 | global batch size: 512 | lm loss: 1.568628E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.667 | TFLOPs: 44.08 | 31: iteration 9830/ 52452 | consumed samples: 5032960 | consumed tokens: 10307502080 | elapsed time per iteration (s): 1.73 | learning rate: 1.861E-04 | global batch size: 512 | lm loss: 1.581337E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.208 | TFLOPs: 44.31 | 31: iteration 9840/ 52452 | consumed samples: 5038080 | consumed tokens: 10317987840 | elapsed time per iteration (s): 1.77 | learning rate: 1.861E-04 | global batch size: 512 | lm loss: 1.574205E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.495 | TFLOPs: 43.30 | 31: iteration 9850/ 52452 | consumed samples: 5043200 | consumed tokens: 10328473600 | elapsed time per iteration (s): 1.78 | learning rate: 1.861E-04 | global batch size: 512 | lm loss: 1.559393E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.216 | TFLOPs: 43.26 | 31: iteration 9860/ 52452 | consumed samples: 5048320 | consumed tokens: 10338959360 | elapsed time per iteration (s): 1.75 | learning rate: 1.860E-04 | global batch size: 512 | lm loss: 1.575911E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.914 | TFLOPs: 43.96 | 31: iteration 9870/ 52452 | consumed samples: 5053440 | consumed tokens: 10349445120 | elapsed time per iteration (s): 1.78 | learning rate: 1.860E-04 | global batch size: 512 | lm loss: 1.584328E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.384 | TFLOPs: 43.13 | 31: iteration 9880/ 52452 | consumed samples: 5058560 | consumed tokens: 10359930880 | elapsed time per iteration (s): 1.73 | learning rate: 1.860E-04 | global batch size: 512 | lm loss: 1.573632E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.355 | TFLOPs: 44.48 | 31: iteration 9890/ 52452 | consumed samples: 5063680 | consumed tokens: 10370416640 | elapsed time per iteration (s): 1.74 | learning rate: 1.859E-04 | global batch size: 512 | lm loss: 1.569667E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.792 | TFLOPs: 44.10 | 31: iteration 9900/ 52452 | consumed samples: 5068800 | consumed tokens: 10380902400 | elapsed time per iteration (s): 1.75 | learning rate: 1.859E-04 | global batch size: 512 | lm loss: 1.567549E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.997 | TFLOPs: 43.83 | 31: -------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 9900 | lm loss value: 4.146947E+00 | lm loss PPL: 6.324063E+01 | 31: -------------------------------------------------------------------------------------------------- 31: ----------------------------------------------------------------------------------------------------- 31: validation_oscar loss at iteration 9900 | lm loss value: 4.160519E+00 | lm loss PPL: 6.410479E+01 | 31: ----------------------------------------------------------------------------------------------------- 31: iteration 9910/ 52452 | consumed samples: 5073920 | consumed tokens: 10391388160 | elapsed time per iteration (s): 9.12 | learning rate: 1.859E-04 | global batch size: 512 | lm loss: 1.555509E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.147 | TFLOPs: 8.43 | 31: iteration 9920/ 52452 | consumed samples: 5079040 | consumed tokens: 10401873920 | elapsed time per iteration (s): 1.73 | learning rate: 1.858E-04 | global batch size: 512 | lm loss: 1.567139E+00 | grad norm: 0.220 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.025 | TFLOPs: 44.43 | 31: iteration 9930/ 52452 | consumed samples: 5084160 | consumed tokens: 10412359680 | elapsed time per iteration (s): 1.73 | learning rate: 1.858E-04 | global batch size: 512 | lm loss: 1.567512E+00 | grad norm: 0.258 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.444 | TFLOPs: 44.49 | 31: iteration 9940/ 52452 | consumed samples: 5089280 | consumed tokens: 10422845440 | elapsed time per iteration (s): 1.76 | learning rate: 1.858E-04 | global batch size: 512 | lm loss: 1.568480E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.441 | TFLOPs: 43.74 | 31: iteration 9950/ 52452 | consumed samples: 5094400 | consumed tokens: 10433331200 | elapsed time per iteration (s): 1.75 | learning rate: 1.858E-04 | global batch size: 512 | lm loss: 1.560948E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.120 | TFLOPs: 43.85 | 31: iteration 9960/ 52452 | consumed samples: 5099520 | consumed tokens: 10443816960 | elapsed time per iteration (s): 1.79 | learning rate: 1.857E-04 | global batch size: 512 | lm loss: 1.562306E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.657 | TFLOPs: 43.03 | 31: iteration 9970/ 52452 | consumed samples: 5104640 | consumed tokens: 10454302720 | elapsed time per iteration (s): 1.73 | learning rate: 1.857E-04 | global batch size: 512 | lm loss: 1.569455E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.333 | TFLOPs: 44.33 | 31: iteration 9980/ 52452 | consumed samples: 5109760 | consumed tokens: 10464788480 | elapsed time per iteration (s): 1.76 | learning rate: 1.857E-04 | global batch size: 512 | lm loss: 1.571941E+00 | grad norm: 1.012 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.071 | TFLOPs: 43.69 | 31: iteration 9990/ 52452 | consumed samples: 5114880 | consumed tokens: 10475274240 | elapsed time per iteration (s): 1.82 | learning rate: 1.856E-04 | global batch size: 512 | lm loss: 1.557991E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.629 | TFLOPs: 42.12 | 0: [2022-12-25 16:04:18,068] [INFO] [logging.py:68:log_dist] [Rank 0] step=10000, skipped=0, lr=[0.00018561226620920443, 0.00018561226620920443, 0.00018561226620920443], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 10000/ 52452 | consumed samples: 5120000 | consumed tokens: 10485760000 | elapsed time per iteration (s): 1.79 | learning rate: 1.856E-04 | global batch size: 512 | lm loss: 1.581514E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.233 | TFLOPs: 42.96 | 0: steps: 10000 loss: 1.5768 iter time (s): 1.794 samples/sec: 285.342 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10000 | lm loss value: 4.246541E+00 | lm loss PPL: 6.986334E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10000 | lm loss value: 4.231962E+00 | lm loss PPL: 6.885217E+01 | 31: ------------------------------------------------------------------------------------------------------ 0: saving checkpoint at iteration 10000 to checkpoints_2b855brealtasky 0: [2022-12-25 16:05:31,958] [INFO] [logging.py:68:log_dist] [Rank 0] [Torch] Checkpoint global_step10000 is begin to save! 0: [2022-12-25 16:05:32,814] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_01-model_00-model_states.pt... 0: [2022-12-25 16:05:33,150] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_01-model_00-model_states.pt. 0: [2022-12-25 16:05:33,150] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_03-model_00-model_states.pt... 0: [2022-12-25 16:05:33,340] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_03-model_00-model_states.pt. 0: [2022-12-25 16:05:33,341] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_04-model_00-model_states.pt... 0: [2022-12-25 16:05:33,524] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_04-model_00-model_states.pt. 0: [2022-12-25 16:05:33,525] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_05-model_00-model_states.pt... 0: [2022-12-25 16:05:33,709] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_05-model_00-model_states.pt. 0: [2022-12-25 16:05:33,709] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_06-model_00-model_states.pt... 0: [2022-12-25 16:05:33,887] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_06-model_00-model_states.pt. 0: [2022-12-25 16:05:33,888] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_07-model_00-model_states.pt... 0: [2022-12-25 16:05:34,073] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_07-model_00-model_states.pt. 0: [2022-12-25 16:05:34,073] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_08-model_00-model_states.pt... 0: [2022-12-25 16:05:34,248] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_08-model_00-model_states.pt. 0: [2022-12-25 16:05:34,249] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_09-model_00-model_states.pt... 0: [2022-12-25 16:05:34,431] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_09-model_00-model_states.pt. 0: [2022-12-25 16:05:34,432] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_10-model_00-model_states.pt... 0: [2022-12-25 16:05:34,611] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_10-model_00-model_states.pt. 0: [2022-12-25 16:05:34,611] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_11-model_00-model_states.pt... 0: [2022-12-25 16:05:34,784] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_11-model_00-model_states.pt. 0: [2022-12-25 16:05:34,784] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_12-model_00-model_states.pt... 0: [2022-12-25 16:05:34,961] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_12-model_00-model_states.pt. 0: [2022-12-25 16:05:34,961] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_13-model_00-model_states.pt... 0: [2022-12-25 16:05:35,139] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_13-model_00-model_states.pt. 0: [2022-12-25 16:05:35,139] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_14-model_00-model_states.pt... 0: [2022-12-25 16:05:35,317] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_14-model_00-model_states.pt. 0: [2022-12-25 16:05:35,317] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_15-model_00-model_states.pt... 0: [2022-12-25 16:05:35,493] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_15-model_00-model_states.pt. 0: [2022-12-25 16:05:35,494] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_16-model_00-model_states.pt... 0: [2022-12-25 16:05:35,672] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_16-model_00-model_states.pt. 0: [2022-12-25 16:05:35,672] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_17-model_00-model_states.pt... 0: [2022-12-25 16:05:35,844] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_17-model_00-model_states.pt. 0: [2022-12-25 16:05:35,844] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_18-model_00-model_states.pt... 0: [2022-12-25 16:05:36,024] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_18-model_00-model_states.pt. 0: [2022-12-25 16:05:36,024] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_19-model_00-model_states.pt... 0: [2022-12-25 16:05:36,199] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_19-model_00-model_states.pt. 0: [2022-12-25 16:05:36,200] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_20-model_00-model_states.pt... 0: [2022-12-25 16:05:36,372] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_20-model_00-model_states.pt. 0: [2022-12-25 16:05:36,372] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_21-model_00-model_states.pt... 0: [2022-12-25 16:05:36,553] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_21-model_00-model_states.pt. 0: [2022-12-25 16:05:36,553] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_22-model_00-model_states.pt... 0: [2022-12-25 16:05:36,725] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_22-model_00-model_states.pt. 0: [2022-12-25 16:05:36,725] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_23-model_00-model_states.pt... 0: [2022-12-25 16:05:36,905] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_23-model_00-model_states.pt. 0: [2022-12-25 16:05:36,906] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_24-model_00-model_states.pt... 0: [2022-12-25 16:05:37,077] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_24-model_00-model_states.pt. 0: [2022-12-25 16:05:37,078] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_25-model_00-model_states.pt... 0: [2022-12-25 16:05:37,258] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_25-model_00-model_states.pt. 0: [2022-12-25 16:05:37,258] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_26-model_00-model_states.pt... 0: [2022-12-25 16:05:37,433] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_26-model_00-model_states.pt. 0: [2022-12-25 16:05:37,434] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_27-model_00-model_states.pt... 0: [2022-12-25 16:05:37,602] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_27-model_00-model_states.pt. 0: [2022-12-25 16:05:37,602] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_28-model_00-model_states.pt... 0: [2022-12-25 16:05:37,775] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_28-model_00-model_states.pt. 0: [2022-12-25 16:05:37,775] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_29-model_00-model_states.pt... 0: [2022-12-25 16:05:37,952] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_29-model_00-model_states.pt. 0: [2022-12-25 16:05:37,953] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_30-model_00-model_states.pt... 0: [2022-12-25 16:05:38,125] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_30-model_00-model_states.pt. 0: [2022-12-25 16:05:38,125] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_31-model_00-model_states.pt... 0: [2022-12-25 16:05:38,298] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_31-model_00-model_states.pt. 0: [2022-12-25 16:05:38,298] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_32-model_00-model_states.pt... 0: [2022-12-25 16:05:38,471] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_32-model_00-model_states.pt. 0: [2022-12-25 16:05:38,471] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_33-model_00-model_states.pt... 0: [2022-12-25 16:05:38,644] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_33-model_00-model_states.pt. 0: [2022-12-25 16:05:38,644] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_34-model_00-model_states.pt... 0: [2022-12-25 16:05:38,816] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_34-model_00-model_states.pt. 0: [2022-12-25 16:05:38,816] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_35-model_00-model_states.pt... 0: [2022-12-25 16:05:38,995] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_35-model_00-model_states.pt. 0: [2022-12-25 16:05:38,996] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_36-model_00-model_states.pt... 0: [2022-12-25 16:05:39,164] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_36-model_00-model_states.pt. 0: [2022-12-25 16:05:39,164] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/layer_38-model_00-model_states.pt... 0: [2022-12-25 16:05:39,169] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/layer_38-model_00-model_states.pt. 0: [2022-12-25 16:05:39,170] [INFO] [logging.py:68:log_dist] [Rank 0] Saving model checkpoint: checkpoints_2b855brealtasky/global_step10000/mp_rank_00_model_states.pt 0: [2022-12-25 16:05:39,170] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/mp_rank_00_model_states.pt... 0: [2022-12-25 16:05:39,177] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/mp_rank_00_model_states.pt. 0: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... 0: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... 16: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt... 27: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt... 0: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... 0: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... 25: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt... 25: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt... 31: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt... 31: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt... 31: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt... 31: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt... 31: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt... 29: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt... 29: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt... 11: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt... 11: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt... 11: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt... 11: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt... 10: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt... 10: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt... 10: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt... 23: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt... 23: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt... 23: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt... 23: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt... 23: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt... 2: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt... 2: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt... 2: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt... 2: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt... 27: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt... 22: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt... 22: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt... 22: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt... 1: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... 1: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... 1: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... 6: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt... 6: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt... 6: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt... 6: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt... 15: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt... 15: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt... 15: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt... 15: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt... 0: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... 16: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt... 16: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt... 14: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt... 14: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt... 14: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt... 12: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt... 3: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt... 3: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt... 3: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt... 3: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt... 3: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt... 18: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt... 18: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt... 8: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt... 8: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt... 8: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt... 8: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt... 5: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt... 5: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt... 5: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt... 5: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt... 5: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt... 4: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt... 7: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt... 7: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt... 7: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt... 7: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt... 7: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt... 9: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt... 9: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt... 9: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt... 9: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt... 13: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt... 13: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt... 13: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt... 24: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt... 24: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt... 24: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt... 17: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt... 17: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt... 17: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt... 17: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt... 17: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt... 28: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt... 28: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt... 28: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt... 28: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt... 28: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt... 21: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt... 21: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt... 21: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt... 21: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt... 21: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt... 30: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt... 30: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt... 30: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt... 30: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt... 30: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt... 26: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt... 26: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt... 19: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt... 19: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt... 25: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt... 25: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt... 31: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt... 29: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt... 29: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt... 11: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt... 11: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt... 11: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt... 10: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt... 23: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt... 23: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt... 20: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt... 20: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt... 20: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt... 20: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt... 20: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt... 2: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt... 2: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt... 27: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt... 27: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt... 27: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt... 22: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt... 22: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt... 1: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... 1: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... 6: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt... 6: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt... 15: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt... 15: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt... 0: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... 0: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... 16: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt... 14: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt... 14: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt... 14: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt... 14: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt... 12: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt... 12: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt... 3: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt... 18: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt... 18: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt... 18: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt... 8: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt... 8: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt... 5: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt... 5: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt... 4: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt... 4: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt... 4: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt... 7: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt... 9: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt... 9: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt... 9: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt... 9: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt... 13: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt... 13: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt... 13: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt... 13: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt... 13: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt... 24: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt... 24: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt... 24: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt... 17: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt... 28: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt... 28: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt... 21: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt... 21: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt... 30: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt... 26: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt... 26: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt... 19: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt... 19: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt... 19: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt... 25: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt... 31: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt... 31: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt... 29: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt... 29: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt... 11: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt... 10: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt... 10: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt... 10: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt... 23: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt... 20: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt... 20: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt... 2: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt... 2: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt... 27: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt... 27: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt... 1: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... 1: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... 6: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt... 6: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt... 15: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt... 0: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... 16: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt... 14: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt... 12: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt... 12: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt... 12: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt... 3: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt... 3: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt... 18: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt... 18: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt... 8: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt... 8: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt... 5: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt... 4: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt... 4: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt... 4: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt... 4: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt... 7: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt... 24: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt... 17: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt... 21: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt... 30: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt... 26: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt... 19: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt... 19: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt... 25: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt... 25: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt... 29: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt... 29: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt... 10: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt... 20: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt... 27: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt... 22: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt... 15: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt... 16: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt... 12: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt... 18: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt... 7: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt... 24: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt... 17: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt... 30: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt... 26: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt... 26: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt... 19: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt... 25: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt... 22: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt... 1: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... 16: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt... 12: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt... 28: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt... 26: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt... 22: [2022-12-25 16:05:39,261] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt... 16: [2022-12-25 16:05:39,260] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt... 0: [2022-12-25 16:05:39,429] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. 0: [2022-12-25 16:05:39,429] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt 0: [2022-12-25 16:05:39,429] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 0: [2022-12-25 16:05:39,430] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. 0: [2022-12-25 16:05:39,430] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt 0: [2022-12-25 16:05:39,430] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. 0: [2022-12-25 16:05:39,430] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 0: [2022-12-25 16:05:39,430] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. 0: [2022-12-25 16:05:39,430] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt 0: [2022-12-25 16:05:39,430] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 0: [2022-12-25 16:05:39,430] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt 0: [2022-12-25 16:05:39,430] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 0: [2022-12-25 16:05:39,430] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. 0: [2022-12-25 16:05:39,430] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt 0: [2022-12-25 16:05:39,430] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 15: [2022-12-25 16:05:39,449] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt. 15: [2022-12-25 16:05:39,449] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt. 15: [2022-12-25 16:05:39,449] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt. 15: [2022-12-25 16:05:39,450] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt 15: [2022-12-25 16:05:39,450] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt 15: [2022-12-25 16:05:39,450] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt 15: [2022-12-25 16:05:39,450] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 15: [2022-12-25 16:05:39,450] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 15: [2022-12-25 16:05:39,450] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 4: [2022-12-25 16:05:39,459] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt. 4: [2022-12-25 16:05:39,459] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt. 4: [2022-12-25 16:05:39,459] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt. 4: [2022-12-25 16:05:39,459] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt 4: [2022-12-25 16:05:39,459] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt 4: [2022-12-25 16:05:39,459] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 4: [2022-12-25 16:05:39,459] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt 4: [2022-12-25 16:05:39,459] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 4: [2022-12-25 16:05:39,460] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. 1: [2022-12-25 16:05:39,464] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt 1: [2022-12-25 16:05:39,464] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 1: [2022-12-25 16:05:39,464] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 1: [2022-12-25 16:05:39,464] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt 1: [2022-12-25 16:05:39,464] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt 1: [2022-12-25 16:05:39,464] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 1: [2022-12-25 16:05:39,464] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 1: [2022-12-25 16:05:39,465] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 15: [2022-12-25 16:05:39,454] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt. 15: [2022-12-25 16:05:39,454] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt 15: [2022-12-25 16:05:39,455] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 15: [2022-12-25 16:05:39,455] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt. 15: [2022-12-25 16:05:39,455] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt 15: [2022-12-25 16:05:39,455] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 15: [2022-12-25 16:05:39,456] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt. 15: [2022-12-25 16:05:39,456] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt 15: [2022-12-25 16:05:39,456] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 21: [2022-12-25 16:05:39,470] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt. 21: [2022-12-25 16:05:39,470] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt. 21: [2022-12-25 16:05:39,470] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt. 21: [2022-12-25 16:05:39,470] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt. 21: [2022-12-25 16:05:39,470] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt 21: [2022-12-25 16:05:39,470] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt 21: [2022-12-25 16:05:39,470] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt 21: [2022-12-25 16:05:39,470] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt 21: [2022-12-25 16:05:39,471] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 21: [2022-12-25 16:05:39,471] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 21: [2022-12-25 16:05:39,471] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 21: [2022-12-25 16:05:39,471] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 25: [2022-12-25 16:05:39,472] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt. 25: [2022-12-25 16:05:39,472] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt. 25: [2022-12-25 16:05:39,472] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt. 25: [2022-12-25 16:05:39,472] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt 25: [2022-12-25 16:05:39,472] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt 25: [2022-12-25 16:05:39,472] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt 25: [2022-12-25 16:05:39,472] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 25: [2022-12-25 16:05:39,472] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 25: [2022-12-25 16:05:39,472] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 13: [2022-12-25 16:05:39,474] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt. 13: [2022-12-25 16:05:39,474] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt. 13: [2022-12-25 16:05:39,474] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt. 13: [2022-12-25 16:05:39,474] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt 13: [2022-12-25 16:05:39,474] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt 13: [2022-12-25 16:05:39,474] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt 13: [2022-12-25 16:05:39,474] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 13: [2022-12-25 16:05:39,474] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 13: [2022-12-25 16:05:39,474] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 13: [2022-12-25 16:05:39,474] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt. 13: [2022-12-25 16:05:39,475] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt 13: [2022-12-25 16:05:39,475] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 13: [2022-12-25 16:05:39,475] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt. 13: [2022-12-25 16:05:39,475] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt 13: [2022-12-25 16:05:39,475] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 13: [2022-12-25 16:05:39,475] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt. 13: [2022-12-25 16:05:39,475] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt 13: [2022-12-25 16:05:39,476] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 13: [2022-12-25 16:05:39,476] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt. 13: [2022-12-25 16:05:39,476] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt 13: [2022-12-25 16:05:39,476] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 25: [2022-12-25 16:05:39,476] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt. 25: [2022-12-25 16:05:39,477] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt 25: [2022-12-25 16:05:39,477] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 25: [2022-12-25 16:05:39,477] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt. 25: [2022-12-25 16:05:39,477] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt 25: [2022-12-25 16:05:39,477] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 25: [2022-12-25 16:05:39,478] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt. 25: [2022-12-25 16:05:39,478] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt 25: [2022-12-25 16:05:39,478] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 4: [2022-12-25 16:05:39,484] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt. 4: [2022-12-25 16:05:39,484] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt 4: [2022-12-25 16:05:39,484] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 4: [2022-12-25 16:05:39,484] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt. 0: [2022-12-25 16:05:39,484] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. 4: [2022-12-25 16:05:39,484] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt 4: [2022-12-25 16:05:39,484] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt. 4: [2022-12-25 16:05:39,484] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 4: [2022-12-25 16:05:39,484] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt 4: [2022-12-25 16:05:39,484] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt. 4: [2022-12-25 16:05:39,485] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 0: [2022-12-25 16:05:39,484] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. 4: [2022-12-25 16:05:39,485] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt 4: [2022-12-25 16:05:39,485] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 0: [2022-12-25 16:05:39,485] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt 0: [2022-12-25 16:05:39,485] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 0: [2022-12-25 16:05:39,485] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. 0: [2022-12-25 16:05:39,485] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt 0: [2022-12-25 16:05:39,485] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt. 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt. 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt. 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt. 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt. 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt. 17: [2022-12-25 16:05:39,486] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt 17: [2022-12-25 16:05:39,486] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt 17: [2022-12-25 16:05:39,486] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt 17: [2022-12-25 16:05:39,486] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt 17: [2022-12-25 16:05:39,486] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt 17: [2022-12-25 16:05:39,486] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 17: [2022-12-25 16:05:39,486] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt. 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt. 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt. 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt. 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt. 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt. 3: [2022-12-25 16:05:39,482] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt 3: [2022-12-25 16:05:39,482] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt 3: [2022-12-25 16:05:39,482] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt 3: [2022-12-25 16:05:39,482] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt 3: [2022-12-25 16:05:39,482] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt 3: [2022-12-25 16:05:39,482] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 3: [2022-12-25 16:05:39,482] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 21: [2022-12-25 16:05:39,499] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt. 21: [2022-12-25 16:05:39,499] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt 21: [2022-12-25 16:05:39,499] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt. 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt. 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt. 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt. 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt. 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt. 28: [2022-12-25 16:05:39,500] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt 28: [2022-12-25 16:05:39,500] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt 28: [2022-12-25 16:05:39,500] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt 28: [2022-12-25 16:05:39,500] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt 28: [2022-12-25 16:05:39,500] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt 28: [2022-12-25 16:05:39,500] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 28: [2022-12-25 16:05:39,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 11: [2022-12-25 16:05:39,505] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt. 11: [2022-12-25 16:05:39,505] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt. 11: [2022-12-25 16:05:39,505] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt. 11: [2022-12-25 16:05:39,505] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt. 11: [2022-12-25 16:05:39,505] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt 11: [2022-12-25 16:05:39,505] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt 11: [2022-12-25 16:05:39,505] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt 11: [2022-12-25 16:05:39,505] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt 11: [2022-12-25 16:05:39,505] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 11: [2022-12-25 16:05:39,505] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 11: [2022-12-25 16:05:39,505] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 11: [2022-12-25 16:05:39,505] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 21: [2022-12-25 16:05:39,515] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt. 21: [2022-12-25 16:05:39,515] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt 21: [2022-12-25 16:05:39,515] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 11: [2022-12-25 16:05:39,520] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt. 11: [2022-12-25 16:05:39,520] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt. 11: [2022-12-25 16:05:39,520] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt 11: [2022-12-25 16:05:39,520] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt 11: [2022-12-25 16:05:39,521] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 11: [2022-12-25 16:05:39,521] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 5: [2022-12-25 16:05:39,494] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt. 5: [2022-12-25 16:05:39,494] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt. 5: [2022-12-25 16:05:39,494] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt. 5: [2022-12-25 16:05:39,495] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt 5: [2022-12-25 16:05:39,495] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 5: [2022-12-25 16:05:39,495] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt 5: [2022-12-25 16:05:39,495] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt 5: [2022-12-25 16:05:39,495] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 5: [2022-12-25 16:05:39,495] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 5: [2022-12-25 16:05:39,495] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt. 5: [2022-12-25 16:05:39,495] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt 5: [2022-12-25 16:05:39,495] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 5: [2022-12-25 16:05:39,495] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt. 5: [2022-12-25 16:05:39,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt 5: [2022-12-25 16:05:39,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 5: [2022-12-25 16:05:39,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt. 5: [2022-12-25 16:05:39,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt 5: [2022-12-25 16:05:39,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 5: [2022-12-25 16:05:39,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt. 5: [2022-12-25 16:05:39,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt 5: [2022-12-25 16:05:39,497] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 17: [2022-12-25 16:05:39,526] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt. 17: [2022-12-25 16:05:39,526] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt 17: [2022-12-25 16:05:39,526] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt. 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt. 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt. 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt. 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt. 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt. 2: [2022-12-25 16:05:39,537] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt 2: [2022-12-25 16:05:39,537] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt 2: [2022-12-25 16:05:39,537] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt 2: [2022-12-25 16:05:39,537] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt 2: [2022-12-25 16:05:39,537] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 2: [2022-12-25 16:05:39,537] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 2: [2022-12-25 16:05:39,537] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt. 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt. 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt. 6: [2022-12-25 16:05:39,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt 6: [2022-12-25 16:05:39,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt 6: [2022-12-25 16:05:39,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt. 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt. 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt. 6: [2022-12-25 16:05:39,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt 6: [2022-12-25 16:05:39,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 6: [2022-12-25 16:05:39,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 6: [2022-12-25 16:05:39,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 7: [2022-12-25 16:05:39,547] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt. 7: [2022-12-25 16:05:39,547] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt. 7: [2022-12-25 16:05:39,547] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt. 7: [2022-12-25 16:05:39,547] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt 7: [2022-12-25 16:05:39,547] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt 7: [2022-12-25 16:05:39,547] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt 7: [2022-12-25 16:05:39,547] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 7: [2022-12-25 16:05:39,547] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 7: [2022-12-25 16:05:39,547] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 7: [2022-12-25 16:05:39,549] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt. 7: [2022-12-25 16:05:39,549] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt. 7: [2022-12-25 16:05:39,549] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt. 7: [2022-12-25 16:05:39,549] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt. 7: [2022-12-25 16:05:39,549] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt 7: [2022-12-25 16:05:39,549] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt 7: [2022-12-25 16:05:39,549] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt 7: [2022-12-25 16:05:39,549] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt 7: [2022-12-25 16:05:39,549] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 7: [2022-12-25 16:05:39,549] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 7: [2022-12-25 16:05:39,549] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 7: [2022-12-25 16:05:39,549] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt. 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt. 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt. 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt. 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt. 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt. 18: [2022-12-25 16:05:39,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt 18: [2022-12-25 16:05:39,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt 18: [2022-12-25 16:05:39,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt 18: [2022-12-25 16:05:39,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt 18: [2022-12-25 16:05:39,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt 18: [2022-12-25 16:05:39,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 18: [2022-12-25 16:05:39,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt. 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt. 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt. 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt. 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt. 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt. 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt. 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt. 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt. 22: [2022-12-25 16:05:39,557] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt. 22: [2022-12-25 16:05:39,557] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt 22: [2022-12-25 16:05:39,557] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 22: [2022-12-25 16:05:39,557] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt 22: [2022-12-25 16:05:39,557] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 22: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 26: [2022-12-25 16:05:39,546] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt. 26: [2022-12-25 16:05:39,546] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt. 26: [2022-12-25 16:05:39,546] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt. 26: [2022-12-25 16:05:39,546] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt. 26: [2022-12-25 16:05:39,546] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt 26: [2022-12-25 16:05:39,546] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt 26: [2022-12-25 16:05:39,546] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt 26: [2022-12-25 16:05:39,546] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt 26: [2022-12-25 16:05:39,546] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 26: [2022-12-25 16:05:39,546] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 26: [2022-12-25 16:05:39,546] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt. 26: [2022-12-25 16:05:39,546] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 30: [2022-12-25 16:05:39,551] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt 26: [2022-12-25 16:05:39,556] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt. 30: [2022-12-25 16:05:39,551] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt 30: [2022-12-25 16:05:39,551] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt 30: [2022-12-25 16:05:39,551] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt 30: [2022-12-25 16:05:39,551] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt 30: [2022-12-25 16:05:39,551] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt 26: [2022-12-25 16:05:39,556] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 26: [2022-12-25 16:05:39,556] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt. 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 30: [2022-12-25 16:05:39,551] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 26: [2022-12-25 16:05:39,556] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 26: [2022-12-25 16:05:39,556] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt 26: [2022-12-25 16:05:39,556] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 26: [2022-12-25 16:05:39,556] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt. 26: [2022-12-25 16:05:39,557] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt 26: [2022-12-25 16:05:39,557] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 16: [2022-12-25 16:05:39,572] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt. 16: [2022-12-25 16:05:39,572] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt. 16: [2022-12-25 16:05:39,572] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt. 16: [2022-12-25 16:05:39,572] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt. 16: [2022-12-25 16:05:39,572] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt. 16: [2022-12-25 16:05:39,572] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt. 16: [2022-12-25 16:05:39,572] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt. 16: [2022-12-25 16:05:39,573] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt 16: [2022-12-25 16:05:39,573] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt 16: [2022-12-25 16:05:39,573] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt 16: [2022-12-25 16:05:39,573] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt 16: [2022-12-25 16:05:39,573] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt 16: [2022-12-25 16:05:39,573] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt 16: [2022-12-25 16:05:39,573] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt 16: [2022-12-25 16:05:39,573] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 16: [2022-12-25 16:05:39,573] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 16: [2022-12-25 16:05:39,573] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 16: [2022-12-25 16:05:39,573] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 16: [2022-12-25 16:05:39,573] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 16: [2022-12-25 16:05:39,573] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 16: [2022-12-25 16:05:39,573] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt. 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt. 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt. 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt. 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt. 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt. 20: [2022-12-25 16:05:39,583] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt 20: [2022-12-25 16:05:39,583] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt 20: [2022-12-25 16:05:39,583] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt 20: [2022-12-25 16:05:39,583] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt 20: [2022-12-25 16:05:39,583] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt 20: [2022-12-25 16:05:39,583] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 20: [2022-12-25 16:05:39,583] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 14: [2022-12-25 16:05:39,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt. 14: [2022-12-25 16:05:39,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt. 14: [2022-12-25 16:05:39,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt. 14: [2022-12-25 16:05:39,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt. 14: [2022-12-25 16:05:39,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt. 14: [2022-12-25 16:05:39,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt. 14: [2022-12-25 16:05:39,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt. 14: [2022-12-25 16:05:39,590] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt 14: [2022-12-25 16:05:39,590] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt 14: [2022-12-25 16:05:39,590] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt 14: [2022-12-25 16:05:39,590] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt 14: [2022-12-25 16:05:39,590] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt 14: [2022-12-25 16:05:39,590] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt 14: [2022-12-25 16:05:39,590] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt 14: [2022-12-25 16:05:39,590] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 14: [2022-12-25 16:05:39,590] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 14: [2022-12-25 16:05:39,590] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 14: [2022-12-25 16:05:39,590] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 14: [2022-12-25 16:05:39,590] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 14: [2022-12-25 16:05:39,590] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 14: [2022-12-25 16:05:39,590] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt. 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt. 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt. 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt. 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt. 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt. 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt. 27: [2022-12-25 16:05:39,597] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt 27: [2022-12-25 16:05:39,597] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt 27: [2022-12-25 16:05:39,597] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt 27: [2022-12-25 16:05:39,597] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt 27: [2022-12-25 16:05:39,597] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 27: [2022-12-25 16:05:39,597] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt 27: [2022-12-25 16:05:39,597] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 27: [2022-12-25 16:05:39,597] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 19: [2022-12-25 16:05:39,608] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt. 19: [2022-12-25 16:05:39,608] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt. 19: [2022-12-25 16:05:39,608] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt. 19: [2022-12-25 16:05:39,608] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt. 19: [2022-12-25 16:05:39,608] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt. 19: [2022-12-25 16:05:39,608] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt. 19: [2022-12-25 16:05:39,608] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt. 19: [2022-12-25 16:05:39,608] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt. 19: [2022-12-25 16:05:39,608] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt 19: [2022-12-25 16:05:39,608] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt 19: [2022-12-25 16:05:39,608] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt 19: [2022-12-25 16:05:39,608] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt 19: [2022-12-25 16:05:39,608] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt 19: [2022-12-25 16:05:39,608] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt 19: [2022-12-25 16:05:39,609] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 19: [2022-12-25 16:05:39,609] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt 19: [2022-12-25 16:05:39,609] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 19: [2022-12-25 16:05:39,609] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 19: [2022-12-25 16:05:39,609] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 19: [2022-12-25 16:05:39,609] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 19: [2022-12-25 16:05:39,609] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 19: [2022-12-25 16:05:39,609] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt 19: [2022-12-25 16:05:39,609] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 19: [2022-12-25 16:05:39,609] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt 23: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 23: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 23: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt 23: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt. 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt. 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt. 23: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt. 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt. 10: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt 10: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt 10: [2022-12-25 16:05:39,612] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt 23: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,612] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 11: [2022-12-25 16:05:39,609] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt. 11: [2022-12-25 16:05:39,610] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt 11: [2022-12-25 16:05:39,610] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 22: [2022-12-25 16:05:39,615] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt. 22: [2022-12-25 16:05:39,615] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt 22: [2022-12-25 16:05:39,616] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt. 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt. 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt. 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt. 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt. 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt. 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt. 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt. 24: [2022-12-25 16:05:39,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt 24: [2022-12-25 16:05:39,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt 24: [2022-12-25 16:05:39,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt 24: [2022-12-25 16:05:39,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt 24: [2022-12-25 16:05:39,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt 24: [2022-12-25 16:05:39,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt 24: [2022-12-25 16:05:39,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 24: [2022-12-25 16:05:39,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 24: [2022-12-25 16:05:39,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 12: [2022-12-25 16:05:39,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt. 12: [2022-12-25 16:05:39,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt. 12: [2022-12-25 16:05:39,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt. 12: [2022-12-25 16:05:39,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt. 12: [2022-12-25 16:05:39,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt. 12: [2022-12-25 16:05:39,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt. 12: [2022-12-25 16:05:39,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt. 12: [2022-12-25 16:05:39,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt. 12: [2022-12-25 16:05:39,623] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt 12: [2022-12-25 16:05:39,623] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt 12: [2022-12-25 16:05:39,623] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt 12: [2022-12-25 16:05:39,623] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt 12: [2022-12-25 16:05:39,623] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt 12: [2022-12-25 16:05:39,623] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt 12: [2022-12-25 16:05:39,623] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt 12: [2022-12-25 16:05:39,623] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 12: [2022-12-25 16:05:39,623] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 12: [2022-12-25 16:05:39,623] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 12: [2022-12-25 16:05:39,623] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt 12: [2022-12-25 16:05:39,623] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 12: [2022-12-25 16:05:39,623] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 12: [2022-12-25 16:05:39,623] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 12: [2022-12-25 16:05:39,623] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 12: [2022-12-25 16:05:39,623] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 28: [2022-12-25 16:05:39,624] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt. 28: [2022-12-25 16:05:39,624] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt 28: [2022-12-25 16:05:39,624] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt. 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt. 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt. 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt. 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt. 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt. 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt. 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt. 9: [2022-12-25 16:05:39,626] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt 9: [2022-12-25 16:05:39,626] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt 9: [2022-12-25 16:05:39,626] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt 9: [2022-12-25 16:05:39,626] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt 9: [2022-12-25 16:05:39,626] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt 9: [2022-12-25 16:05:39,626] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt 9: [2022-12-25 16:05:39,626] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt 9: [2022-12-25 16:05:39,626] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 9: [2022-12-25 16:05:39,626] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 25: [2022-12-25 16:05:39,628] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt. 25: [2022-12-25 16:05:39,628] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt 25: [2022-12-25 16:05:39,628] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 0: [2022-12-25 16:05:39,628] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt 0: [2022-12-25 16:05:39,628] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 3: [2022-12-25 16:05:39,631] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt. 3: [2022-12-25 16:05:39,631] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt 3: [2022-12-25 16:05:39,631] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 15: [2022-12-25 16:05:39,633] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt. 15: [2022-12-25 16:05:39,633] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt 15: [2022-12-25 16:05:39,633] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 21: [2022-12-25 16:05:39,636] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt. 21: [2022-12-25 16:05:39,636] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt 21: [2022-12-25 16:05:39,636] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 1: [2022-12-25 16:05:39,652] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. 1: [2022-12-25 16:05:39,652] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt 1: [2022-12-25 16:05:39,652] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 6: [2022-12-25 16:05:39,653] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt. 6: [2022-12-25 16:05:39,653] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt 6: [2022-12-25 16:05:39,653] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt. 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt. 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt. 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt. 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt. 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt. 31: [2022-12-25 16:05:39,658] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt 31: [2022-12-25 16:05:39,658] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt 31: [2022-12-25 16:05:39,658] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt 31: [2022-12-25 16:05:39,658] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt 31: [2022-12-25 16:05:39,658] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt 31: [2022-12-25 16:05:39,658] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt. 31: [2022-12-25 16:05:39,658] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt 31: [2022-12-25 16:05:39,658] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 30: [2022-12-25 16:05:39,660] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt. 30: [2022-12-25 16:05:39,660] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt 30: [2022-12-25 16:05:39,660] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 29: [2022-12-25 16:05:39,660] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt. 29: [2022-12-25 16:05:39,660] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt. 29: [2022-12-25 16:05:39,660] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt. 29: [2022-12-25 16:05:39,660] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt 29: [2022-12-25 16:05:39,660] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt 29: [2022-12-25 16:05:39,660] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt 29: [2022-12-25 16:05:39,660] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 29: [2022-12-25 16:05:39,660] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 29: [2022-12-25 16:05:39,660] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 29: [2022-12-25 16:05:39,666] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt. 29: [2022-12-25 16:05:39,666] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt 29: [2022-12-25 16:05:39,666] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 29: [2022-12-25 16:05:39,666] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt. 29: [2022-12-25 16:05:39,666] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt 29: [2022-12-25 16:05:39,666] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt. 29: [2022-12-25 16:05:39,666] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 29: [2022-12-25 16:05:39,666] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt. 29: [2022-12-25 16:05:39,666] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt 29: [2022-12-25 16:05:39,667] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 29: [2022-12-25 16:05:39,667] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt 29: [2022-12-25 16:05:39,667] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 20: [2022-12-25 16:05:39,672] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt. 20: [2022-12-25 16:05:39,672] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt 20: [2022-12-25 16:05:39,673] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,685] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt. 10: [2022-12-25 16:05:39,686] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt 10: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 13: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt. 13: [2022-12-25 16:05:39,686] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt 13: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt. 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt. 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt. 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt. 8: [2022-12-25 16:05:39,686] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt. 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt. 8: [2022-12-25 16:05:39,686] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt. 8: [2022-12-25 16:05:39,686] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt 8: [2022-12-25 16:05:39,686] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 8: [2022-12-25 16:05:39,686] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt 8: [2022-12-25 16:05:39,686] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 8: [2022-12-25 16:05:39,686] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 8: [2022-12-25 16:05:39,686] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 18: [2022-12-25 16:05:39,687] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt. 18: [2022-12-25 16:05:39,687] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt 18: [2022-12-25 16:05:39,687] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 25: [2022-12-25 16:05:39,687] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt. 25: [2022-12-25 16:05:39,687] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt 25: [2022-12-25 16:05:39,687] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 16: [2022-12-25 16:05:39,687] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt. 30: [2022-12-25 16:05:39,687] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt. 16: [2022-12-25 16:05:39,687] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt 30: [2022-12-25 16:05:39,687] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt 30: [2022-12-25 16:05:39,687] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 16: [2022-12-25 16:05:39,687] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 22: [2022-12-25 16:05:39,688] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt. 22: [2022-12-25 16:05:39,688] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt 22: [2022-12-25 16:05:39,688] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 17: [2022-12-25 16:05:39,689] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt. 17: [2022-12-25 16:05:39,689] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt 17: [2022-12-25 16:05:39,689] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 1: [2022-12-25 16:05:39,689] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. 1: [2022-12-25 16:05:39,689] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt 1: [2022-12-25 16:05:39,689] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 2: [2022-12-25 16:05:39,690] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt. 2: [2022-12-25 16:05:39,690] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt 2: [2022-12-25 16:05:39,690] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 15: [2022-12-25 16:05:39,690] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt. 15: [2022-12-25 16:05:39,691] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt 5: [2022-12-25 16:05:39,690] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt. 15: [2022-12-25 16:05:39,691] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 5: [2022-12-25 16:05:39,691] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt 5: [2022-12-25 16:05:39,691] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 28: [2022-12-25 16:05:39,691] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt. 28: [2022-12-25 16:05:39,691] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt 28: [2022-12-25 16:05:39,691] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 18: [2022-12-25 16:05:39,693] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt. 18: [2022-12-25 16:05:39,693] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt 18: [2022-12-25 16:05:39,693] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 31: [2022-12-25 16:05:39,693] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt. 7: [2022-12-25 16:05:39,693] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt. 7: [2022-12-25 16:05:39,694] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt 7: [2022-12-25 16:05:39,694] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 3: [2022-12-25 16:05:39,694] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt. 3: [2022-12-25 16:05:39,694] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt 3: [2022-12-25 16:05:39,694] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 2: [2022-12-25 16:05:39,694] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt. 8: [2022-12-25 16:05:39,694] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt. 2: [2022-12-25 16:05:39,694] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt 2: [2022-12-25 16:05:39,694] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 8: [2022-12-25 16:05:39,694] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt 8: [2022-12-25 16:05:39,694] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 31: [2022-12-25 16:05:39,693] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt 31: [2022-12-25 16:05:39,693] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 29: [2022-12-25 16:05:39,695] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt. 29: [2022-12-25 16:05:39,696] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt 29: [2022-12-25 16:05:39,696] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 20: [2022-12-25 16:05:39,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt. 20: [2022-12-25 16:05:39,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt 20: [2022-12-25 16:05:39,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 10: [2022-12-25 16:05:39,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt. 10: [2022-12-25 16:05:39,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt 10: [2022-12-25 16:05:39,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 11: [2022-12-25 16:05:39,698] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt. 11: [2022-12-25 16:05:39,698] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt 11: [2022-12-25 16:05:39,698] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 26: [2022-12-25 16:05:39,699] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt. 26: [2022-12-25 16:05:39,699] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt 26: [2022-12-25 16:05:39,699] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 21: [2022-12-25 16:05:39,702] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt. 21: [2022-12-25 16:05:39,702] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt 21: [2022-12-25 16:05:39,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 14: [2022-12-25 16:05:39,705] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt. 14: [2022-12-25 16:05:39,705] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt 14: [2022-12-25 16:05:39,705] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 4: [2022-12-25 16:05:39,706] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt. 4: [2022-12-25 16:05:39,706] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt 4: [2022-12-25 16:05:39,706] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 6: [2022-12-25 16:05:39,706] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt. 6: [2022-12-25 16:05:39,706] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt 6: [2022-12-25 16:05:39,706] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 22: [2022-12-25 16:05:39,707] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt. 22: [2022-12-25 16:05:39,707] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt 22: [2022-12-25 16:05:39,707] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 27: [2022-12-25 16:05:39,728] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt. 27: [2022-12-25 16:05:39,729] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step10000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt 27: [2022-12-25 16:05:39,729] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step10000 is ready now! 0: successfully saved checkpoint at iteration 10000 to checkpoints_2b855brealtasky 31: time (ms) | save-checkpoint: 7865.94 31: iteration 10010/ 52452 | consumed samples: 5125120 | consumed tokens: 10496245760 | elapsed time per iteration (s): 9.87 | learning rate: 1.856E-04 | global batch size: 512 | lm loss: 1.578129E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 51.882 | TFLOPs: 7.79 | 31: iteration 10020/ 52452 | consumed samples: 5130240 | consumed tokens: 10506731520 | elapsed time per iteration (s): 1.74 | learning rate: 1.856E-04 | global batch size: 512 | lm loss: 1.580219E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.945 | TFLOPs: 44.12 | 31: iteration 10030/ 52452 | consumed samples: 5135360 | consumed tokens: 10517217280 | elapsed time per iteration (s): 1.73 | learning rate: 1.855E-04 | global batch size: 512 | lm loss: 1.583683E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.958 | TFLOPs: 44.42 | 31: iteration 10040/ 52452 | consumed samples: 5140480 | consumed tokens: 10527703040 | elapsed time per iteration (s): 1.72 | learning rate: 1.855E-04 | global batch size: 512 | lm loss: 1.569477E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.078 | TFLOPs: 44.59 | 31: iteration 10050/ 52452 | consumed samples: 5145600 | consumed tokens: 10538188800 | elapsed time per iteration (s): 1.80 | learning rate: 1.855E-04 | global batch size: 512 | lm loss: 1.578111E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.721 | TFLOPs: 42.58 | 31: iteration 10060/ 52452 | consumed samples: 5150720 | consumed tokens: 10548674560 | elapsed time per iteration (s): 1.72 | learning rate: 1.854E-04 | global batch size: 512 | lm loss: 1.577003E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.394 | TFLOPs: 44.64 | 31: iteration 10070/ 52452 | consumed samples: 5155840 | consumed tokens: 10559160320 | elapsed time per iteration (s): 1.74 | learning rate: 1.854E-04 | global batch size: 512 | lm loss: 1.568890E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.631 | TFLOPs: 44.22 | 31: iteration 10080/ 52452 | consumed samples: 5160960 | consumed tokens: 10569646080 | elapsed time per iteration (s): 1.75 | learning rate: 1.854E-04 | global batch size: 512 | lm loss: 1.568625E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.928 | TFLOPs: 43.82 | 31: iteration 10090/ 52452 | consumed samples: 5166080 | consumed tokens: 10580131840 | elapsed time per iteration (s): 1.74 | learning rate: 1.853E-04 | global batch size: 512 | lm loss: 1.580320E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.482 | TFLOPs: 44.20 | 31: iteration 10100/ 52452 | consumed samples: 5171200 | consumed tokens: 10590617600 | elapsed time per iteration (s): 1.72 | learning rate: 1.853E-04 | global batch size: 512 | lm loss: 1.577904E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.258 | TFLOPs: 44.62 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10100 | lm loss value: 4.253605E+00 | lm loss PPL: 7.035862E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10100 | lm loss value: 4.264658E+00 | lm loss PPL: 7.114058E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 10110/ 52452 | consumed samples: 5176320 | consumed tokens: 10601103360 | elapsed time per iteration (s): 9.14 | learning rate: 1.853E-04 | global batch size: 512 | lm loss: 1.573345E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.041 | TFLOPs: 8.41 | 31: iteration 10120/ 52452 | consumed samples: 5181440 | consumed tokens: 10611589120 | elapsed time per iteration (s): 1.75 | learning rate: 1.853E-04 | global batch size: 512 | lm loss: 1.561880E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.640 | TFLOPs: 43.92 | 31: iteration 10130/ 52452 | consumed samples: 5186560 | consumed tokens: 10622074880 | elapsed time per iteration (s): 1.76 | learning rate: 1.852E-04 | global batch size: 512 | lm loss: 1.583967E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.688 | TFLOPs: 43.78 | 31: iteration 10140/ 52452 | consumed samples: 5191680 | consumed tokens: 10632560640 | elapsed time per iteration (s): 1.78 | learning rate: 1.852E-04 | global batch size: 512 | lm loss: 1.571173E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.216 | TFLOPs: 43.26 | 31: iteration 10150/ 52452 | consumed samples: 5196800 | consumed tokens: 10643046400 | elapsed time per iteration (s): 1.79 | learning rate: 1.852E-04 | global batch size: 512 | lm loss: 1.560880E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.246 | TFLOPs: 42.81 | 31: iteration 10160/ 52452 | consumed samples: 5201920 | consumed tokens: 10653532160 | elapsed time per iteration (s): 1.73 | learning rate: 1.851E-04 | global batch size: 512 | lm loss: 1.579787E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.173 | TFLOPs: 44.45 | 31: iteration 10170/ 52452 | consumed samples: 5207040 | consumed tokens: 10664017920 | elapsed time per iteration (s): 1.78 | learning rate: 1.851E-04 | global batch size: 512 | lm loss: 1.564269E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.081 | TFLOPs: 43.24 | 31: iteration 10180/ 52452 | consumed samples: 5212160 | consumed tokens: 10674503680 | elapsed time per iteration (s): 1.81 | learning rate: 1.851E-04 | global batch size: 512 | lm loss: 1.566649E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.611 | TFLOPs: 42.57 | 31: iteration 10190/ 52452 | consumed samples: 5217280 | consumed tokens: 10684989440 | elapsed time per iteration (s): 1.72 | learning rate: 1.850E-04 | global batch size: 512 | lm loss: 1.565385E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.836 | TFLOPs: 44.55 | 31: iteration 10200/ 52452 | consumed samples: 5222400 | consumed tokens: 10695475200 | elapsed time per iteration (s): 1.76 | learning rate: 1.850E-04 | global batch size: 512 | lm loss: 1.574117E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.312 | TFLOPs: 43.57 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10200 | lm loss value: 4.158999E+00 | lm loss PPL: 6.400745E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10200 | lm loss value: 4.162475E+00 | lm loss PPL: 6.423027E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 10210/ 52452 | consumed samples: 5227520 | consumed tokens: 10705960960 | elapsed time per iteration (s): 9.15 | learning rate: 1.850E-04 | global batch size: 512 | lm loss: 1.559307E+00 | grad norm: 0.206 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.960 | TFLOPs: 8.40 | 31: iteration 10220/ 52452 | consumed samples: 5232640 | consumed tokens: 10716446720 | elapsed time per iteration (s): 1.75 | learning rate: 1.850E-04 | global batch size: 512 | lm loss: 1.554784E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.822 | TFLOPs: 43.80 | 31: iteration 10230/ 52452 | consumed samples: 5237760 | consumed tokens: 10726932480 | elapsed time per iteration (s): 1.74 | learning rate: 1.849E-04 | global batch size: 512 | lm loss: 1.558546E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.829 | TFLOPs: 44.25 | 31: iteration 10240/ 52452 | consumed samples: 5242880 | consumed tokens: 10737418240 | elapsed time per iteration (s): 1.74 | learning rate: 1.849E-04 | global batch size: 512 | lm loss: 1.570164E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.777 | TFLOPs: 44.09 | 31: iteration 10250/ 52452 | consumed samples: 5248000 | consumed tokens: 10747904000 | elapsed time per iteration (s): 1.75 | learning rate: 1.849E-04 | global batch size: 512 | lm loss: 1.559528E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.073 | TFLOPs: 43.84 | 31: iteration 10260/ 52452 | consumed samples: 5253120 | consumed tokens: 10758389760 | elapsed time per iteration (s): 1.72 | learning rate: 1.848E-04 | global batch size: 512 | lm loss: 1.565816E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.196 | TFLOPs: 44.61 | 31: iteration 10270/ 52452 | consumed samples: 5258240 | consumed tokens: 10768875520 | elapsed time per iteration (s): 1.78 | learning rate: 1.848E-04 | global batch size: 512 | lm loss: 1.561027E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.341 | TFLOPs: 43.13 | 31: iteration 10280/ 52452 | consumed samples: 5263360 | consumed tokens: 10779361280 | elapsed time per iteration (s): 1.80 | learning rate: 1.848E-04 | global batch size: 512 | lm loss: 1.557283E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.972 | TFLOPs: 42.77 | 31: iteration 10290/ 52452 | consumed samples: 5268480 | consumed tokens: 10789847040 | elapsed time per iteration (s): 1.74 | learning rate: 1.847E-04 | global batch size: 512 | lm loss: 1.568390E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.766 | TFLOPs: 44.24 | 31: iteration 10300/ 52452 | consumed samples: 5273600 | consumed tokens: 10800332800 | elapsed time per iteration (s): 1.73 | learning rate: 1.847E-04 | global batch size: 512 | lm loss: 1.564304E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.859 | TFLOPs: 44.41 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10300 | lm loss value: 3.967503E+00 | lm loss PPL: 5.285240E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10300 | lm loss value: 4.015564E+00 | lm loss PPL: 5.545456E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 10310/ 52452 | consumed samples: 5278720 | consumed tokens: 10810818560 | elapsed time per iteration (s): 9.19 | learning rate: 1.847E-04 | global batch size: 512 | lm loss: 1.548211E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.730 | TFLOPs: 8.36 | 31: iteration 10320/ 52452 | consumed samples: 5283840 | consumed tokens: 10821304320 | elapsed time per iteration (s): 1.79 | learning rate: 1.847E-04 | global batch size: 512 | lm loss: 1.564247E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.695 | TFLOPs: 42.88 | 31: iteration 10330/ 52452 | consumed samples: 5288960 | consumed tokens: 10831790080 | elapsed time per iteration (s): 1.77 | learning rate: 1.846E-04 | global batch size: 512 | lm loss: 1.567467E+00 | grad norm: 0.227 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.845 | TFLOPs: 43.50 | 31: iteration 10340/ 52452 | consumed samples: 5294080 | consumed tokens: 10842275840 | elapsed time per iteration (s): 1.78 | learning rate: 1.846E-04 | global batch size: 512 | lm loss: 1.577948E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.050 | TFLOPs: 43.23 | 31: iteration 10350/ 52452 | consumed samples: 5299200 | consumed tokens: 10852761600 | elapsed time per iteration (s): 1.74 | learning rate: 1.846E-04 | global batch size: 512 | lm loss: 1.571877E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.433 | TFLOPs: 44.19 | 31: iteration 10360/ 52452 | consumed samples: 5304320 | consumed tokens: 10863247360 | elapsed time per iteration (s): 1.77 | learning rate: 1.845E-04 | global batch size: 512 | lm loss: 1.558771E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.939 | TFLOPs: 43.37 | 31: iteration 10370/ 52452 | consumed samples: 5309440 | consumed tokens: 10873733120 | elapsed time per iteration (s): 1.78 | learning rate: 1.845E-04 | global batch size: 512 | lm loss: 1.567828E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.024 | TFLOPs: 43.08 | 31: iteration 10380/ 52452 | consumed samples: 5314560 | consumed tokens: 10884218880 | elapsed time per iteration (s): 1.85 | learning rate: 1.845E-04 | global batch size: 512 | lm loss: 1.560984E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.362 | TFLOPs: 41.48 | 31: iteration 10390/ 52452 | consumed samples: 5319680 | consumed tokens: 10894704640 | elapsed time per iteration (s): 1.74 | learning rate: 1.844E-04 | global batch size: 512 | lm loss: 1.568132E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.044 | TFLOPs: 44.13 | 31: iteration 10400/ 52452 | consumed samples: 5324800 | consumed tokens: 10905190400 | elapsed time per iteration (s): 1.81 | learning rate: 1.844E-04 | global batch size: 512 | lm loss: 1.563019E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.840 | TFLOPs: 42.45 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10400 | lm loss value: 3.987554E+00 | lm loss PPL: 5.392281E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10400 | lm loss value: 4.059594E+00 | lm loss PPL: 5.795079E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 10410/ 52452 | consumed samples: 5329920 | consumed tokens: 10915676160 | elapsed time per iteration (s): 9.16 | learning rate: 1.844E-04 | global batch size: 512 | lm loss: 1.560621E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.923 | TFLOPs: 8.39 | 31: iteration 10420/ 52452 | consumed samples: 5335040 | consumed tokens: 10926161920 | elapsed time per iteration (s): 1.73 | learning rate: 1.843E-04 | global batch size: 512 | lm loss: 1.578620E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.622 | TFLOPs: 44.37 | 31: iteration 10430/ 52452 | consumed samples: 5340160 | consumed tokens: 10936647680 | elapsed time per iteration (s): 1.76 | learning rate: 1.843E-04 | global batch size: 512 | lm loss: 1.565841E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.413 | TFLOPs: 43.74 | 31: iteration 10440/ 52452 | consumed samples: 5345280 | consumed tokens: 10947133440 | elapsed time per iteration (s): 1.74 | learning rate: 1.843E-04 | global batch size: 512 | lm loss: 1.567519E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.501 | TFLOPs: 44.05 | 31: iteration 10450/ 52452 | consumed samples: 5350400 | consumed tokens: 10957619200 | elapsed time per iteration (s): 1.72 | learning rate: 1.843E-04 | global batch size: 512 | lm loss: 1.569333E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.105 | TFLOPs: 44.59 | 31: iteration 10460/ 52452 | consumed samples: 5355520 | consumed tokens: 10968104960 | elapsed time per iteration (s): 1.74 | learning rate: 1.842E-04 | global batch size: 512 | lm loss: 1.556275E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.383 | TFLOPs: 44.19 | 31: iteration 10470/ 52452 | consumed samples: 5360640 | consumed tokens: 10978590720 | elapsed time per iteration (s): 1.79 | learning rate: 1.842E-04 | global batch size: 512 | lm loss: 1.562517E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.802 | TFLOPs: 42.90 | 31: iteration 10480/ 52452 | consumed samples: 5365760 | consumed tokens: 10989076480 | elapsed time per iteration (s): 1.73 | learning rate: 1.842E-04 | global batch size: 512 | lm loss: 1.559051E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.142 | TFLOPs: 44.30 | 31: iteration 10490/ 52452 | consumed samples: 5370880 | consumed tokens: 10999562240 | elapsed time per iteration (s): 1.76 | learning rate: 1.841E-04 | global batch size: 512 | lm loss: 1.569016E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.916 | TFLOPs: 43.66 | 31: iteration 10500/ 52452 | consumed samples: 5376000 | consumed tokens: 11010048000 | elapsed time per iteration (s): 1.74 | learning rate: 1.841E-04 | global batch size: 512 | lm loss: 1.562648E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.352 | TFLOPs: 44.18 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10500 | lm loss value: 4.266769E+00 | lm loss PPL: 7.129095E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10500 | lm loss value: 4.274693E+00 | lm loss PPL: 7.185808E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 10510/ 52452 | consumed samples: 5381120 | consumed tokens: 11020533760 | elapsed time per iteration (s): 9.17 | learning rate: 1.841E-04 | global batch size: 512 | lm loss: 1.552425E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.862 | TFLOPs: 8.38 | 31: iteration 10520/ 52452 | consumed samples: 5386240 | consumed tokens: 11031019520 | elapsed time per iteration (s): 1.75 | learning rate: 1.840E-04 | global batch size: 512 | lm loss: 1.549240E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.894 | TFLOPs: 43.96 | 31: iteration 10530/ 52452 | consumed samples: 5391360 | consumed tokens: 11041505280 | elapsed time per iteration (s): 1.80 | learning rate: 1.840E-04 | global batch size: 512 | lm loss: 1.568592E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.860 | TFLOPs: 42.61 | 31: iteration 10540/ 52452 | consumed samples: 5396480 | consumed tokens: 11051991040 | elapsed time per iteration (s): 1.76 | learning rate: 1.840E-04 | global batch size: 512 | lm loss: 1.556851E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.532 | TFLOPs: 43.61 | 31: iteration 10550/ 52452 | consumed samples: 5401600 | consumed tokens: 11062476800 | elapsed time per iteration (s): 1.75 | learning rate: 1.839E-04 | global batch size: 512 | lm loss: 1.568187E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.963 | TFLOPs: 43.97 | 31: iteration 10560/ 52452 | consumed samples: 5406720 | consumed tokens: 11072962560 | elapsed time per iteration (s): 1.77 | learning rate: 1.839E-04 | global batch size: 512 | lm loss: 1.557046E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.429 | TFLOPs: 43.44 | 31: iteration 10570/ 52452 | consumed samples: 5411840 | consumed tokens: 11083448320 | elapsed time per iteration (s): 1.85 | learning rate: 1.839E-04 | global batch size: 512 | lm loss: 1.553626E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.401 | TFLOPs: 41.49 | 31: iteration 10580/ 52452 | consumed samples: 5416960 | consumed tokens: 11093934080 | elapsed time per iteration (s): 1.76 | learning rate: 1.839E-04 | global batch size: 512 | lm loss: 1.566025E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.744 | TFLOPs: 43.64 | 31: iteration 10590/ 52452 | consumed samples: 5422080 | consumed tokens: 11104419840 | elapsed time per iteration (s): 1.81 | learning rate: 1.838E-04 | global batch size: 512 | lm loss: 1.544063E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.156 | TFLOPs: 42.35 | 31: iteration 10600/ 52452 | consumed samples: 5427200 | consumed tokens: 11114905600 | elapsed time per iteration (s): 1.74 | learning rate: 1.838E-04 | global batch size: 512 | lm loss: 1.568932E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.945 | TFLOPs: 44.27 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10600 | lm loss value: 4.137158E+00 | lm loss PPL: 6.262458E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10600 | lm loss value: 4.150613E+00 | lm loss PPL: 6.347292E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 10610/ 52452 | consumed samples: 5432320 | consumed tokens: 11125391360 | elapsed time per iteration (s): 9.13 | learning rate: 1.838E-04 | global batch size: 512 | lm loss: 1.547341E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.092 | TFLOPs: 8.42 | 31: iteration 10620/ 52452 | consumed samples: 5437440 | consumed tokens: 11135877120 | elapsed time per iteration (s): 1.77 | learning rate: 1.837E-04 | global batch size: 512 | lm loss: 1.564036E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.474 | TFLOPs: 43.30 | 31: iteration 10630/ 52452 | consumed samples: 5442560 | consumed tokens: 11146362880 | elapsed time per iteration (s): 1.80 | learning rate: 1.837E-04 | global batch size: 512 | lm loss: 1.564347E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.541 | TFLOPs: 42.71 | 31: iteration 10640/ 52452 | consumed samples: 5447680 | consumed tokens: 11156848640 | elapsed time per iteration (s): 1.75 | learning rate: 1.837E-04 | global batch size: 512 | lm loss: 1.573969E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.945 | TFLOPs: 43.82 | 31: iteration 10650/ 52452 | consumed samples: 5452800 | consumed tokens: 11167334400 | elapsed time per iteration (s): 1.78 | learning rate: 1.836E-04 | global batch size: 512 | lm loss: 1.557993E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.586 | TFLOPs: 43.17 | 31: iteration 10660/ 52452 | consumed samples: 5457920 | consumed tokens: 11177820160 | elapsed time per iteration (s): 1.77 | learning rate: 1.836E-04 | global batch size: 512 | lm loss: 1.560673E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.255 | TFLOPs: 43.42 | 31: iteration 10670/ 52452 | consumed samples: 5463040 | consumed tokens: 11188305920 | elapsed time per iteration (s): 1.75 | learning rate: 1.836E-04 | global batch size: 512 | lm loss: 1.563670E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.079 | TFLOPs: 43.99 | 31: iteration 10680/ 52452 | consumed samples: 5468160 | consumed tokens: 11198791680 | elapsed time per iteration (s): 1.85 | learning rate: 1.835E-04 | global batch size: 512 | lm loss: 1.568504E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.710 | TFLOPs: 41.53 | 31: iteration 10690/ 52452 | consumed samples: 5473280 | consumed tokens: 11209277440 | elapsed time per iteration (s): 1.80 | learning rate: 1.835E-04 | global batch size: 512 | lm loss: 1.575546E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.137 | TFLOPs: 42.65 | 31: iteration 10700/ 52452 | consumed samples: 5478400 | consumed tokens: 11219763200 | elapsed time per iteration (s): 1.75 | learning rate: 1.835E-04 | global batch size: 512 | lm loss: 1.549814E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.910 | TFLOPs: 43.96 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10700 | lm loss value: 4.306633E+00 | lm loss PPL: 7.419027E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10700 | lm loss value: 4.285619E+00 | lm loss PPL: 7.264749E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 10710/ 52452 | consumed samples: 5483520 | consumed tokens: 11230248960 | elapsed time per iteration (s): 9.23 | learning rate: 1.834E-04 | global batch size: 512 | lm loss: 1.557939E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.493 | TFLOPs: 8.33 | 31: iteration 10720/ 52452 | consumed samples: 5488640 | consumed tokens: 11240734720 | elapsed time per iteration (s): 1.75 | learning rate: 1.834E-04 | global batch size: 512 | lm loss: 1.549837E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.379 | TFLOPs: 44.03 | 31: iteration 10730/ 52452 | consumed samples: 5493760 | consumed tokens: 11251220480 | elapsed time per iteration (s): 1.73 | learning rate: 1.834E-04 | global batch size: 512 | lm loss: 1.554006E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.791 | TFLOPs: 44.40 | 31: iteration 10740/ 52452 | consumed samples: 5498880 | consumed tokens: 11261706240 | elapsed time per iteration (s): 1.80 | learning rate: 1.834E-04 | global batch size: 512 | lm loss: 1.554894E+00 | grad norm: 0.213 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.880 | TFLOPs: 42.61 | 31: iteration 10750/ 52452 | consumed samples: 5504000 | consumed tokens: 11272192000 | elapsed time per iteration (s): 1.77 | learning rate: 1.833E-04 | global batch size: 512 | lm loss: 1.551375E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.521 | TFLOPs: 43.46 | 31: iteration 10760/ 52452 | consumed samples: 5509120 | consumed tokens: 11282677760 | elapsed time per iteration (s): 1.77 | learning rate: 1.833E-04 | global batch size: 512 | lm loss: 1.542996E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.710 | TFLOPs: 43.48 | 31: iteration 10770/ 52452 | consumed samples: 5514240 | consumed tokens: 11293163520 | elapsed time per iteration (s): 1.79 | learning rate: 1.833E-04 | global batch size: 512 | lm loss: 1.543262E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.651 | TFLOPs: 43.02 | 31: iteration 10780/ 52452 | consumed samples: 5519360 | consumed tokens: 11303649280 | elapsed time per iteration (s): 1.77 | learning rate: 1.832E-04 | global batch size: 512 | lm loss: 1.543139E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.808 | TFLOPs: 43.50 | 31: iteration 10790/ 52452 | consumed samples: 5524480 | consumed tokens: 11314135040 | elapsed time per iteration (s): 1.74 | learning rate: 1.832E-04 | global batch size: 512 | lm loss: 1.550226E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.467 | TFLOPs: 44.20 | 31: iteration 10800/ 52452 | consumed samples: 5529600 | consumed tokens: 11324620800 | elapsed time per iteration (s): 1.77 | learning rate: 1.832E-04 | global batch size: 512 | lm loss: 1.559549E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.819 | TFLOPs: 43.50 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10800 | lm loss value: 4.025786E+00 | lm loss PPL: 5.602432E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10800 | lm loss value: 4.067913E+00 | lm loss PPL: 5.843486E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 10810/ 52452 | consumed samples: 5534720 | consumed tokens: 11335106560 | elapsed time per iteration (s): 9.17 | learning rate: 1.831E-04 | global batch size: 512 | lm loss: 1.553927E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.815 | TFLOPs: 8.38 | 31: iteration 10820/ 52452 | consumed samples: 5539840 | consumed tokens: 11345592320 | elapsed time per iteration (s): 1.77 | learning rate: 1.831E-04 | global batch size: 512 | lm loss: 1.534750E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.987 | TFLOPs: 43.53 | 31: iteration 10830/ 52452 | consumed samples: 5544960 | consumed tokens: 11356078080 | elapsed time per iteration (s): 1.77 | learning rate: 1.831E-04 | global batch size: 512 | lm loss: 1.551419E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.600 | TFLOPs: 43.32 | 31: iteration 10840/ 52452 | consumed samples: 5550080 | consumed tokens: 11366563840 | elapsed time per iteration (s): 1.75 | learning rate: 1.830E-04 | global batch size: 512 | lm loss: 1.552326E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.745 | TFLOPs: 43.79 | 31: iteration 10850/ 52452 | consumed samples: 5555200 | consumed tokens: 11377049600 | elapsed time per iteration (s): 1.80 | learning rate: 1.830E-04 | global batch size: 512 | lm loss: 1.549252E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.224 | TFLOPs: 42.81 | 31: iteration 10860/ 52452 | consumed samples: 5560320 | consumed tokens: 11387535360 | elapsed time per iteration (s): 1.84 | learning rate: 1.830E-04 | global batch size: 512 | lm loss: 1.566967E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.906 | TFLOPs: 41.86 | 31: iteration 10870/ 52452 | consumed samples: 5565440 | consumed tokens: 11398021120 | elapsed time per iteration (s): 1.76 | learning rate: 1.829E-04 | global batch size: 512 | lm loss: 1.560746E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.285 | TFLOPs: 43.72 | 31: iteration 10880/ 52452 | consumed samples: 5570560 | consumed tokens: 11408506880 | elapsed time per iteration (s): 1.80 | learning rate: 1.829E-04 | global batch size: 512 | lm loss: 1.563474E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.012 | TFLOPs: 42.63 | 31: iteration 10890/ 52452 | consumed samples: 5575680 | consumed tokens: 11418992640 | elapsed time per iteration (s): 1.80 | learning rate: 1.829E-04 | global batch size: 512 | lm loss: 1.548703E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.224 | TFLOPs: 42.66 | 31: iteration 10900/ 52452 | consumed samples: 5580800 | consumed tokens: 11429478400 | elapsed time per iteration (s): 1.73 | learning rate: 1.828E-04 | global batch size: 512 | lm loss: 1.549279E+00 | grad norm: 0.199 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.211 | TFLOPs: 44.46 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 10900 | lm loss value: 4.092970E+00 | lm loss PPL: 5.991761E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 10900 | lm loss value: 4.093364E+00 | lm loss PPL: 5.994118E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 10910/ 52452 | consumed samples: 5585920 | consumed tokens: 11439964160 | elapsed time per iteration (s): 9.11 | learning rate: 1.828E-04 | global batch size: 512 | lm loss: 1.538445E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.183 | TFLOPs: 8.43 | 31: iteration 10920/ 52452 | consumed samples: 5591040 | consumed tokens: 11450449920 | elapsed time per iteration (s): 1.82 | learning rate: 1.828E-04 | global batch size: 512 | lm loss: 1.554471E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.181 | TFLOPs: 42.20 | 31: iteration 10930/ 52452 | consumed samples: 5596160 | consumed tokens: 11460935680 | elapsed time per iteration (s): 1.75 | learning rate: 1.827E-04 | global batch size: 512 | lm loss: 1.540247E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.145 | TFLOPs: 43.85 | 31: iteration 10940/ 52452 | consumed samples: 5601280 | consumed tokens: 11471421440 | elapsed time per iteration (s): 1.84 | learning rate: 1.827E-04 | global batch size: 512 | lm loss: 1.548825E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.083 | TFLOPs: 41.74 | 31: iteration 10950/ 52452 | consumed samples: 5606400 | consumed tokens: 11481907200 | elapsed time per iteration (s): 1.77 | learning rate: 1.827E-04 | global batch size: 512 | lm loss: 1.560018E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.615 | TFLOPs: 43.32 | 31: iteration 10960/ 52452 | consumed samples: 5611520 | consumed tokens: 11492392960 | elapsed time per iteration (s): 1.78 | learning rate: 1.827E-04 | global batch size: 512 | lm loss: 1.565165E+00 | grad norm: 0.198 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.870 | TFLOPs: 43.06 | 31: iteration 10970/ 52452 | consumed samples: 5616640 | consumed tokens: 11502878720 | elapsed time per iteration (s): 1.85 | learning rate: 1.826E-04 | global batch size: 512 | lm loss: 1.547416E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.760 | TFLOPs: 41.54 | 31: iteration 10980/ 52452 | consumed samples: 5621760 | consumed tokens: 11513364480 | elapsed time per iteration (s): 1.82 | learning rate: 1.826E-04 | global batch size: 512 | lm loss: 1.554439E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.785 | TFLOPs: 42.29 | 31: iteration 10990/ 52452 | consumed samples: 5626880 | consumed tokens: 11523850240 | elapsed time per iteration (s): 1.77 | learning rate: 1.826E-04 | global batch size: 512 | lm loss: 1.540223E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.552 | TFLOPs: 43.31 | 31: iteration 11000/ 52452 | consumed samples: 5632000 | consumed tokens: 11534336000 | elapsed time per iteration (s): 1.75 | learning rate: 1.825E-04 | global batch size: 512 | lm loss: 1.551560E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.726 | TFLOPs: 43.94 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11000 | lm loss value: 4.160478E+00 | lm loss PPL: 6.410213E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11000 | lm loss value: 4.157666E+00 | lm loss PPL: 6.392214E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11010/ 52452 | consumed samples: 5637120 | consumed tokens: 11544821760 | elapsed time per iteration (s): 9.18 | learning rate: 1.825E-04 | global batch size: 512 | lm loss: 1.552900E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.774 | TFLOPs: 8.37 | 31: iteration 11020/ 52452 | consumed samples: 5642240 | consumed tokens: 11555307520 | elapsed time per iteration (s): 1.76 | learning rate: 1.825E-04 | global batch size: 512 | lm loss: 1.543755E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.228 | TFLOPs: 43.71 | 31: iteration 11030/ 52452 | consumed samples: 5647360 | consumed tokens: 11565793280 | elapsed time per iteration (s): 1.75 | learning rate: 1.824E-04 | global batch size: 512 | lm loss: 1.551267E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.365 | TFLOPs: 43.88 | 31: iteration 11040/ 52452 | consumed samples: 5652480 | consumed tokens: 11576279040 | elapsed time per iteration (s): 1.82 | learning rate: 1.824E-04 | global batch size: 512 | lm loss: 1.541873E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.934 | TFLOPs: 42.17 | 31: iteration 11050/ 52452 | consumed samples: 5657600 | consumed tokens: 11586764800 | elapsed time per iteration (s): 1.82 | learning rate: 1.824E-04 | global batch size: 512 | lm loss: 1.543221E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.093 | TFLOPs: 42.19 | 31: iteration 11060/ 52452 | consumed samples: 5662720 | consumed tokens: 11597250560 | elapsed time per iteration (s): 1.78 | learning rate: 1.823E-04 | global batch size: 512 | lm loss: 1.548942E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.753 | TFLOPs: 43.19 | 31: iteration 11070/ 52452 | consumed samples: 5667840 | consumed tokens: 11607736320 | elapsed time per iteration (s): 1.78 | learning rate: 1.823E-04 | global batch size: 512 | lm loss: 1.554433E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.092 | TFLOPs: 43.24 | 31: iteration 11080/ 52452 | consumed samples: 5672960 | consumed tokens: 11618222080 | elapsed time per iteration (s): 1.85 | learning rate: 1.823E-04 | global batch size: 512 | lm loss: 1.539502E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.980 | TFLOPs: 41.57 | 31: iteration 11090/ 52452 | consumed samples: 5678080 | consumed tokens: 11628707840 | elapsed time per iteration (s): 1.76 | learning rate: 1.822E-04 | global batch size: 512 | lm loss: 1.543482E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.340 | TFLOPs: 43.73 | 31: iteration 11100/ 52452 | consumed samples: 5683200 | consumed tokens: 11639193600 | elapsed time per iteration (s): 1.78 | learning rate: 1.822E-04 | global batch size: 512 | lm loss: 1.542863E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.003 | TFLOPs: 43.08 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11100 | lm loss value: 4.180393E+00 | lm loss PPL: 6.539153E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11100 | lm loss value: 4.179310E+00 | lm loss PPL: 6.532079E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11110/ 52452 | consumed samples: 5688320 | consumed tokens: 11649679360 | elapsed time per iteration (s): 9.19 | learning rate: 1.822E-04 | global batch size: 512 | lm loss: 1.545234E+00 | grad norm: 0.209 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.718 | TFLOPs: 8.36 | 31: iteration 11120/ 52452 | consumed samples: 5693440 | consumed tokens: 11660165120 | elapsed time per iteration (s): 1.80 | learning rate: 1.821E-04 | global batch size: 512 | lm loss: 1.556518E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.177 | TFLOPs: 42.65 | 31: iteration 11130/ 52452 | consumed samples: 5698560 | consumed tokens: 11670650880 | elapsed time per iteration (s): 1.75 | learning rate: 1.821E-04 | global batch size: 512 | lm loss: 1.536694E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.193 | TFLOPs: 43.86 | 31: iteration 11140/ 52452 | consumed samples: 5703680 | consumed tokens: 11681136640 | elapsed time per iteration (s): 1.79 | learning rate: 1.821E-04 | global batch size: 512 | lm loss: 1.538666E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.560 | TFLOPs: 42.86 | 31: iteration 11150/ 52452 | consumed samples: 5708800 | consumed tokens: 11691622400 | elapsed time per iteration (s): 1.81 | learning rate: 1.820E-04 | global batch size: 512 | lm loss: 1.539933E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.447 | TFLOPs: 42.54 | 31: iteration 11160/ 52452 | consumed samples: 5713920 | consumed tokens: 11702108160 | elapsed time per iteration (s): 1.79 | learning rate: 1.820E-04 | global batch size: 512 | lm loss: 1.546176E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.585 | TFLOPs: 43.01 | 31: iteration 11170/ 52452 | consumed samples: 5719040 | consumed tokens: 11712593920 | elapsed time per iteration (s): 1.73 | learning rate: 1.820E-04 | global batch size: 512 | lm loss: 1.539483E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.584 | TFLOPs: 44.37 | 31: iteration 11180/ 52452 | consumed samples: 5724160 | consumed tokens: 11723079680 | elapsed time per iteration (s): 1.72 | learning rate: 1.819E-04 | global batch size: 512 | lm loss: 1.541525E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.996 | TFLOPs: 44.58 | 31: iteration 11190/ 52452 | consumed samples: 5729280 | consumed tokens: 11733565440 | elapsed time per iteration (s): 1.73 | learning rate: 1.819E-04 | global batch size: 512 | lm loss: 1.546466E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.265 | TFLOPs: 44.32 | 31: iteration 11200/ 52452 | consumed samples: 5734400 | consumed tokens: 11744051200 | elapsed time per iteration (s): 1.76 | learning rate: 1.819E-04 | global batch size: 512 | lm loss: 1.526680E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.325 | TFLOPs: 43.58 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11200 | lm loss value: 4.212658E+00 | lm loss PPL: 6.753581E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11200 | lm loss value: 4.181877E+00 | lm loss PPL: 6.548867E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11210/ 52452 | consumed samples: 5739520 | consumed tokens: 11754536960 | elapsed time per iteration (s): 9.22 | learning rate: 1.818E-04 | global batch size: 512 | lm loss: 1.542254E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.541 | TFLOPs: 8.34 | 31: iteration 11220/ 52452 | consumed samples: 5744640 | consumed tokens: 11765022720 | elapsed time per iteration (s): 1.74 | learning rate: 1.818E-04 | global batch size: 512 | lm loss: 1.537759E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.510 | TFLOPs: 44.05 | 31: iteration 11230/ 52452 | consumed samples: 5749760 | consumed tokens: 11775508480 | elapsed time per iteration (s): 1.73 | learning rate: 1.818E-04 | global batch size: 512 | lm loss: 1.526462E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.225 | TFLOPs: 44.46 | 31: iteration 11240/ 52452 | consumed samples: 5754880 | consumed tokens: 11785994240 | elapsed time per iteration (s): 1.74 | learning rate: 1.817E-04 | global batch size: 512 | lm loss: 1.556037E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.189 | TFLOPs: 44.16 | 31: iteration 11250/ 52452 | consumed samples: 5760000 | consumed tokens: 11796480000 | elapsed time per iteration (s): 1.73 | learning rate: 1.817E-04 | global batch size: 512 | lm loss: 1.551783E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.846 | TFLOPs: 44.40 | 31: iteration 11260/ 52452 | consumed samples: 5765120 | consumed tokens: 11806965760 | elapsed time per iteration (s): 1.76 | learning rate: 1.817E-04 | global batch size: 512 | lm loss: 1.546193E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.657 | TFLOPs: 43.63 | 31: iteration 11270/ 52452 | consumed samples: 5770240 | consumed tokens: 11817451520 | elapsed time per iteration (s): 1.76 | learning rate: 1.816E-04 | global batch size: 512 | lm loss: 1.529964E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.467 | TFLOPs: 43.60 | 31: iteration 11280/ 52452 | consumed samples: 5775360 | consumed tokens: 11827937280 | elapsed time per iteration (s): 1.89 | learning rate: 1.816E-04 | global batch size: 512 | lm loss: 1.543965E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.512 | TFLOPs: 40.75 | 31: iteration 11290/ 52452 | consumed samples: 5780480 | consumed tokens: 11838423040 | elapsed time per iteration (s): 1.77 | learning rate: 1.816E-04 | global batch size: 512 | lm loss: 1.537678E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.074 | TFLOPs: 43.39 | 31: iteration 11300/ 52452 | consumed samples: 5785600 | consumed tokens: 11848908800 | elapsed time per iteration (s): 1.75 | learning rate: 1.815E-04 | global batch size: 512 | lm loss: 1.533647E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.098 | TFLOPs: 43.84 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11300 | lm loss value: 4.131279E+00 | lm loss PPL: 6.225750E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11300 | lm loss value: 4.117549E+00 | lm loss PPL: 6.140857E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11310/ 52452 | consumed samples: 5790720 | consumed tokens: 11859394560 | elapsed time per iteration (s): 9.13 | learning rate: 1.815E-04 | global batch size: 512 | lm loss: 1.539902E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.050 | TFLOPs: 8.41 | 31: iteration 11320/ 52452 | consumed samples: 5795840 | consumed tokens: 11869880320 | elapsed time per iteration (s): 1.73 | learning rate: 1.815E-04 | global batch size: 512 | lm loss: 1.542241E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.155 | TFLOPs: 44.30 | 31: iteration 11330/ 52452 | consumed samples: 5800960 | consumed tokens: 11880366080 | elapsed time per iteration (s): 1.76 | learning rate: 1.814E-04 | global batch size: 512 | lm loss: 1.530272E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.761 | TFLOPs: 43.64 | 31: iteration 11340/ 52452 | consumed samples: 5806080 | consumed tokens: 11890851840 | elapsed time per iteration (s): 1.74 | learning rate: 1.814E-04 | global batch size: 512 | lm loss: 1.542339E+00 | grad norm: 0.203 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.048 | TFLOPs: 44.14 | 31: iteration 11350/ 52452 | consumed samples: 5811200 | consumed tokens: 11901337600 | elapsed time per iteration (s): 1.76 | learning rate: 1.814E-04 | global batch size: 512 | lm loss: 1.539291E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.957 | TFLOPs: 43.67 | 31: iteration 11360/ 52452 | consumed samples: 5816320 | consumed tokens: 11911823360 | elapsed time per iteration (s): 1.75 | learning rate: 1.813E-04 | global batch size: 512 | lm loss: 1.538105E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.507 | TFLOPs: 43.90 | 31: iteration 11370/ 52452 | consumed samples: 5821440 | consumed tokens: 11922309120 | elapsed time per iteration (s): 1.74 | learning rate: 1.813E-04 | global batch size: 512 | lm loss: 1.530604E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.649 | TFLOPs: 44.08 | 31: iteration 11380/ 52452 | consumed samples: 5826560 | consumed tokens: 11932794880 | elapsed time per iteration (s): 1.81 | learning rate: 1.813E-04 | global batch size: 512 | lm loss: 1.540566E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.328 | TFLOPs: 42.38 | 31: iteration 11390/ 52452 | consumed samples: 5831680 | consumed tokens: 11943280640 | elapsed time per iteration (s): 1.77 | learning rate: 1.812E-04 | global batch size: 512 | lm loss: 1.541941E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.981 | TFLOPs: 43.37 | 31: iteration 11400/ 52452 | consumed samples: 5836800 | consumed tokens: 11953766400 | elapsed time per iteration (s): 1.75 | learning rate: 1.812E-04 | global batch size: 512 | lm loss: 1.551111E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.015 | TFLOPs: 43.98 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11400 | lm loss value: 4.071367E+00 | lm loss PPL: 5.863708E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11400 | lm loss value: 4.104965E+00 | lm loss PPL: 6.064061E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11410/ 52452 | consumed samples: 5841920 | consumed tokens: 11964252160 | elapsed time per iteration (s): 9.14 | learning rate: 1.812E-04 | global batch size: 512 | lm loss: 1.530876E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.038 | TFLOPs: 8.41 | 31: iteration 11420/ 52452 | consumed samples: 5847040 | consumed tokens: 11974737920 | elapsed time per iteration (s): 1.74 | learning rate: 1.811E-04 | global batch size: 512 | lm loss: 1.537629E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.765 | TFLOPs: 44.09 | 31: iteration 11430/ 52452 | consumed samples: 5852160 | consumed tokens: 11985223680 | elapsed time per iteration (s): 1.73 | learning rate: 1.811E-04 | global batch size: 512 | lm loss: 1.529215E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.539 | TFLOPs: 44.51 | 31: iteration 11440/ 52452 | consumed samples: 5857280 | consumed tokens: 11995709440 | elapsed time per iteration (s): 1.81 | learning rate: 1.811E-04 | global batch size: 512 | lm loss: 1.542155E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.615 | TFLOPs: 42.57 | 31: iteration 11450/ 52452 | consumed samples: 5862400 | consumed tokens: 12006195200 | elapsed time per iteration (s): 1.78 | learning rate: 1.810E-04 | global batch size: 512 | lm loss: 1.539638E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.509 | TFLOPs: 43.15 | 31: iteration 11460/ 52452 | consumed samples: 5867520 | consumed tokens: 12016680960 | elapsed time per iteration (s): 1.77 | learning rate: 1.810E-04 | global batch size: 512 | lm loss: 1.529493E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.690 | TFLOPs: 43.48 | 31: iteration 11470/ 52452 | consumed samples: 5872640 | consumed tokens: 12027166720 | elapsed time per iteration (s): 1.77 | learning rate: 1.810E-04 | global batch size: 512 | lm loss: 1.539619E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.186 | TFLOPs: 43.41 | 31: iteration 11480/ 52452 | consumed samples: 5877760 | consumed tokens: 12037652480 | elapsed time per iteration (s): 1.78 | learning rate: 1.809E-04 | global batch size: 512 | lm loss: 1.518918E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.055 | TFLOPs: 43.09 | 31: iteration 11490/ 52452 | consumed samples: 5882880 | consumed tokens: 12048138240 | elapsed time per iteration (s): 1.74 | learning rate: 1.809E-04 | global batch size: 512 | lm loss: 1.530916E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.902 | TFLOPs: 44.26 | 31: iteration 11500/ 52452 | consumed samples: 5888000 | consumed tokens: 12058624000 | elapsed time per iteration (s): 1.74 | learning rate: 1.809E-04 | global batch size: 512 | lm loss: 1.527754E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.040 | TFLOPs: 44.13 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11500 | lm loss value: 4.092633E+00 | lm loss PPL: 5.989738E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11500 | lm loss value: 4.093918E+00 | lm loss PPL: 5.997440E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11510/ 52452 | consumed samples: 5893120 | consumed tokens: 12069109760 | elapsed time per iteration (s): 9.14 | learning rate: 1.808E-04 | global batch size: 512 | lm loss: 1.536810E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.015 | TFLOPs: 8.41 | 31: iteration 11520/ 52452 | consumed samples: 5898240 | consumed tokens: 12079595520 | elapsed time per iteration (s): 1.80 | learning rate: 1.808E-04 | global batch size: 512 | lm loss: 1.537269E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.310 | TFLOPs: 42.67 | 31: iteration 11530/ 52452 | consumed samples: 5903360 | consumed tokens: 12090081280 | elapsed time per iteration (s): 1.78 | learning rate: 1.808E-04 | global batch size: 512 | lm loss: 1.538815E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.838 | TFLOPs: 43.20 | 31: iteration 11540/ 52452 | consumed samples: 5908480 | consumed tokens: 12100567040 | elapsed time per iteration (s): 1.72 | learning rate: 1.807E-04 | global batch size: 512 | lm loss: 1.526773E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.255 | TFLOPs: 44.62 | 31: iteration 11550/ 52452 | consumed samples: 5913600 | consumed tokens: 12111052800 | elapsed time per iteration (s): 1.77 | learning rate: 1.807E-04 | global batch size: 512 | lm loss: 1.550907E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.193 | TFLOPs: 43.41 | 31: iteration 11560/ 52452 | consumed samples: 5918720 | consumed tokens: 12121538560 | elapsed time per iteration (s): 1.76 | learning rate: 1.807E-04 | global batch size: 512 | lm loss: 1.533594E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.637 | TFLOPs: 43.77 | 31: iteration 11570/ 52452 | consumed samples: 5923840 | consumed tokens: 12132024320 | elapsed time per iteration (s): 1.74 | learning rate: 1.806E-04 | global batch size: 512 | lm loss: 1.534393E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.567 | TFLOPs: 44.21 | 31: iteration 11580/ 52452 | consumed samples: 5928960 | consumed tokens: 12142510080 | elapsed time per iteration (s): 1.82 | learning rate: 1.806E-04 | global batch size: 512 | lm loss: 1.539727E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.987 | TFLOPs: 42.32 | 31: iteration 11590/ 52452 | consumed samples: 5934080 | consumed tokens: 12152995840 | elapsed time per iteration (s): 1.77 | learning rate: 1.806E-04 | global batch size: 512 | lm loss: 1.523511E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.476 | TFLOPs: 43.45 | 31: iteration 11600/ 52452 | consumed samples: 5939200 | consumed tokens: 12163481600 | elapsed time per iteration (s): 1.74 | learning rate: 1.805E-04 | global batch size: 512 | lm loss: 1.536137E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.694 | TFLOPs: 44.08 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11600 | lm loss value: 4.035996E+00 | lm loss PPL: 5.659926E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11600 | lm loss value: 4.040091E+00 | lm loss PPL: 5.683152E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11610/ 52452 | consumed samples: 5944320 | consumed tokens: 12173967360 | elapsed time per iteration (s): 9.20 | learning rate: 1.805E-04 | global batch size: 512 | lm loss: 1.533068E+00 | grad norm: 0.196 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.654 | TFLOPs: 8.35 | 31: iteration 11620/ 52452 | consumed samples: 5949440 | consumed tokens: 12184453120 | elapsed time per iteration (s): 1.81 | learning rate: 1.805E-04 | global batch size: 512 | lm loss: 1.538786E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.049 | TFLOPs: 42.48 | 31: iteration 11630/ 52452 | consumed samples: 5954560 | consumed tokens: 12194938880 | elapsed time per iteration (s): 1.86 | learning rate: 1.804E-04 | global batch size: 512 | lm loss: 1.546064E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.348 | TFLOPs: 41.33 | 31: iteration 11640/ 52452 | consumed samples: 5959680 | consumed tokens: 12205424640 | elapsed time per iteration (s): 1.85 | learning rate: 1.804E-04 | global batch size: 512 | lm loss: 1.542821E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.904 | TFLOPs: 41.56 | 31: iteration 11650/ 52452 | consumed samples: 5964800 | consumed tokens: 12215910400 | elapsed time per iteration (s): 1.82 | learning rate: 1.804E-04 | global batch size: 512 | lm loss: 1.524177E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.566 | TFLOPs: 42.26 | 31: iteration 11660/ 52452 | consumed samples: 5969920 | consumed tokens: 12226396160 | elapsed time per iteration (s): 1.79 | learning rate: 1.803E-04 | global batch size: 512 | lm loss: 1.546383E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.030 | TFLOPs: 42.93 | 31: iteration 11670/ 52452 | consumed samples: 5975040 | consumed tokens: 12236881920 | elapsed time per iteration (s): 1.83 | learning rate: 1.803E-04 | global batch size: 512 | lm loss: 1.524558E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.530 | TFLOPs: 41.96 | 31: iteration 11680/ 52452 | consumed samples: 5980160 | consumed tokens: 12247367680 | elapsed time per iteration (s): 1.79 | learning rate: 1.803E-04 | global batch size: 512 | lm loss: 1.536379E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.496 | TFLOPs: 42.85 | 31: iteration 11690/ 52452 | consumed samples: 5985280 | consumed tokens: 12257853440 | elapsed time per iteration (s): 1.77 | learning rate: 1.802E-04 | global batch size: 512 | lm loss: 1.541983E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.256 | TFLOPs: 43.42 | 31: iteration 11700/ 52452 | consumed samples: 5990400 | consumed tokens: 12268339200 | elapsed time per iteration (s): 1.78 | learning rate: 1.802E-04 | global batch size: 512 | lm loss: 1.523014E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.879 | TFLOPs: 43.06 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11700 | lm loss value: 4.169176E+00 | lm loss PPL: 6.466216E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11700 | lm loss value: 4.147730E+00 | lm loss PPL: 6.329016E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11710/ 52452 | consumed samples: 5995520 | consumed tokens: 12278824960 | elapsed time per iteration (s): 9.23 | learning rate: 1.802E-04 | global batch size: 512 | lm loss: 1.540252E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.490 | TFLOPs: 8.33 | 31: iteration 11720/ 52452 | consumed samples: 6000640 | consumed tokens: 12289310720 | elapsed time per iteration (s): 1.74 | learning rate: 1.801E-04 | global batch size: 512 | lm loss: 1.523736E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.770 | TFLOPs: 44.24 | 31: iteration 11730/ 52452 | consumed samples: 6005760 | consumed tokens: 12299796480 | elapsed time per iteration (s): 1.79 | learning rate: 1.801E-04 | global batch size: 512 | lm loss: 1.523466E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.505 | TFLOPs: 42.85 | 31: iteration 11740/ 52452 | consumed samples: 6010880 | consumed tokens: 12310282240 | elapsed time per iteration (s): 1.80 | learning rate: 1.801E-04 | global batch size: 512 | lm loss: 1.521747E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.050 | TFLOPs: 42.63 | 31: iteration 11750/ 52452 | consumed samples: 6016000 | consumed tokens: 12320768000 | elapsed time per iteration (s): 1.83 | learning rate: 1.800E-04 | global batch size: 512 | lm loss: 1.528519E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.161 | TFLOPs: 41.90 | 31: iteration 11760/ 52452 | consumed samples: 6021120 | consumed tokens: 12331253760 | elapsed time per iteration (s): 1.79 | learning rate: 1.800E-04 | global batch size: 512 | lm loss: 1.532921E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.834 | TFLOPs: 42.90 | 31: iteration 11770/ 52452 | consumed samples: 6026240 | consumed tokens: 12341739520 | elapsed time per iteration (s): 1.79 | learning rate: 1.800E-04 | global batch size: 512 | lm loss: 1.536936E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.881 | TFLOPs: 42.91 | 31: iteration 11780/ 52452 | consumed samples: 6031360 | consumed tokens: 12352225280 | elapsed time per iteration (s): 1.82 | learning rate: 1.799E-04 | global batch size: 512 | lm loss: 1.535039E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.611 | TFLOPs: 42.12 | 31: iteration 11790/ 52452 | consumed samples: 6036480 | consumed tokens: 12362711040 | elapsed time per iteration (s): 1.80 | learning rate: 1.799E-04 | global batch size: 512 | lm loss: 1.526871E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.065 | TFLOPs: 42.64 | 31: iteration 11800/ 52452 | consumed samples: 6041600 | consumed tokens: 12373196800 | elapsed time per iteration (s): 1.79 | learning rate: 1.799E-04 | global batch size: 512 | lm loss: 1.538982E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.103 | TFLOPs: 42.94 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11800 | lm loss value: 4.209925E+00 | lm loss PPL: 6.735150E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11800 | lm loss value: 4.185045E+00 | lm loss PPL: 6.569644E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11810/ 52452 | consumed samples: 6046720 | consumed tokens: 12383682560 | elapsed time per iteration (s): 9.27 | learning rate: 1.798E-04 | global batch size: 512 | lm loss: 1.534773E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.242 | TFLOPs: 8.29 | 31: iteration 11820/ 52452 | consumed samples: 6051840 | consumed tokens: 12394168320 | elapsed time per iteration (s): 1.81 | learning rate: 1.798E-04 | global batch size: 512 | lm loss: 1.525737E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.739 | TFLOPs: 42.44 | 31: iteration 11830/ 52452 | consumed samples: 6056960 | consumed tokens: 12404654080 | elapsed time per iteration (s): 1.87 | learning rate: 1.798E-04 | global batch size: 512 | lm loss: 1.516484E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.412 | TFLOPs: 41.04 | 31: iteration 11840/ 52452 | consumed samples: 6062080 | consumed tokens: 12415139840 | elapsed time per iteration (s): 1.83 | learning rate: 1.797E-04 | global batch size: 512 | lm loss: 1.527676E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.116 | TFLOPs: 42.04 | 31: iteration 11850/ 52452 | consumed samples: 6067200 | consumed tokens: 12425625600 | elapsed time per iteration (s): 1.85 | learning rate: 1.797E-04 | global batch size: 512 | lm loss: 1.524682E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.399 | TFLOPs: 41.49 | 31: iteration 11860/ 52452 | consumed samples: 6072320 | consumed tokens: 12436111360 | elapsed time per iteration (s): 1.98 | learning rate: 1.797E-04 | global batch size: 512 | lm loss: 1.539178E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 258.488 | TFLOPs: 38.80 | 31: iteration 11870/ 52452 | consumed samples: 6077440 | consumed tokens: 12446597120 | elapsed time per iteration (s): 1.90 | learning rate: 1.796E-04 | global batch size: 512 | lm loss: 1.520546E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.177 | TFLOPs: 40.40 | 31: iteration 11880/ 52452 | consumed samples: 6082560 | consumed tokens: 12457082880 | elapsed time per iteration (s): 1.77 | learning rate: 1.796E-04 | global batch size: 512 | lm loss: 1.526277E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.674 | TFLOPs: 43.33 | 31: iteration 11890/ 52452 | consumed samples: 6087680 | consumed tokens: 12467568640 | elapsed time per iteration (s): 1.78 | learning rate: 1.795E-04 | global batch size: 512 | lm loss: 1.518620E+00 | grad norm: 0.194 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.391 | TFLOPs: 43.14 | 31: iteration 11900/ 52452 | consumed samples: 6092800 | consumed tokens: 12478054400 | elapsed time per iteration (s): 1.77 | learning rate: 1.795E-04 | global batch size: 512 | lm loss: 1.532143E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.744 | TFLOPs: 43.34 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 11900 | lm loss value: 4.188198E+00 | lm loss PPL: 6.590393E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 11900 | lm loss value: 4.203684E+00 | lm loss PPL: 6.693245E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 11910/ 52452 | consumed samples: 6097920 | consumed tokens: 12488540160 | elapsed time per iteration (s): 9.14 | learning rate: 1.795E-04 | global batch size: 512 | lm loss: 1.531560E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.013 | TFLOPs: 8.41 | 31: iteration 11920/ 52452 | consumed samples: 6103040 | consumed tokens: 12499025920 | elapsed time per iteration (s): 1.76 | learning rate: 1.794E-04 | global batch size: 512 | lm loss: 1.538768E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.123 | TFLOPs: 43.55 | 31: iteration 11930/ 52452 | consumed samples: 6108160 | consumed tokens: 12509511680 | elapsed time per iteration (s): 1.74 | learning rate: 1.794E-04 | global batch size: 512 | lm loss: 1.513697E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.495 | TFLOPs: 44.20 | 31: iteration 11940/ 52452 | consumed samples: 6113280 | consumed tokens: 12519997440 | elapsed time per iteration (s): 1.82 | learning rate: 1.794E-04 | global batch size: 512 | lm loss: 1.530338E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.192 | TFLOPs: 42.21 | 31: iteration 11950/ 52452 | consumed samples: 6118400 | consumed tokens: 12530483200 | elapsed time per iteration (s): 1.80 | learning rate: 1.793E-04 | global batch size: 512 | lm loss: 1.519640E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.179 | TFLOPs: 42.80 | 31: iteration 11960/ 52452 | consumed samples: 6123520 | consumed tokens: 12540968960 | elapsed time per iteration (s): 1.76 | learning rate: 1.793E-04 | global batch size: 512 | lm loss: 1.527696E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.444 | TFLOPs: 43.74 | 31: iteration 11970/ 52452 | consumed samples: 6128640 | consumed tokens: 12551454720 | elapsed time per iteration (s): 1.92 | learning rate: 1.793E-04 | global batch size: 512 | lm loss: 1.522836E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.333 | TFLOPs: 40.13 | 31: iteration 11980/ 52452 | consumed samples: 6133760 | consumed tokens: 12561940480 | elapsed time per iteration (s): 1.76 | learning rate: 1.792E-04 | global batch size: 512 | lm loss: 1.542734E+00 | grad norm: 0.210 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.573 | TFLOPs: 43.61 | 31: iteration 11990/ 52452 | consumed samples: 6138880 | consumed tokens: 12572426240 | elapsed time per iteration (s): 1.82 | learning rate: 1.792E-04 | global batch size: 512 | lm loss: 1.532918E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.238 | TFLOPs: 42.21 | 0: [2022-12-25 17:28:17,842] [INFO] [logging.py:68:log_dist] [Rank 0] step=12000, skipped=0, lr=[0.0001791674120615615, 0.0001791674120615615, 0.0001791674120615615], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 12000/ 52452 | consumed samples: 6144000 | consumed tokens: 12582912000 | elapsed time per iteration (s): 1.84 | learning rate: 1.792E-04 | global batch size: 512 | lm loss: 1.521848E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.634 | TFLOPs: 41.82 | 0: steps: 12000 loss: 1.5088 iter time (s): 1.777 samples/sec: 288.178 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12000 | lm loss value: 4.253230E+00 | lm loss PPL: 7.033219E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12000 | lm loss value: 4.251334E+00 | lm loss PPL: 7.019901E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12010/ 52452 | consumed samples: 6149120 | consumed tokens: 12593397760 | elapsed time per iteration (s): 9.23 | learning rate: 1.791E-04 | global batch size: 512 | lm loss: 1.525082E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.471 | TFLOPs: 8.33 | 31: iteration 12020/ 52452 | consumed samples: 6154240 | consumed tokens: 12603883520 | elapsed time per iteration (s): 1.77 | learning rate: 1.791E-04 | global batch size: 512 | lm loss: 1.534735E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.951 | TFLOPs: 43.52 | 31: iteration 12030/ 52452 | consumed samples: 6159360 | consumed tokens: 12614369280 | elapsed time per iteration (s): 1.79 | learning rate: 1.791E-04 | global batch size: 512 | lm loss: 1.522891E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.131 | TFLOPs: 42.95 | 31: iteration 12040/ 52452 | consumed samples: 6164480 | consumed tokens: 12624855040 | elapsed time per iteration (s): 1.79 | learning rate: 1.790E-04 | global batch size: 512 | lm loss: 1.528751E+00 | grad norm: 0.197 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.709 | TFLOPs: 43.03 | 31: iteration 12050/ 52452 | consumed samples: 6169600 | consumed tokens: 12635340800 | elapsed time per iteration (s): 1.74 | learning rate: 1.790E-04 | global batch size: 512 | lm loss: 1.533398E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.899 | TFLOPs: 44.26 | 31: iteration 12060/ 52452 | consumed samples: 6174720 | consumed tokens: 12645826560 | elapsed time per iteration (s): 1.81 | learning rate: 1.790E-04 | global batch size: 512 | lm loss: 1.522400E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.630 | TFLOPs: 42.57 | 31: iteration 12070/ 52452 | consumed samples: 6179840 | consumed tokens: 12656312320 | elapsed time per iteration (s): 1.85 | learning rate: 1.789E-04 | global batch size: 512 | lm loss: 1.519752E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.698 | TFLOPs: 41.53 | 31: iteration 12080/ 52452 | consumed samples: 6184960 | consumed tokens: 12666798080 | elapsed time per iteration (s): 1.76 | learning rate: 1.789E-04 | global batch size: 512 | lm loss: 1.524164E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.980 | TFLOPs: 43.67 | 31: iteration 12090/ 52452 | consumed samples: 6190080 | consumed tokens: 12677283840 | elapsed time per iteration (s): 1.73 | learning rate: 1.789E-04 | global batch size: 512 | lm loss: 1.535979E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.501 | TFLOPs: 44.35 | 31: iteration 12100/ 52452 | consumed samples: 6195200 | consumed tokens: 12687769600 | elapsed time per iteration (s): 1.80 | learning rate: 1.788E-04 | global batch size: 512 | lm loss: 1.524001E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.859 | TFLOPs: 42.61 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12100 | lm loss value: 3.984103E+00 | lm loss PPL: 5.373705E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12100 | lm loss value: 4.023091E+00 | lm loss PPL: 5.587356E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12110/ 52452 | consumed samples: 6200320 | consumed tokens: 12698255360 | elapsed time per iteration (s): 9.18 | learning rate: 1.788E-04 | global batch size: 512 | lm loss: 1.524686E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.745 | TFLOPs: 8.37 | 31: iteration 12120/ 52452 | consumed samples: 6205440 | consumed tokens: 12708741120 | elapsed time per iteration (s): 1.77 | learning rate: 1.787E-04 | global batch size: 512 | lm loss: 1.515942E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.584 | TFLOPs: 43.31 | 31: iteration 12130/ 52452 | consumed samples: 6210560 | consumed tokens: 12719226880 | elapsed time per iteration (s): 1.80 | learning rate: 1.787E-04 | global batch size: 512 | lm loss: 1.512472E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.931 | TFLOPs: 42.62 | 31: iteration 12140/ 52452 | consumed samples: 6215680 | consumed tokens: 12729712640 | elapsed time per iteration (s): 1.80 | learning rate: 1.787E-04 | global batch size: 512 | lm loss: 1.518582E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.150 | TFLOPs: 42.80 | 31: iteration 12150/ 52452 | consumed samples: 6220800 | consumed tokens: 12740198400 | elapsed time per iteration (s): 1.86 | learning rate: 1.786E-04 | global batch size: 512 | lm loss: 1.531770E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.710 | TFLOPs: 41.38 | 31: iteration 12160/ 52452 | consumed samples: 6225920 | consumed tokens: 12750684160 | elapsed time per iteration (s): 1.88 | learning rate: 1.786E-04 | global batch size: 512 | lm loss: 1.506833E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.759 | TFLOPs: 40.94 | 31: iteration 12170/ 52452 | consumed samples: 6231040 | consumed tokens: 12761169920 | elapsed time per iteration (s): 1.79 | learning rate: 1.786E-04 | global batch size: 512 | lm loss: 1.534327E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.376 | TFLOPs: 42.98 | 31: iteration 12180/ 52452 | consumed samples: 6236160 | consumed tokens: 12771655680 | elapsed time per iteration (s): 1.79 | learning rate: 1.785E-04 | global batch size: 512 | lm loss: 1.505319E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.006 | TFLOPs: 42.93 | 31: iteration 12190/ 52452 | consumed samples: 6241280 | consumed tokens: 12782141440 | elapsed time per iteration (s): 1.83 | learning rate: 1.785E-04 | global batch size: 512 | lm loss: 1.499508E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.229 | TFLOPs: 41.91 | 31: iteration 12200/ 52452 | consumed samples: 6246400 | consumed tokens: 12792627200 | elapsed time per iteration (s): 1.74 | learning rate: 1.785E-04 | global batch size: 512 | lm loss: 1.521115E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.418 | TFLOPs: 44.04 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12200 | lm loss value: 4.091706E+00 | lm loss PPL: 5.984188E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12200 | lm loss value: 4.114633E+00 | lm loss PPL: 6.122974E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12210/ 52452 | consumed samples: 6251520 | consumed tokens: 12803112960 | elapsed time per iteration (s): 9.19 | learning rate: 1.784E-04 | global batch size: 512 | lm loss: 1.528200E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.743 | TFLOPs: 8.37 | 31: iteration 12220/ 52452 | consumed samples: 6256640 | consumed tokens: 12813598720 | elapsed time per iteration (s): 1.88 | learning rate: 1.784E-04 | global batch size: 512 | lm loss: 1.510218E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.031 | TFLOPs: 40.83 | 31: iteration 12230/ 52452 | consumed samples: 6261760 | consumed tokens: 12824084480 | elapsed time per iteration (s): 1.83 | learning rate: 1.784E-04 | global batch size: 512 | lm loss: 1.517807E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.033 | TFLOPs: 42.03 | 31: iteration 12240/ 52452 | consumed samples: 6266880 | consumed tokens: 12834570240 | elapsed time per iteration (s): 1.75 | learning rate: 1.783E-04 | global batch size: 512 | lm loss: 1.500334E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.828 | TFLOPs: 43.80 | 31: iteration 12250/ 52452 | consumed samples: 6272000 | consumed tokens: 12845056000 | elapsed time per iteration (s): 1.84 | learning rate: 1.783E-04 | global batch size: 512 | lm loss: 1.530057E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.745 | TFLOPs: 41.69 | 31: iteration 12260/ 52452 | consumed samples: 6277120 | consumed tokens: 12855541760 | elapsed time per iteration (s): 1.79 | learning rate: 1.783E-04 | global batch size: 512 | lm loss: 1.521237E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.409 | TFLOPs: 42.99 | 31: iteration 12270/ 52452 | consumed samples: 6282240 | consumed tokens: 12866027520 | elapsed time per iteration (s): 1.84 | learning rate: 1.782E-04 | global batch size: 512 | lm loss: 1.519010E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.396 | TFLOPs: 41.79 | 31: iteration 12280/ 52452 | consumed samples: 6287360 | consumed tokens: 12876513280 | elapsed time per iteration (s): 1.85 | learning rate: 1.782E-04 | global batch size: 512 | lm loss: 1.521747E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.048 | TFLOPs: 41.43 | 31: iteration 12290/ 52452 | consumed samples: 6292480 | consumed tokens: 12886999040 | elapsed time per iteration (s): 1.77 | learning rate: 1.781E-04 | global batch size: 512 | lm loss: 1.516858E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.941 | TFLOPs: 43.37 | 31: iteration 12300/ 52452 | consumed samples: 6297600 | consumed tokens: 12897484800 | elapsed time per iteration (s): 1.78 | learning rate: 1.781E-04 | global batch size: 512 | lm loss: 1.515810E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.219 | TFLOPs: 43.11 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12300 | lm loss value: 4.155126E+00 | lm loss PPL: 6.375997E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12300 | lm loss value: 4.181764E+00 | lm loss PPL: 6.548124E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12310/ 52452 | consumed samples: 6302720 | consumed tokens: 12907970560 | elapsed time per iteration (s): 9.14 | learning rate: 1.781E-04 | global batch size: 512 | lm loss: 1.531349E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.998 | TFLOPs: 8.40 | 31: iteration 12320/ 52452 | consumed samples: 6307840 | consumed tokens: 12918456320 | elapsed time per iteration (s): 1.75 | learning rate: 1.780E-04 | global batch size: 512 | lm loss: 1.513328E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.399 | TFLOPs: 44.04 | 31: iteration 12330/ 52452 | consumed samples: 6312960 | consumed tokens: 12928942080 | elapsed time per iteration (s): 1.77 | learning rate: 1.780E-04 | global batch size: 512 | lm loss: 1.518379E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.029 | TFLOPs: 43.38 | 31: iteration 12340/ 52452 | consumed samples: 6318080 | consumed tokens: 12939427840 | elapsed time per iteration (s): 1.73 | learning rate: 1.780E-04 | global batch size: 512 | lm loss: 1.530414E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.750 | TFLOPs: 44.39 | 31: iteration 12350/ 52452 | consumed samples: 6323200 | consumed tokens: 12949913600 | elapsed time per iteration (s): 1.82 | learning rate: 1.779E-04 | global batch size: 512 | lm loss: 1.510457E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.028 | TFLOPs: 42.33 | 31: iteration 12360/ 52452 | consumed samples: 6328320 | consumed tokens: 12960399360 | elapsed time per iteration (s): 1.78 | learning rate: 1.779E-04 | global batch size: 512 | lm loss: 1.514356E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.891 | TFLOPs: 43.21 | 31: iteration 12370/ 52452 | consumed samples: 6333440 | consumed tokens: 12970885120 | elapsed time per iteration (s): 1.75 | learning rate: 1.779E-04 | global batch size: 512 | lm loss: 1.532081E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.777 | TFLOPs: 43.94 | 31: iteration 12380/ 52452 | consumed samples: 6338560 | consumed tokens: 12981370880 | elapsed time per iteration (s): 1.77 | learning rate: 1.778E-04 | global batch size: 512 | lm loss: 1.504462E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.011 | TFLOPs: 43.38 | 31: iteration 12390/ 52452 | consumed samples: 6343680 | consumed tokens: 12991856640 | elapsed time per iteration (s): 1.73 | learning rate: 1.778E-04 | global batch size: 512 | lm loss: 1.509801E+00 | grad norm: 0.200 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.825 | TFLOPs: 44.40 | 31: iteration 12400/ 52452 | consumed samples: 6348800 | consumed tokens: 13002342400 | elapsed time per iteration (s): 1.76 | learning rate: 1.778E-04 | global batch size: 512 | lm loss: 1.534336E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.695 | TFLOPs: 43.63 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12400 | lm loss value: 4.267299E+00 | lm loss PPL: 7.132873E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12400 | lm loss value: 4.220800E+00 | lm loss PPL: 6.808793E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12410/ 52452 | consumed samples: 6353920 | consumed tokens: 13012828160 | elapsed time per iteration (s): 9.20 | learning rate: 1.777E-04 | global batch size: 512 | lm loss: 1.502903E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.636 | TFLOPs: 8.35 | 31: iteration 12420/ 52452 | consumed samples: 6359040 | consumed tokens: 13023313920 | elapsed time per iteration (s): 1.85 | learning rate: 1.777E-04 | global batch size: 512 | lm loss: 1.508943E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.678 | TFLOPs: 41.53 | 31: iteration 12430/ 52452 | consumed samples: 6364160 | consumed tokens: 13033799680 | elapsed time per iteration (s): 1.78 | learning rate: 1.776E-04 | global batch size: 512 | lm loss: 1.517630E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.183 | TFLOPs: 43.25 | 31: iteration 12440/ 52452 | consumed samples: 6369280 | consumed tokens: 13044285440 | elapsed time per iteration (s): 1.74 | learning rate: 1.776E-04 | global batch size: 512 | lm loss: 1.516616E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.146 | TFLOPs: 44.15 | 31: iteration 12450/ 52452 | consumed samples: 6374400 | consumed tokens: 13054771200 | elapsed time per iteration (s): 1.80 | learning rate: 1.776E-04 | global batch size: 512 | lm loss: 1.504535E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.297 | TFLOPs: 42.67 | 31: iteration 12460/ 52452 | consumed samples: 6379520 | consumed tokens: 13065256960 | elapsed time per iteration (s): 1.81 | learning rate: 1.775E-04 | global batch size: 512 | lm loss: 1.519078E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.410 | TFLOPs: 42.39 | 31: iteration 12470/ 52452 | consumed samples: 6384640 | consumed tokens: 13075742720 | elapsed time per iteration (s): 1.78 | learning rate: 1.775E-04 | global batch size: 512 | lm loss: 1.527302E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.381 | TFLOPs: 43.13 | 31: iteration 12480/ 52452 | consumed samples: 6389760 | consumed tokens: 13086228480 | elapsed time per iteration (s): 1.83 | learning rate: 1.775E-04 | global batch size: 512 | lm loss: 1.524105E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.187 | TFLOPs: 42.05 | 31: iteration 12490/ 52452 | consumed samples: 6394880 | consumed tokens: 13096714240 | elapsed time per iteration (s): 1.81 | learning rate: 1.774E-04 | global batch size: 512 | lm loss: 1.518510E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.433 | TFLOPs: 42.39 | 31: iteration 12500/ 52452 | consumed samples: 6400000 | consumed tokens: 13107200000 | elapsed time per iteration (s): 1.77 | learning rate: 1.774E-04 | global batch size: 512 | lm loss: 1.511837E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.526 | TFLOPs: 43.31 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12500 | lm loss value: 4.196791E+00 | lm loss PPL: 6.647269E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12500 | lm loss value: 4.174174E+00 | lm loss PPL: 6.498616E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12510/ 52452 | consumed samples: 6405120 | consumed tokens: 13117685760 | elapsed time per iteration (s): 9.19 | learning rate: 1.774E-04 | global batch size: 512 | lm loss: 1.501749E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.713 | TFLOPs: 8.36 | 31: iteration 12520/ 52452 | consumed samples: 6410240 | consumed tokens: 13128171520 | elapsed time per iteration (s): 1.83 | learning rate: 1.773E-04 | global batch size: 512 | lm loss: 1.503993E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.262 | TFLOPs: 42.07 | 31: iteration 12530/ 52452 | consumed samples: 6415360 | consumed tokens: 13138657280 | elapsed time per iteration (s): 1.76 | learning rate: 1.773E-04 | global batch size: 512 | lm loss: 1.509160E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.613 | TFLOPs: 43.62 | 31: iteration 12540/ 52452 | consumed samples: 6420480 | consumed tokens: 13149143040 | elapsed time per iteration (s): 1.76 | learning rate: 1.772E-04 | global batch size: 512 | lm loss: 1.502453E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.392 | TFLOPs: 43.59 | 31: iteration 12550/ 52452 | consumed samples: 6425600 | consumed tokens: 13159628800 | elapsed time per iteration (s): 1.80 | learning rate: 1.772E-04 | global batch size: 512 | lm loss: 1.522045E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.107 | TFLOPs: 42.64 | 31: iteration 12560/ 52452 | consumed samples: 6430720 | consumed tokens: 13170114560 | elapsed time per iteration (s): 1.82 | learning rate: 1.772E-04 | global batch size: 512 | lm loss: 1.509155E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.724 | TFLOPs: 42.29 | 31: iteration 12570/ 52452 | consumed samples: 6435840 | consumed tokens: 13180600320 | elapsed time per iteration (s): 1.87 | learning rate: 1.771E-04 | global batch size: 512 | lm loss: 1.506025E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.596 | TFLOPs: 41.07 | 31: iteration 12580/ 52452 | consumed samples: 6440960 | consumed tokens: 13191086080 | elapsed time per iteration (s): 1.86 | learning rate: 1.771E-04 | global batch size: 512 | lm loss: 1.506963E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.815 | TFLOPs: 41.25 | 31: iteration 12590/ 52452 | consumed samples: 6446080 | consumed tokens: 13201571840 | elapsed time per iteration (s): 1.78 | learning rate: 1.771E-04 | global batch size: 512 | lm loss: 1.508786E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.698 | TFLOPs: 43.18 | 31: iteration 12600/ 52452 | consumed samples: 6451200 | consumed tokens: 13212057600 | elapsed time per iteration (s): 1.82 | learning rate: 1.770E-04 | global batch size: 512 | lm loss: 1.509711E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.719 | TFLOPs: 42.28 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12600 | lm loss value: 4.069237E+00 | lm loss PPL: 5.851231E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12600 | lm loss value: 4.057358E+00 | lm loss PPL: 5.782133E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12610/ 52452 | consumed samples: 6456320 | consumed tokens: 13222543360 | elapsed time per iteration (s): 9.17 | learning rate: 1.770E-04 | global batch size: 512 | lm loss: 1.507218E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.830 | TFLOPs: 8.38 | 31: iteration 12620/ 52452 | consumed samples: 6461440 | consumed tokens: 13233029120 | elapsed time per iteration (s): 1.87 | learning rate: 1.770E-04 | global batch size: 512 | lm loss: 1.507018E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.848 | TFLOPs: 41.10 | 31: iteration 12630/ 52452 | consumed samples: 6466560 | consumed tokens: 13243514880 | elapsed time per iteration (s): 1.86 | learning rate: 1.769E-04 | global batch size: 512 | lm loss: 1.521450E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.838 | TFLOPs: 41.40 | 31: iteration 12640/ 52452 | consumed samples: 6471680 | consumed tokens: 13254000640 | elapsed time per iteration (s): 1.86 | learning rate: 1.769E-04 | global batch size: 512 | lm loss: 1.530446E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.878 | TFLOPs: 41.41 | 31: iteration 12650/ 52452 | consumed samples: 6476800 | consumed tokens: 13264486400 | elapsed time per iteration (s): 1.85 | learning rate: 1.769E-04 | global batch size: 512 | lm loss: 1.516946E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.269 | TFLOPs: 41.62 | 31: iteration 12660/ 52452 | consumed samples: 6481920 | consumed tokens: 13274972160 | elapsed time per iteration (s): 1.85 | learning rate: 1.768E-04 | global batch size: 512 | lm loss: 1.515673E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.222 | TFLOPs: 41.61 | 31: iteration 12670/ 52452 | consumed samples: 6487040 | consumed tokens: 13285457920 | elapsed time per iteration (s): 1.89 | learning rate: 1.768E-04 | global batch size: 512 | lm loss: 1.514305E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.605 | TFLOPs: 40.77 | 31: iteration 12680/ 52452 | consumed samples: 6492160 | consumed tokens: 13295943680 | elapsed time per iteration (s): 1.79 | learning rate: 1.767E-04 | global batch size: 512 | lm loss: 1.511016E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.768 | TFLOPs: 43.04 | 31: iteration 12690/ 52452 | consumed samples: 6497280 | consumed tokens: 13306429440 | elapsed time per iteration (s): 1.99 | learning rate: 1.767E-04 | global batch size: 512 | lm loss: 1.515761E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 256.888 | TFLOPs: 38.56 | 31: iteration 12700/ 52452 | consumed samples: 6502400 | consumed tokens: 13316915200 | elapsed time per iteration (s): 1.86 | learning rate: 1.767E-04 | global batch size: 512 | lm loss: 1.516706E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.933 | TFLOPs: 41.42 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12700 | lm loss value: 4.070642E+00 | lm loss PPL: 5.859460E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12700 | lm loss value: 4.084880E+00 | lm loss PPL: 5.943483E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12710/ 52452 | consumed samples: 6507520 | consumed tokens: 13327400960 | elapsed time per iteration (s): 9.19 | learning rate: 1.766E-04 | global batch size: 512 | lm loss: 1.504251E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.727 | TFLOPs: 8.36 | 31: iteration 12720/ 52452 | consumed samples: 6512640 | consumed tokens: 13337886720 | elapsed time per iteration (s): 1.84 | learning rate: 1.766E-04 | global batch size: 512 | lm loss: 1.500674E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.537 | TFLOPs: 41.66 | 31: iteration 12730/ 52452 | consumed samples: 6517760 | consumed tokens: 13348372480 | elapsed time per iteration (s): 1.78 | learning rate: 1.766E-04 | global batch size: 512 | lm loss: 1.498174E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.946 | TFLOPs: 43.07 | 31: iteration 12740/ 52452 | consumed samples: 6522880 | consumed tokens: 13358858240 | elapsed time per iteration (s): 1.83 | learning rate: 1.765E-04 | global batch size: 512 | lm loss: 1.526463E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.920 | TFLOPs: 42.01 | 31: iteration 12750/ 52452 | consumed samples: 6528000 | consumed tokens: 13369344000 | elapsed time per iteration (s): 1.79 | learning rate: 1.765E-04 | global batch size: 512 | lm loss: 1.511620E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.168 | TFLOPs: 42.95 | 31: iteration 12760/ 52452 | consumed samples: 6533120 | consumed tokens: 13379829760 | elapsed time per iteration (s): 1.84 | learning rate: 1.764E-04 | global batch size: 512 | lm loss: 1.507499E+00 | grad norm: 0.190 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.589 | TFLOPs: 41.81 | 31: iteration 12770/ 52452 | consumed samples: 6538240 | consumed tokens: 13390315520 | elapsed time per iteration (s): 1.88 | learning rate: 1.764E-04 | global batch size: 512 | lm loss: 1.513721E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.511 | TFLOPs: 40.90 | 31: iteration 12780/ 52452 | consumed samples: 6543360 | consumed tokens: 13400801280 | elapsed time per iteration (s): 1.81 | learning rate: 1.764E-04 | global batch size: 512 | lm loss: 1.511707E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.668 | TFLOPs: 42.43 | 31: iteration 12790/ 52452 | consumed samples: 6548480 | consumed tokens: 13411287040 | elapsed time per iteration (s): 1.79 | learning rate: 1.763E-04 | global batch size: 512 | lm loss: 1.499216E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.040 | TFLOPs: 42.93 | 31: iteration 12800/ 52452 | consumed samples: 6553600 | consumed tokens: 13421772800 | elapsed time per iteration (s): 1.91 | learning rate: 1.763E-04 | global batch size: 512 | lm loss: 1.511183E+00 | grad norm: 0.186 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.027 | TFLOPs: 40.23 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12800 | lm loss value: 4.315019E+00 | lm loss PPL: 7.481502E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12800 | lm loss value: 4.278359E+00 | lm loss PPL: 7.212199E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12810/ 52452 | consumed samples: 6558720 | consumed tokens: 13432258560 | elapsed time per iteration (s): 9.20 | learning rate: 1.763E-04 | global batch size: 512 | lm loss: 1.501463E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.643 | TFLOPs: 8.35 | 31: iteration 12820/ 52452 | consumed samples: 6563840 | consumed tokens: 13442744320 | elapsed time per iteration (s): 1.81 | learning rate: 1.762E-04 | global batch size: 512 | lm loss: 1.502146E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.371 | TFLOPs: 42.53 | 31: iteration 12830/ 52452 | consumed samples: 6568960 | consumed tokens: 13453230080 | elapsed time per iteration (s): 1.78 | learning rate: 1.762E-04 | global batch size: 512 | lm loss: 1.496554E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.313 | TFLOPs: 43.27 | 31: iteration 12840/ 52452 | consumed samples: 6574080 | consumed tokens: 13463715840 | elapsed time per iteration (s): 1.84 | learning rate: 1.762E-04 | global batch size: 512 | lm loss: 1.504117E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.837 | TFLOPs: 41.85 | 31: iteration 12850/ 52452 | consumed samples: 6579200 | consumed tokens: 13474201600 | elapsed time per iteration (s): 1.75 | learning rate: 1.761E-04 | global batch size: 512 | lm loss: 1.498283E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.467 | TFLOPs: 43.90 | 31: iteration 12860/ 52452 | consumed samples: 6584320 | consumed tokens: 13484687360 | elapsed time per iteration (s): 1.81 | learning rate: 1.761E-04 | global batch size: 512 | lm loss: 1.503103E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.317 | TFLOPs: 42.52 | 31: iteration 12870/ 52452 | consumed samples: 6589440 | consumed tokens: 13495173120 | elapsed time per iteration (s): 1.77 | learning rate: 1.760E-04 | global batch size: 512 | lm loss: 1.509767E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.668 | TFLOPs: 43.33 | 31: iteration 12880/ 52452 | consumed samples: 6594560 | consumed tokens: 13505658880 | elapsed time per iteration (s): 1.80 | learning rate: 1.760E-04 | global batch size: 512 | lm loss: 1.510662E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.579 | TFLOPs: 42.71 | 31: iteration 12890/ 52452 | consumed samples: 6599680 | consumed tokens: 13516144640 | elapsed time per iteration (s): 1.80 | learning rate: 1.760E-04 | global batch size: 512 | lm loss: 1.511136E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.848 | TFLOPs: 42.60 | 31: iteration 12900/ 52452 | consumed samples: 6604800 | consumed tokens: 13526630400 | elapsed time per iteration (s): 1.81 | learning rate: 1.759E-04 | global batch size: 512 | lm loss: 1.494169E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.823 | TFLOPs: 42.45 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 12900 | lm loss value: 4.269450E+00 | lm loss PPL: 7.148232E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 12900 | lm loss value: 4.218218E+00 | lm loss PPL: 6.791238E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 12910/ 52452 | consumed samples: 6609920 | consumed tokens: 13537116160 | elapsed time per iteration (s): 9.28 | learning rate: 1.759E-04 | global batch size: 512 | lm loss: 1.489729E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.163 | TFLOPs: 8.28 | 31: iteration 12920/ 52452 | consumed samples: 6615040 | consumed tokens: 13547601920 | elapsed time per iteration (s): 1.81 | learning rate: 1.759E-04 | global batch size: 512 | lm loss: 1.497192E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.694 | TFLOPs: 42.43 | 31: iteration 12930/ 52452 | consumed samples: 6620160 | consumed tokens: 13558087680 | elapsed time per iteration (s): 1.85 | learning rate: 1.758E-04 | global batch size: 512 | lm loss: 1.510340E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.113 | TFLOPs: 41.59 | 31: iteration 12940/ 52452 | consumed samples: 6625280 | consumed tokens: 13568573440 | elapsed time per iteration (s): 1.88 | learning rate: 1.758E-04 | global batch size: 512 | lm loss: 1.509552E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.612 | TFLOPs: 40.92 | 31: iteration 12950/ 52452 | consumed samples: 6630400 | consumed tokens: 13579059200 | elapsed time per iteration (s): 1.87 | learning rate: 1.757E-04 | global batch size: 512 | lm loss: 1.498194E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.401 | TFLOPs: 41.19 | 31: iteration 12960/ 52452 | consumed samples: 6635520 | consumed tokens: 13589544960 | elapsed time per iteration (s): 1.88 | learning rate: 1.757E-04 | global batch size: 512 | lm loss: 1.516413E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.258 | TFLOPs: 40.86 | 31: iteration 12970/ 52452 | consumed samples: 6640640 | consumed tokens: 13600030720 | elapsed time per iteration (s): 1.86 | learning rate: 1.757E-04 | global batch size: 512 | lm loss: 1.514473E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.487 | TFLOPs: 41.35 | 31: iteration 12980/ 52452 | consumed samples: 6645760 | consumed tokens: 13610516480 | elapsed time per iteration (s): 1.83 | learning rate: 1.756E-04 | global batch size: 512 | lm loss: 1.499491E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.221 | TFLOPs: 42.06 | 31: iteration 12990/ 52452 | consumed samples: 6650880 | consumed tokens: 13621002240 | elapsed time per iteration (s): 1.91 | learning rate: 1.756E-04 | global batch size: 512 | lm loss: 1.513880E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.408 | TFLOPs: 40.14 | 31: iteration 13000/ 52452 | consumed samples: 6656000 | consumed tokens: 13631488000 | elapsed time per iteration (s): 1.94 | learning rate: 1.756E-04 | global batch size: 512 | lm loss: 1.504972E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 264.323 | TFLOPs: 39.67 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13000 | lm loss value: 4.244615E+00 | lm loss PPL: 6.972888E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13000 | lm loss value: 4.179721E+00 | lm loss PPL: 6.534764E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13010/ 52452 | consumed samples: 6661120 | consumed tokens: 13641973760 | elapsed time per iteration (s): 9.19 | learning rate: 1.755E-04 | global batch size: 512 | lm loss: 1.510746E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.690 | TFLOPs: 8.36 | 31: iteration 13020/ 52452 | consumed samples: 6666240 | consumed tokens: 13652459520 | elapsed time per iteration (s): 1.88 | learning rate: 1.755E-04 | global batch size: 512 | lm loss: 1.516775E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.996 | TFLOPs: 40.98 | 31: iteration 13030/ 52452 | consumed samples: 6671360 | consumed tokens: 13662945280 | elapsed time per iteration (s): 1.86 | learning rate: 1.754E-04 | global batch size: 512 | lm loss: 1.512034E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.168 | TFLOPs: 41.30 | 31: iteration 13040/ 52452 | consumed samples: 6676480 | consumed tokens: 13673431040 | elapsed time per iteration (s): 1.81 | learning rate: 1.754E-04 | global batch size: 512 | lm loss: 1.501118E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.093 | TFLOPs: 42.49 | 31: iteration 13050/ 52452 | consumed samples: 6681600 | consumed tokens: 13683916800 | elapsed time per iteration (s): 1.79 | learning rate: 1.754E-04 | global batch size: 512 | lm loss: 1.500856E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.647 | TFLOPs: 43.02 | 31: iteration 13060/ 52452 | consumed samples: 6686720 | consumed tokens: 13694402560 | elapsed time per iteration (s): 1.83 | learning rate: 1.753E-04 | global batch size: 512 | lm loss: 1.498729E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.974 | TFLOPs: 42.02 | 31: iteration 13070/ 52452 | consumed samples: 6691840 | consumed tokens: 13704888320 | elapsed time per iteration (s): 1.80 | learning rate: 1.753E-04 | global batch size: 512 | lm loss: 1.498077E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.444 | TFLOPs: 42.69 | 31: iteration 13080/ 52452 | consumed samples: 6696960 | consumed tokens: 13715374080 | elapsed time per iteration (s): 1.86 | learning rate: 1.753E-04 | global batch size: 512 | lm loss: 1.501296E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.511 | TFLOPs: 41.35 | 31: iteration 13090/ 52452 | consumed samples: 6702080 | consumed tokens: 13725859840 | elapsed time per iteration (s): 1.77 | learning rate: 1.752E-04 | global batch size: 512 | lm loss: 1.502312E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.499 | TFLOPs: 43.45 | 31: iteration 13100/ 52452 | consumed samples: 6707200 | consumed tokens: 13736345600 | elapsed time per iteration (s): 1.82 | learning rate: 1.752E-04 | global batch size: 512 | lm loss: 1.495673E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.043 | TFLOPs: 42.18 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13100 | lm loss value: 4.049468E+00 | lm loss PPL: 5.736690E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13100 | lm loss value: 4.037417E+00 | lm loss PPL: 5.667977E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13110/ 52452 | consumed samples: 6712320 | consumed tokens: 13746831360 | elapsed time per iteration (s): 9.20 | learning rate: 1.751E-04 | global batch size: 512 | lm loss: 1.496192E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.667 | TFLOPs: 8.36 | 31: iteration 13120/ 52452 | consumed samples: 6717440 | consumed tokens: 13757317120 | elapsed time per iteration (s): 1.85 | learning rate: 1.751E-04 | global batch size: 512 | lm loss: 1.509992E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.329 | TFLOPs: 41.63 | 31: iteration 13130/ 52452 | consumed samples: 6722560 | consumed tokens: 13767802880 | elapsed time per iteration (s): 1.82 | learning rate: 1.751E-04 | global batch size: 512 | lm loss: 1.493477E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.810 | TFLOPs: 42.15 | 31: iteration 13140/ 52452 | consumed samples: 6727680 | consumed tokens: 13778288640 | elapsed time per iteration (s): 1.75 | learning rate: 1.750E-04 | global batch size: 512 | lm loss: 1.505994E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.273 | TFLOPs: 43.87 | 31: iteration 13150/ 52452 | consumed samples: 6732800 | consumed tokens: 13788774400 | elapsed time per iteration (s): 1.87 | learning rate: 1.750E-04 | global batch size: 512 | lm loss: 1.506168E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.577 | TFLOPs: 41.06 | 31: iteration 13160/ 52452 | consumed samples: 6737920 | consumed tokens: 13799260160 | elapsed time per iteration (s): 1.82 | learning rate: 1.750E-04 | global batch size: 512 | lm loss: 1.495496E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.217 | TFLOPs: 42.21 | 31: iteration 13170/ 52452 | consumed samples: 6743040 | consumed tokens: 13809745920 | elapsed time per iteration (s): 1.83 | learning rate: 1.749E-04 | global batch size: 512 | lm loss: 1.507514E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.313 | TFLOPs: 42.07 | 31: iteration 13180/ 52452 | consumed samples: 6748160 | consumed tokens: 13820231680 | elapsed time per iteration (s): 1.77 | learning rate: 1.749E-04 | global batch size: 512 | lm loss: 1.506247E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.203 | TFLOPs: 43.41 | 31: iteration 13190/ 52452 | consumed samples: 6753280 | consumed tokens: 13830717440 | elapsed time per iteration (s): 1.85 | learning rate: 1.748E-04 | global batch size: 512 | lm loss: 1.502866E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.035 | TFLOPs: 41.43 | 31: iteration 13200/ 52452 | consumed samples: 6758400 | consumed tokens: 13841203200 | elapsed time per iteration (s): 1.74 | learning rate: 1.748E-04 | global batch size: 512 | lm loss: 1.493625E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.430 | TFLOPs: 44.04 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13200 | lm loss value: 4.253604E+00 | lm loss PPL: 7.035856E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13200 | lm loss value: 4.192797E+00 | lm loss PPL: 6.620770E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13210/ 52452 | consumed samples: 6763520 | consumed tokens: 13851688960 | elapsed time per iteration (s): 9.17 | learning rate: 1.748E-04 | global batch size: 512 | lm loss: 1.495434E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.834 | TFLOPs: 8.38 | 31: iteration 13220/ 52452 | consumed samples: 6768640 | consumed tokens: 13862174720 | elapsed time per iteration (s): 1.78 | learning rate: 1.747E-04 | global batch size: 512 | lm loss: 1.502332E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.151 | TFLOPs: 43.25 | 31: iteration 13230/ 52452 | consumed samples: 6773760 | consumed tokens: 13872660480 | elapsed time per iteration (s): 1.75 | learning rate: 1.747E-04 | global batch size: 512 | lm loss: 1.499720E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.455 | TFLOPs: 43.90 | 31: iteration 13240/ 52452 | consumed samples: 6778880 | consumed tokens: 13883146240 | elapsed time per iteration (s): 1.88 | learning rate: 1.747E-04 | global batch size: 512 | lm loss: 1.495571E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.790 | TFLOPs: 40.94 | 31: iteration 13250/ 52452 | consumed samples: 6784000 | consumed tokens: 13893632000 | elapsed time per iteration (s): 1.80 | learning rate: 1.746E-04 | global batch size: 512 | lm loss: 1.492271E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.869 | TFLOPs: 42.76 | 31: iteration 13260/ 52452 | consumed samples: 6789120 | consumed tokens: 13904117760 | elapsed time per iteration (s): 1.76 | learning rate: 1.746E-04 | global batch size: 512 | lm loss: 1.495740E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.892 | TFLOPs: 43.66 | 31: iteration 13270/ 52452 | consumed samples: 6794240 | consumed tokens: 13914603520 | elapsed time per iteration (s): 1.81 | learning rate: 1.745E-04 | global batch size: 512 | lm loss: 1.502286E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.274 | TFLOPs: 42.37 | 31: iteration 13280/ 52452 | consumed samples: 6799360 | consumed tokens: 13925089280 | elapsed time per iteration (s): 1.86 | learning rate: 1.745E-04 | global batch size: 512 | lm loss: 1.497894E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.306 | TFLOPs: 41.32 | 31: iteration 13290/ 52452 | consumed samples: 6804480 | consumed tokens: 13935575040 | elapsed time per iteration (s): 1.80 | learning rate: 1.745E-04 | global batch size: 512 | lm loss: 1.501711E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.407 | TFLOPs: 42.69 | 31: iteration 13300/ 52452 | consumed samples: 6809600 | consumed tokens: 13946060800 | elapsed time per iteration (s): 1.75 | learning rate: 1.744E-04 | global batch size: 512 | lm loss: 1.483171E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.925 | TFLOPs: 43.82 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13300 | lm loss value: 4.062509E+00 | lm loss PPL: 5.811993E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13300 | lm loss value: 4.073118E+00 | lm loss PPL: 5.873981E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13310/ 52452 | consumed samples: 6814720 | consumed tokens: 13956546560 | elapsed time per iteration (s): 9.18 | learning rate: 1.744E-04 | global batch size: 512 | lm loss: 1.497989E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.774 | TFLOPs: 8.37 | 31: iteration 13320/ 52452 | consumed samples: 6819840 | consumed tokens: 13967032320 | elapsed time per iteration (s): 1.78 | learning rate: 1.744E-04 | global batch size: 512 | lm loss: 1.504283E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.093 | TFLOPs: 43.24 | 31: iteration 13330/ 52452 | consumed samples: 6824960 | consumed tokens: 13977518080 | elapsed time per iteration (s): 1.78 | learning rate: 1.743E-04 | global batch size: 512 | lm loss: 1.497374E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.065 | TFLOPs: 43.09 | 31: iteration 13340/ 52452 | consumed samples: 6830080 | consumed tokens: 13988003840 | elapsed time per iteration (s): 1.75 | learning rate: 1.743E-04 | global batch size: 512 | lm loss: 1.515757E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.009 | TFLOPs: 43.98 | 31: iteration 13350/ 52452 | consumed samples: 6835200 | consumed tokens: 13998489600 | elapsed time per iteration (s): 1.77 | learning rate: 1.742E-04 | global batch size: 512 | lm loss: 1.503798E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.843 | TFLOPs: 43.50 | 31: iteration 13360/ 52452 | consumed samples: 6840320 | consumed tokens: 14008975360 | elapsed time per iteration (s): 1.78 | learning rate: 1.742E-04 | global batch size: 512 | lm loss: 1.490927E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.896 | TFLOPs: 43.06 | 31: iteration 13370/ 52452 | consumed samples: 6845440 | consumed tokens: 14019461120 | elapsed time per iteration (s): 1.76 | learning rate: 1.742E-04 | global batch size: 512 | lm loss: 1.489874E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.202 | TFLOPs: 43.56 | 31: iteration 13380/ 52452 | consumed samples: 6850560 | consumed tokens: 14029946880 | elapsed time per iteration (s): 1.82 | learning rate: 1.741E-04 | global batch size: 512 | lm loss: 1.510918E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.796 | TFLOPs: 42.30 | 31: iteration 13390/ 52452 | consumed samples: 6855680 | consumed tokens: 14040432640 | elapsed time per iteration (s): 1.80 | learning rate: 1.741E-04 | global batch size: 512 | lm loss: 1.492352E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.101 | TFLOPs: 42.64 | 31: iteration 13400/ 52452 | consumed samples: 6860800 | consumed tokens: 14050918400 | elapsed time per iteration (s): 1.80 | learning rate: 1.740E-04 | global batch size: 512 | lm loss: 1.487860E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.742 | TFLOPs: 42.74 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13400 | lm loss value: 4.099704E+00 | lm loss PPL: 6.032242E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13400 | lm loss value: 4.089785E+00 | lm loss PPL: 5.972703E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13410/ 52452 | consumed samples: 6865920 | consumed tokens: 14061404160 | elapsed time per iteration (s): 9.18 | learning rate: 1.740E-04 | global batch size: 512 | lm loss: 1.496966E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.744 | TFLOPs: 8.37 | 31: iteration 13420/ 52452 | consumed samples: 6871040 | consumed tokens: 14071889920 | elapsed time per iteration (s): 1.90 | learning rate: 1.740E-04 | global batch size: 512 | lm loss: 1.496672E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.154 | TFLOPs: 40.55 | 31: iteration 13430/ 52452 | consumed samples: 6876160 | consumed tokens: 14082375680 | elapsed time per iteration (s): 1.76 | learning rate: 1.739E-04 | global batch size: 512 | lm loss: 1.498861E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.136 | TFLOPs: 43.70 | 31: iteration 13440/ 52452 | consumed samples: 6881280 | consumed tokens: 14092861440 | elapsed time per iteration (s): 1.81 | learning rate: 1.739E-04 | global batch size: 512 | lm loss: 1.500651E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.980 | TFLOPs: 42.47 | 31: iteration 13450/ 52452 | consumed samples: 6886400 | consumed tokens: 14103347200 | elapsed time per iteration (s): 1.85 | learning rate: 1.739E-04 | global batch size: 512 | lm loss: 1.495104E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.690 | TFLOPs: 41.53 | 31: iteration 13460/ 52452 | consumed samples: 6891520 | consumed tokens: 14113832960 | elapsed time per iteration (s): 1.84 | learning rate: 1.738E-04 | global batch size: 512 | lm loss: 1.478642E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.605 | TFLOPs: 41.67 | 31: iteration 13470/ 52452 | consumed samples: 6896640 | consumed tokens: 14124318720 | elapsed time per iteration (s): 1.76 | learning rate: 1.738E-04 | global batch size: 512 | lm loss: 1.494325E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.781 | TFLOPs: 43.64 | 31: iteration 13480/ 52452 | consumed samples: 6901760 | consumed tokens: 14134804480 | elapsed time per iteration (s): 1.76 | learning rate: 1.737E-04 | global batch size: 512 | lm loss: 1.487527E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.643 | TFLOPs: 43.62 | 31: iteration 13490/ 52452 | consumed samples: 6906880 | consumed tokens: 14145290240 | elapsed time per iteration (s): 1.79 | learning rate: 1.737E-04 | global batch size: 512 | lm loss: 1.492851E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.685 | TFLOPs: 42.88 | 31: iteration 13500/ 52452 | consumed samples: 6912000 | consumed tokens: 14155776000 | elapsed time per iteration (s): 1.76 | learning rate: 1.737E-04 | global batch size: 512 | lm loss: 1.481405E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.253 | TFLOPs: 43.57 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13500 | lm loss value: 4.099172E+00 | lm loss PPL: 6.029032E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13500 | lm loss value: 4.092809E+00 | lm loss PPL: 5.990792E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13510/ 52452 | consumed samples: 6917120 | consumed tokens: 14166261760 | elapsed time per iteration (s): 9.18 | learning rate: 1.736E-04 | global batch size: 512 | lm loss: 1.493272E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.769 | TFLOPs: 8.37 | 31: iteration 13520/ 52452 | consumed samples: 6922240 | consumed tokens: 14176747520 | elapsed time per iteration (s): 1.86 | learning rate: 1.736E-04 | global batch size: 512 | lm loss: 1.485003E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.587 | TFLOPs: 41.36 | 31: iteration 13530/ 52452 | consumed samples: 6927360 | consumed tokens: 14187233280 | elapsed time per iteration (s): 1.74 | learning rate: 1.735E-04 | global batch size: 512 | lm loss: 1.495647E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.096 | TFLOPs: 44.14 | 31: iteration 13540/ 52452 | consumed samples: 6932480 | consumed tokens: 14197719040 | elapsed time per iteration (s): 1.80 | learning rate: 1.735E-04 | global batch size: 512 | lm loss: 1.496407E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.139 | TFLOPs: 42.80 | 31: iteration 13550/ 52452 | consumed samples: 6937600 | consumed tokens: 14208204800 | elapsed time per iteration (s): 1.79 | learning rate: 1.735E-04 | global batch size: 512 | lm loss: 1.497499E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.477 | TFLOPs: 42.85 | 31: iteration 13560/ 52452 | consumed samples: 6942720 | consumed tokens: 14218690560 | elapsed time per iteration (s): 1.77 | learning rate: 1.734E-04 | global batch size: 512 | lm loss: 1.492921E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.960 | TFLOPs: 43.52 | 31: iteration 13570/ 52452 | consumed samples: 6947840 | consumed tokens: 14229176320 | elapsed time per iteration (s): 1.74 | learning rate: 1.734E-04 | global batch size: 512 | lm loss: 1.496049E+00 | grad norm: 3.100 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.231 | TFLOPs: 44.16 | 31: iteration 13580/ 52452 | consumed samples: 6952960 | consumed tokens: 14239662080 | elapsed time per iteration (s): 1.77 | learning rate: 1.734E-04 | global batch size: 512 | lm loss: 1.497679E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.202 | TFLOPs: 43.41 | 31: iteration 13590/ 52452 | consumed samples: 6958080 | consumed tokens: 14250147840 | elapsed time per iteration (s): 1.82 | learning rate: 1.733E-04 | global batch size: 512 | lm loss: 1.495784E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.974 | TFLOPs: 42.32 | 31: iteration 13600/ 52452 | consumed samples: 6963200 | consumed tokens: 14260633600 | elapsed time per iteration (s): 1.78 | learning rate: 1.733E-04 | global batch size: 512 | lm loss: 1.483063E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.999 | TFLOPs: 43.08 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13600 | lm loss value: 4.343745E+00 | lm loss PPL: 7.699537E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13600 | lm loss value: 4.276152E+00 | lm loss PPL: 7.196297E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13610/ 52452 | consumed samples: 6968320 | consumed tokens: 14271119360 | elapsed time per iteration (s): 9.12 | learning rate: 1.732E-04 | global batch size: 512 | lm loss: 1.490518E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.155 | TFLOPs: 8.43 | 31: iteration 13620/ 52452 | consumed samples: 6973440 | consumed tokens: 14281605120 | elapsed time per iteration (s): 1.75 | learning rate: 1.732E-04 | global batch size: 512 | lm loss: 1.487209E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.577 | TFLOPs: 43.91 | 31: iteration 13630/ 52452 | consumed samples: 6978560 | consumed tokens: 14292090880 | elapsed time per iteration (s): 1.77 | learning rate: 1.732E-04 | global batch size: 512 | lm loss: 1.492133E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.645 | TFLOPs: 43.32 | 31: iteration 13640/ 52452 | consumed samples: 6983680 | consumed tokens: 14302576640 | elapsed time per iteration (s): 1.73 | learning rate: 1.731E-04 | global batch size: 512 | lm loss: 1.490305E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.974 | TFLOPs: 44.42 | 31: iteration 13650/ 52452 | consumed samples: 6988800 | consumed tokens: 14313062400 | elapsed time per iteration (s): 1.76 | learning rate: 1.731E-04 | global batch size: 512 | lm loss: 1.504885E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.613 | TFLOPs: 43.77 | 31: iteration 13660/ 52452 | consumed samples: 6993920 | consumed tokens: 14323548160 | elapsed time per iteration (s): 1.76 | learning rate: 1.730E-04 | global batch size: 512 | lm loss: 1.488545E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.527 | TFLOPs: 43.61 | 31: iteration 13670/ 52452 | consumed samples: 6999040 | consumed tokens: 14334033920 | elapsed time per iteration (s): 1.76 | learning rate: 1.730E-04 | global batch size: 512 | lm loss: 1.493188E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.603 | TFLOPs: 43.62 | 31: iteration 13680/ 52452 | consumed samples: 7004160 | consumed tokens: 14344519680 | elapsed time per iteration (s): 1.77 | learning rate: 1.730E-04 | global batch size: 512 | lm loss: 1.486521E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.783 | TFLOPs: 43.49 | 31: iteration 13690/ 52452 | consumed samples: 7009280 | consumed tokens: 14355005440 | elapsed time per iteration (s): 1.76 | learning rate: 1.729E-04 | global batch size: 512 | lm loss: 1.485477E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.264 | TFLOPs: 43.57 | 31: iteration 13700/ 52452 | consumed samples: 7014400 | consumed tokens: 14365491200 | elapsed time per iteration (s): 1.75 | learning rate: 1.729E-04 | global batch size: 512 | lm loss: 1.479793E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.029 | TFLOPs: 43.98 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13700 | lm loss value: 4.161635E+00 | lm loss PPL: 6.417636E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13700 | lm loss value: 4.152108E+00 | lm loss PPL: 6.356787E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13710/ 52452 | consumed samples: 7019520 | consumed tokens: 14375976960 | elapsed time per iteration (s): 9.24 | learning rate: 1.729E-04 | global batch size: 512 | lm loss: 1.485519E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.390 | TFLOPs: 8.31 | 31: iteration 13720/ 52452 | consumed samples: 7024640 | consumed tokens: 14386462720 | elapsed time per iteration (s): 1.82 | learning rate: 1.728E-04 | global batch size: 512 | lm loss: 1.483219E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.621 | TFLOPs: 42.12 | 31: iteration 13730/ 52452 | consumed samples: 7029760 | consumed tokens: 14396948480 | elapsed time per iteration (s): 1.79 | learning rate: 1.728E-04 | global batch size: 512 | lm loss: 1.496321E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.290 | TFLOPs: 42.97 | 31: iteration 13740/ 52452 | consumed samples: 7034880 | consumed tokens: 14407434240 | elapsed time per iteration (s): 1.83 | learning rate: 1.727E-04 | global batch size: 512 | lm loss: 1.495686E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.194 | TFLOPs: 41.91 | 31: iteration 13750/ 52452 | consumed samples: 7040000 | consumed tokens: 14417920000 | elapsed time per iteration (s): 1.86 | learning rate: 1.727E-04 | global batch size: 512 | lm loss: 1.491920E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.551 | TFLOPs: 41.21 | 31: iteration 13760/ 52452 | consumed samples: 7045120 | consumed tokens: 14428405760 | elapsed time per iteration (s): 1.83 | learning rate: 1.727E-04 | global batch size: 512 | lm loss: 1.500186E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.548 | TFLOPs: 42.11 | 31: iteration 13770/ 52452 | consumed samples: 7050240 | consumed tokens: 14438891520 | elapsed time per iteration (s): 1.81 | learning rate: 1.726E-04 | global batch size: 512 | lm loss: 1.499221E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.267 | TFLOPs: 42.37 | 31: iteration 13780/ 52452 | consumed samples: 7055360 | consumed tokens: 14449377280 | elapsed time per iteration (s): 1.79 | learning rate: 1.726E-04 | global batch size: 512 | lm loss: 1.488336E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.757 | TFLOPs: 43.04 | 31: iteration 13790/ 52452 | consumed samples: 7060480 | consumed tokens: 14459863040 | elapsed time per iteration (s): 1.89 | learning rate: 1.725E-04 | global batch size: 512 | lm loss: 1.486407E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.475 | TFLOPs: 40.75 | 31: iteration 13800/ 52452 | consumed samples: 7065600 | consumed tokens: 14470348800 | elapsed time per iteration (s): 1.78 | learning rate: 1.725E-04 | global batch size: 512 | lm loss: 1.503858E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.264 | TFLOPs: 43.12 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13800 | lm loss value: 4.039279E+00 | lm loss PPL: 5.678538E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13800 | lm loss value: 4.063271E+00 | lm loss PPL: 5.816426E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13810/ 52452 | consumed samples: 7070720 | consumed tokens: 14480834560 | elapsed time per iteration (s): 9.27 | learning rate: 1.725E-04 | global batch size: 512 | lm loss: 1.490295E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.226 | TFLOPs: 8.29 | 31: iteration 13820/ 52452 | consumed samples: 7075840 | consumed tokens: 14491320320 | elapsed time per iteration (s): 1.76 | learning rate: 1.724E-04 | global batch size: 512 | lm loss: 1.483867E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.389 | TFLOPs: 43.59 | 31: iteration 13830/ 52452 | consumed samples: 7080960 | consumed tokens: 14501806080 | elapsed time per iteration (s): 1.76 | learning rate: 1.724E-04 | global batch size: 512 | lm loss: 1.501903E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.647 | TFLOPs: 43.77 | 31: iteration 13840/ 52452 | consumed samples: 7086080 | consumed tokens: 14512291840 | elapsed time per iteration (s): 1.85 | learning rate: 1.723E-04 | global batch size: 512 | lm loss: 1.488697E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.366 | TFLOPs: 41.63 | 31: iteration 13850/ 52452 | consumed samples: 7091200 | consumed tokens: 14522777600 | elapsed time per iteration (s): 1.77 | learning rate: 1.723E-04 | global batch size: 512 | lm loss: 1.486655E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.431 | TFLOPs: 43.44 | 31: iteration 13860/ 52452 | consumed samples: 7096320 | consumed tokens: 14533263360 | elapsed time per iteration (s): 1.80 | learning rate: 1.723E-04 | global batch size: 512 | lm loss: 1.488400E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.588 | TFLOPs: 42.72 | 31: iteration 13870/ 52452 | consumed samples: 7101440 | consumed tokens: 14543749120 | elapsed time per iteration (s): 1.78 | learning rate: 1.722E-04 | global batch size: 512 | lm loss: 1.482482E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.839 | TFLOPs: 43.05 | 31: iteration 13880/ 52452 | consumed samples: 7106560 | consumed tokens: 14554234880 | elapsed time per iteration (s): 1.81 | learning rate: 1.722E-04 | global batch size: 512 | lm loss: 1.489997E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.193 | TFLOPs: 42.36 | 31: iteration 13890/ 52452 | consumed samples: 7111680 | consumed tokens: 14564720640 | elapsed time per iteration (s): 1.82 | learning rate: 1.721E-04 | global batch size: 512 | lm loss: 1.488558E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.659 | TFLOPs: 42.28 | 31: iteration 13900/ 52452 | consumed samples: 7116800 | consumed tokens: 14575206400 | elapsed time per iteration (s): 1.80 | learning rate: 1.721E-04 | global batch size: 512 | lm loss: 1.494274E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.422 | TFLOPs: 42.69 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 13900 | lm loss value: 4.126352E+00 | lm loss PPL: 6.195150E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 13900 | lm loss value: 4.097576E+00 | lm loss PPL: 6.019421E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 13910/ 52452 | consumed samples: 7121920 | consumed tokens: 14585692160 | elapsed time per iteration (s): 9.17 | learning rate: 1.721E-04 | global batch size: 512 | lm loss: 1.484882E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.852 | TFLOPs: 8.38 | 31: iteration 13920/ 52452 | consumed samples: 7127040 | consumed tokens: 14596177920 | elapsed time per iteration (s): 1.88 | learning rate: 1.720E-04 | global batch size: 512 | lm loss: 1.483157E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.974 | TFLOPs: 40.82 | 31: iteration 13930/ 52452 | consumed samples: 7132160 | consumed tokens: 14606663680 | elapsed time per iteration (s): 1.87 | learning rate: 1.720E-04 | global batch size: 512 | lm loss: 1.479561E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.302 | TFLOPs: 41.02 | 31: iteration 13940/ 52452 | consumed samples: 7137280 | consumed tokens: 14617149440 | elapsed time per iteration (s): 1.76 | learning rate: 1.719E-04 | global batch size: 512 | lm loss: 1.485569E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.779 | TFLOPs: 43.64 | 31: iteration 13950/ 52452 | consumed samples: 7142400 | consumed tokens: 14627635200 | elapsed time per iteration (s): 1.82 | learning rate: 1.719E-04 | global batch size: 512 | lm loss: 1.488549E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.766 | TFLOPs: 42.29 | 31: iteration 13960/ 52452 | consumed samples: 7147520 | consumed tokens: 14638120960 | elapsed time per iteration (s): 1.78 | learning rate: 1.719E-04 | global batch size: 512 | lm loss: 1.471734E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.983 | TFLOPs: 43.07 | 31: iteration 13970/ 52452 | consumed samples: 7152640 | consumed tokens: 14648606720 | elapsed time per iteration (s): 1.80 | learning rate: 1.718E-04 | global batch size: 512 | lm loss: 1.482625E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.609 | TFLOPs: 42.72 | 31: iteration 13980/ 52452 | consumed samples: 7157760 | consumed tokens: 14659092480 | elapsed time per iteration (s): 1.77 | learning rate: 1.718E-04 | global batch size: 512 | lm loss: 1.482640E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.046 | TFLOPs: 43.38 | 31: iteration 13990/ 52452 | consumed samples: 7162880 | consumed tokens: 14669578240 | elapsed time per iteration (s): 1.77 | learning rate: 1.718E-04 | global batch size: 512 | lm loss: 1.477357E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.822 | TFLOPs: 43.35 | 0: [2022-12-25 18:53:07,281] [INFO] [logging.py:68:log_dist] [Rank 0] step=14000, skipped=0, lr=[0.00017171112893629768, 0.00017171112893629768, 0.00017171112893629768], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 14000/ 52452 | consumed samples: 7168000 | consumed tokens: 14680064000 | elapsed time per iteration (s): 1.75 | learning rate: 1.717E-04 | global batch size: 512 | lm loss: 1.479162E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.639 | TFLOPs: 43.92 | 0: steps: 14000 loss: 1.5000 iter time (s): 1.805 samples/sec: 283.589 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14000 | lm loss value: 4.124949E+00 | lm loss PPL: 6.186468E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14000 | lm loss value: 4.098033E+00 | lm loss PPL: 6.022171E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14010/ 52452 | consumed samples: 7173120 | consumed tokens: 14690549760 | elapsed time per iteration (s): 9.24 | learning rate: 1.717E-04 | global batch size: 512 | lm loss: 1.481164E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.393 | TFLOPs: 8.31 | 31: iteration 14020/ 52452 | consumed samples: 7178240 | consumed tokens: 14701035520 | elapsed time per iteration (s): 1.79 | learning rate: 1.716E-04 | global batch size: 512 | lm loss: 1.487118E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.866 | TFLOPs: 42.91 | 31: iteration 14030/ 52452 | consumed samples: 7183360 | consumed tokens: 14711521280 | elapsed time per iteration (s): 1.83 | learning rate: 1.716E-04 | global batch size: 512 | lm loss: 1.481483E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.182 | TFLOPs: 41.90 | 31: iteration 14040/ 52452 | consumed samples: 7188480 | consumed tokens: 14722007040 | elapsed time per iteration (s): 1.77 | learning rate: 1.716E-04 | global batch size: 512 | lm loss: 1.490210E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.264 | TFLOPs: 43.42 | 31: iteration 14050/ 52452 | consumed samples: 7193600 | consumed tokens: 14732492800 | elapsed time per iteration (s): 1.79 | learning rate: 1.715E-04 | global batch size: 512 | lm loss: 1.477724E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.035 | TFLOPs: 42.93 | 31: iteration 14060/ 52452 | consumed samples: 7198720 | consumed tokens: 14742978560 | elapsed time per iteration (s): 1.76 | learning rate: 1.715E-04 | global batch size: 512 | lm loss: 1.483339E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.764 | TFLOPs: 43.64 | 31: iteration 14070/ 52452 | consumed samples: 7203840 | consumed tokens: 14753464320 | elapsed time per iteration (s): 1.78 | learning rate: 1.714E-04 | global batch size: 512 | lm loss: 1.477997E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.304 | TFLOPs: 43.27 | 31: iteration 14080/ 52452 | consumed samples: 7208960 | consumed tokens: 14763950080 | elapsed time per iteration (s): 1.84 | learning rate: 1.714E-04 | global batch size: 512 | lm loss: 1.468316E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.613 | TFLOPs: 41.67 | 31: iteration 14090/ 52452 | consumed samples: 7214080 | consumed tokens: 14774435840 | elapsed time per iteration (s): 1.87 | learning rate: 1.714E-04 | global batch size: 512 | lm loss: 1.473152E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.318 | TFLOPs: 41.02 | 31: iteration 14100/ 52452 | consumed samples: 7219200 | consumed tokens: 14784921600 | elapsed time per iteration (s): 1.77 | learning rate: 1.713E-04 | global batch size: 512 | lm loss: 1.468101E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.010 | TFLOPs: 43.53 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14100 | lm loss value: 4.018673E+00 | lm loss PPL: 5.562724E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14100 | lm loss value: 4.026964E+00 | lm loss PPL: 5.609035E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14110/ 52452 | consumed samples: 7224320 | consumed tokens: 14795407360 | elapsed time per iteration (s): 9.13 | learning rate: 1.713E-04 | global batch size: 512 | lm loss: 1.480424E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.070 | TFLOPs: 8.42 | 31: iteration 14120/ 52452 | consumed samples: 7229440 | consumed tokens: 14805893120 | elapsed time per iteration (s): 1.89 | learning rate: 1.712E-04 | global batch size: 512 | lm loss: 1.475764E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.473 | TFLOPs: 40.75 | 31: iteration 14130/ 52452 | consumed samples: 7234560 | consumed tokens: 14816378880 | elapsed time per iteration (s): 1.81 | learning rate: 1.712E-04 | global batch size: 512 | lm loss: 1.474847E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.330 | TFLOPs: 42.38 | 31: iteration 14140/ 52452 | consumed samples: 7239680 | consumed tokens: 14826864640 | elapsed time per iteration (s): 1.81 | learning rate: 1.712E-04 | global batch size: 512 | lm loss: 1.486782E+00 | grad norm: 0.195 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.928 | TFLOPs: 42.47 | 31: iteration 14150/ 52452 | consumed samples: 7244800 | consumed tokens: 14837350400 | elapsed time per iteration (s): 1.84 | learning rate: 1.711E-04 | global batch size: 512 | lm loss: 1.482998E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.012 | TFLOPs: 41.73 | 31: iteration 14160/ 52452 | consumed samples: 7249920 | consumed tokens: 14847836160 | elapsed time per iteration (s): 1.77 | learning rate: 1.711E-04 | global batch size: 512 | lm loss: 1.490451E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.093 | TFLOPs: 43.39 | 31: iteration 14170/ 52452 | consumed samples: 7255040 | consumed tokens: 14858321920 | elapsed time per iteration (s): 1.77 | learning rate: 1.710E-04 | global batch size: 512 | lm loss: 1.483544E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.159 | TFLOPs: 43.40 | 31: iteration 14180/ 52452 | consumed samples: 7260160 | consumed tokens: 14868807680 | elapsed time per iteration (s): 1.75 | learning rate: 1.710E-04 | global batch size: 512 | lm loss: 1.480252E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.945 | TFLOPs: 43.97 | 31: iteration 14190/ 52452 | consumed samples: 7265280 | consumed tokens: 14879293440 | elapsed time per iteration (s): 1.82 | learning rate: 1.710E-04 | global batch size: 512 | lm loss: 1.485617E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.723 | TFLOPs: 42.29 | 31: iteration 14200/ 52452 | consumed samples: 7270400 | consumed tokens: 14889779200 | elapsed time per iteration (s): 1.78 | learning rate: 1.709E-04 | global batch size: 512 | lm loss: 1.475798E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.622 | TFLOPs: 43.17 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14200 | lm loss value: 4.195465E+00 | lm loss PPL: 6.638457E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14200 | lm loss value: 4.175059E+00 | lm loss PPL: 6.504367E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14210/ 52452 | consumed samples: 7275520 | consumed tokens: 14900264960 | elapsed time per iteration (s): 9.14 | learning rate: 1.709E-04 | global batch size: 512 | lm loss: 1.475971E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.999 | TFLOPs: 8.41 | 31: iteration 14220/ 52452 | consumed samples: 7280640 | consumed tokens: 14910750720 | elapsed time per iteration (s): 1.80 | learning rate: 1.708E-04 | global batch size: 512 | lm loss: 1.465506E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.273 | TFLOPs: 42.67 | 31: iteration 14230/ 52452 | consumed samples: 7285760 | consumed tokens: 14921236480 | elapsed time per iteration (s): 1.82 | learning rate: 1.708E-04 | global batch size: 512 | lm loss: 1.478170E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.577 | TFLOPs: 42.26 | 31: iteration 14240/ 52452 | consumed samples: 7290880 | consumed tokens: 14931722240 | elapsed time per iteration (s): 1.78 | learning rate: 1.708E-04 | global batch size: 512 | lm loss: 1.476296E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.413 | TFLOPs: 43.29 | 31: iteration 14250/ 52452 | consumed samples: 7296000 | consumed tokens: 14942208000 | elapsed time per iteration (s): 1.81 | learning rate: 1.707E-04 | global batch size: 512 | lm loss: 1.480669E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.645 | TFLOPs: 42.57 | 31: iteration 14260/ 52452 | consumed samples: 7301120 | consumed tokens: 14952693760 | elapsed time per iteration (s): 1.77 | learning rate: 1.707E-04 | global batch size: 512 | lm loss: 1.478361E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.805 | TFLOPs: 43.50 | 31: iteration 14270/ 52452 | consumed samples: 7306240 | consumed tokens: 14963179520 | elapsed time per iteration (s): 1.82 | learning rate: 1.706E-04 | global batch size: 512 | lm loss: 1.469818E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.997 | TFLOPs: 42.18 | 31: iteration 14280/ 52452 | consumed samples: 7311360 | consumed tokens: 14973665280 | elapsed time per iteration (s): 1.78 | learning rate: 1.706E-04 | global batch size: 512 | lm loss: 1.488728E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.563 | TFLOPs: 43.16 | 31: iteration 14290/ 52452 | consumed samples: 7316480 | consumed tokens: 14984151040 | elapsed time per iteration (s): 1.74 | learning rate: 1.706E-04 | global batch size: 512 | lm loss: 1.484186E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.180 | TFLOPs: 44.15 | 31: iteration 14300/ 52452 | consumed samples: 7321600 | consumed tokens: 14994636800 | elapsed time per iteration (s): 1.76 | learning rate: 1.705E-04 | global batch size: 512 | lm loss: 1.477459E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.667 | TFLOPs: 43.63 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14300 | lm loss value: 4.105038E+00 | lm loss PPL: 6.064506E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14300 | lm loss value: 4.104886E+00 | lm loss PPL: 6.063584E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14310/ 52452 | consumed samples: 7326720 | consumed tokens: 15005122560 | elapsed time per iteration (s): 9.24 | learning rate: 1.705E-04 | global batch size: 512 | lm loss: 1.473023E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.425 | TFLOPs: 8.32 | 31: iteration 14320/ 52452 | consumed samples: 7331840 | consumed tokens: 15015608320 | elapsed time per iteration (s): 1.84 | learning rate: 1.704E-04 | global batch size: 512 | lm loss: 1.471271E+00 | grad norm: 0.201 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.506 | TFLOPs: 41.80 | 31: iteration 14330/ 52452 | consumed samples: 7336960 | consumed tokens: 15026094080 | elapsed time per iteration (s): 1.74 | learning rate: 1.704E-04 | global batch size: 512 | lm loss: 1.471040E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.830 | TFLOPs: 44.10 | 31: iteration 14340/ 52452 | consumed samples: 7342080 | consumed tokens: 15036579840 | elapsed time per iteration (s): 1.75 | learning rate: 1.704E-04 | global batch size: 512 | lm loss: 1.482529E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.384 | TFLOPs: 43.89 | 31: iteration 14350/ 52452 | consumed samples: 7347200 | consumed tokens: 15047065600 | elapsed time per iteration (s): 1.79 | learning rate: 1.703E-04 | global batch size: 512 | lm loss: 1.477273E+00 | grad norm: 0.205 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.687 | TFLOPs: 42.88 | 31: iteration 14360/ 52452 | consumed samples: 7352320 | consumed tokens: 15057551360 | elapsed time per iteration (s): 1.78 | learning rate: 1.703E-04 | global batch size: 512 | lm loss: 1.488057E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.769 | TFLOPs: 43.19 | 31: iteration 14370/ 52452 | consumed samples: 7357440 | consumed tokens: 15068037120 | elapsed time per iteration (s): 1.84 | learning rate: 1.702E-04 | global batch size: 512 | lm loss: 1.478227E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.773 | TFLOPs: 41.69 | 31: iteration 14380/ 52452 | consumed samples: 7362560 | consumed tokens: 15078522880 | elapsed time per iteration (s): 1.74 | learning rate: 1.702E-04 | global batch size: 512 | lm loss: 1.486393E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.930 | TFLOPs: 44.12 | 31: iteration 14390/ 52452 | consumed samples: 7367680 | consumed tokens: 15089008640 | elapsed time per iteration (s): 1.81 | learning rate: 1.701E-04 | global batch size: 512 | lm loss: 1.470875E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.307 | TFLOPs: 42.52 | 31: iteration 14400/ 52452 | consumed samples: 7372800 | consumed tokens: 15099494400 | elapsed time per iteration (s): 1.80 | learning rate: 1.701E-04 | global batch size: 512 | lm loss: 1.484824E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.021 | TFLOPs: 42.78 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14400 | lm loss value: 4.079036E+00 | lm loss PPL: 5.908850E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14400 | lm loss value: 4.055215E+00 | lm loss PPL: 5.769756E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14410/ 52452 | consumed samples: 7377920 | consumed tokens: 15109980160 | elapsed time per iteration (s): 9.18 | learning rate: 1.701E-04 | global batch size: 512 | lm loss: 1.478399E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.804 | TFLOPs: 8.38 | 31: iteration 14420/ 52452 | consumed samples: 7383040 | consumed tokens: 15120465920 | elapsed time per iteration (s): 1.86 | learning rate: 1.700E-04 | global batch size: 512 | lm loss: 1.461195E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.492 | TFLOPs: 41.35 | 31: iteration 14430/ 52452 | consumed samples: 7388160 | consumed tokens: 15130951680 | elapsed time per iteration (s): 1.75 | learning rate: 1.700E-04 | global batch size: 512 | lm loss: 1.471244E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.790 | TFLOPs: 43.95 | 31: iteration 14440/ 52452 | consumed samples: 7393280 | consumed tokens: 15141437440 | elapsed time per iteration (s): 1.79 | learning rate: 1.699E-04 | global batch size: 512 | lm loss: 1.475015E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.703 | TFLOPs: 43.03 | 31: iteration 14450/ 52452 | consumed samples: 7398400 | consumed tokens: 15151923200 | elapsed time per iteration (s): 1.74 | learning rate: 1.699E-04 | global batch size: 512 | lm loss: 1.464347E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.686 | TFLOPs: 44.23 | 31: iteration 14460/ 52452 | consumed samples: 7403520 | consumed tokens: 15162408960 | elapsed time per iteration (s): 1.87 | learning rate: 1.699E-04 | global batch size: 512 | lm loss: 1.470416E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.224 | TFLOPs: 41.16 | 31: iteration 14470/ 52452 | consumed samples: 7408640 | consumed tokens: 15172894720 | elapsed time per iteration (s): 1.75 | learning rate: 1.698E-04 | global batch size: 512 | lm loss: 1.463593E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.960 | TFLOPs: 43.82 | 31: iteration 14480/ 52452 | consumed samples: 7413760 | consumed tokens: 15183380480 | elapsed time per iteration (s): 1.78 | learning rate: 1.698E-04 | global batch size: 512 | lm loss: 1.489414E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.950 | TFLOPs: 43.22 | 31: iteration 14490/ 52452 | consumed samples: 7418880 | consumed tokens: 15193866240 | elapsed time per iteration (s): 1.78 | learning rate: 1.697E-04 | global batch size: 512 | lm loss: 1.463738E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.851 | TFLOPs: 43.20 | 31: iteration 14500/ 52452 | consumed samples: 7424000 | consumed tokens: 15204352000 | elapsed time per iteration (s): 1.77 | learning rate: 1.697E-04 | global batch size: 512 | lm loss: 1.466177E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.736 | TFLOPs: 43.34 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14500 | lm loss value: 4.019043E+00 | lm loss PPL: 5.564782E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14500 | lm loss value: 4.025020E+00 | lm loss PPL: 5.598144E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14510/ 52452 | consumed samples: 7429120 | consumed tokens: 15214837760 | elapsed time per iteration (s): 9.20 | learning rate: 1.697E-04 | global batch size: 512 | lm loss: 1.469623E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.633 | TFLOPs: 8.35 | 31: iteration 14520/ 52452 | consumed samples: 7434240 | consumed tokens: 15225323520 | elapsed time per iteration (s): 1.75 | learning rate: 1.696E-04 | global batch size: 512 | lm loss: 1.470692E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.149 | TFLOPs: 43.85 | 31: iteration 14530/ 52452 | consumed samples: 7439360 | consumed tokens: 15235809280 | elapsed time per iteration (s): 1.77 | learning rate: 1.696E-04 | global batch size: 512 | lm loss: 1.477491E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.220 | TFLOPs: 43.41 | 31: iteration 14540/ 52452 | consumed samples: 7444480 | consumed tokens: 15246295040 | elapsed time per iteration (s): 1.74 | learning rate: 1.695E-04 | global batch size: 512 | lm loss: 1.471591E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.092 | TFLOPs: 44.29 | 31: iteration 14550/ 52452 | consumed samples: 7449600 | consumed tokens: 15256780800 | elapsed time per iteration (s): 1.77 | learning rate: 1.695E-04 | global batch size: 512 | lm loss: 1.491580E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.700 | TFLOPs: 43.33 | 31: iteration 14560/ 52452 | consumed samples: 7454720 | consumed tokens: 15267266560 | elapsed time per iteration (s): 1.78 | learning rate: 1.695E-04 | global batch size: 512 | lm loss: 1.486120E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.433 | TFLOPs: 43.14 | 31: iteration 14570/ 52452 | consumed samples: 7459840 | consumed tokens: 15277752320 | elapsed time per iteration (s): 1.76 | learning rate: 1.694E-04 | global batch size: 512 | lm loss: 1.468335E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.500 | TFLOPs: 43.75 | 31: iteration 14580/ 52452 | consumed samples: 7464960 | consumed tokens: 15288238080 | elapsed time per iteration (s): 1.80 | learning rate: 1.694E-04 | global batch size: 512 | lm loss: 1.475598E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.831 | TFLOPs: 42.75 | 31: iteration 14590/ 52452 | consumed samples: 7470080 | consumed tokens: 15298723840 | elapsed time per iteration (s): 1.75 | learning rate: 1.693E-04 | global batch size: 512 | lm loss: 1.480962E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.468 | TFLOPs: 43.90 | 31: iteration 14600/ 52452 | consumed samples: 7475200 | consumed tokens: 15309209600 | elapsed time per iteration (s): 1.78 | learning rate: 1.693E-04 | global batch size: 512 | lm loss: 1.467100E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.099 | TFLOPs: 43.09 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14600 | lm loss value: 4.110553E+00 | lm loss PPL: 6.098042E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14600 | lm loss value: 4.073909E+00 | lm loss PPL: 5.878633E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14610/ 52452 | consumed samples: 7480320 | consumed tokens: 15319695360 | elapsed time per iteration (s): 9.21 | learning rate: 1.693E-04 | global batch size: 512 | lm loss: 1.479011E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.613 | TFLOPs: 8.35 | 31: iteration 14620/ 52452 | consumed samples: 7485440 | consumed tokens: 15330181120 | elapsed time per iteration (s): 1.74 | learning rate: 1.692E-04 | global batch size: 512 | lm loss: 1.459992E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.894 | TFLOPs: 44.26 | 31: iteration 14630/ 52452 | consumed samples: 7490560 | consumed tokens: 15340666880 | elapsed time per iteration (s): 1.80 | learning rate: 1.692E-04 | global batch size: 512 | lm loss: 1.479127E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.624 | TFLOPs: 42.72 | 31: iteration 14640/ 52452 | consumed samples: 7495680 | consumed tokens: 15351152640 | elapsed time per iteration (s): 1.77 | learning rate: 1.691E-04 | global batch size: 512 | lm loss: 1.469738E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.741 | TFLOPs: 43.34 | 31: iteration 14650/ 52452 | consumed samples: 7500800 | consumed tokens: 15361638400 | elapsed time per iteration (s): 1.76 | learning rate: 1.691E-04 | global batch size: 512 | lm loss: 1.457380E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.352 | TFLOPs: 43.73 | 31: iteration 14660/ 52452 | consumed samples: 7505920 | consumed tokens: 15372124160 | elapsed time per iteration (s): 1.84 | learning rate: 1.690E-04 | global batch size: 512 | lm loss: 1.472187E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.811 | TFLOPs: 41.70 | 31: iteration 14670/ 52452 | consumed samples: 7511040 | consumed tokens: 15382609920 | elapsed time per iteration (s): 1.72 | learning rate: 1.690E-04 | global batch size: 512 | lm loss: 1.477685E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.064 | TFLOPs: 44.59 | 31: iteration 14680/ 52452 | consumed samples: 7516160 | consumed tokens: 15393095680 | elapsed time per iteration (s): 1.77 | learning rate: 1.690E-04 | global batch size: 512 | lm loss: 1.477557E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.903 | TFLOPs: 43.51 | 31: iteration 14690/ 52452 | consumed samples: 7521280 | consumed tokens: 15403581440 | elapsed time per iteration (s): 1.74 | learning rate: 1.689E-04 | global batch size: 512 | lm loss: 1.466782E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.243 | TFLOPs: 44.16 | 31: iteration 14700/ 52452 | consumed samples: 7526400 | consumed tokens: 15414067200 | elapsed time per iteration (s): 1.75 | learning rate: 1.689E-04 | global batch size: 512 | lm loss: 1.470297E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.138 | TFLOPs: 44.00 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14700 | lm loss value: 4.173102E+00 | lm loss PPL: 6.491654E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14700 | lm loss value: 4.147566E+00 | lm loss PPL: 6.327981E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14710/ 52452 | consumed samples: 7531520 | consumed tokens: 15424552960 | elapsed time per iteration (s): 9.11 | learning rate: 1.688E-04 | global batch size: 512 | lm loss: 1.463702E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.181 | TFLOPs: 8.43 | 31: iteration 14720/ 52452 | consumed samples: 7536640 | consumed tokens: 15435038720 | elapsed time per iteration (s): 1.76 | learning rate: 1.688E-04 | global batch size: 512 | lm loss: 1.472215E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.498 | TFLOPs: 43.60 | 31: iteration 14730/ 52452 | consumed samples: 7541760 | consumed tokens: 15445524480 | elapsed time per iteration (s): 1.77 | learning rate: 1.688E-04 | global batch size: 512 | lm loss: 1.471080E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.718 | TFLOPs: 43.33 | 31: iteration 14740/ 52452 | consumed samples: 7546880 | consumed tokens: 15456010240 | elapsed time per iteration (s): 1.75 | learning rate: 1.687E-04 | global batch size: 512 | lm loss: 1.471737E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.871 | TFLOPs: 43.81 | 31: iteration 14750/ 52452 | consumed samples: 7552000 | consumed tokens: 15466496000 | elapsed time per iteration (s): 1.76 | learning rate: 1.687E-04 | global batch size: 512 | lm loss: 1.469363E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.195 | TFLOPs: 43.71 | 31: iteration 14760/ 52452 | consumed samples: 7557120 | consumed tokens: 15476981760 | elapsed time per iteration (s): 1.77 | learning rate: 1.686E-04 | global batch size: 512 | lm loss: 1.469234E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.747 | TFLOPs: 43.49 | 31: iteration 14770/ 52452 | consumed samples: 7562240 | consumed tokens: 15487467520 | elapsed time per iteration (s): 1.75 | learning rate: 1.686E-04 | global batch size: 512 | lm loss: 1.466274E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.051 | TFLOPs: 43.84 | 31: iteration 14780/ 52452 | consumed samples: 7567360 | consumed tokens: 15497953280 | elapsed time per iteration (s): 1.74 | learning rate: 1.686E-04 | global batch size: 512 | lm loss: 1.461143E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.991 | TFLOPs: 44.28 | 31: iteration 14790/ 52452 | consumed samples: 7572480 | consumed tokens: 15508439040 | elapsed time per iteration (s): 1.76 | learning rate: 1.685E-04 | global batch size: 512 | lm loss: 1.465189E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.349 | TFLOPs: 43.58 | 31: iteration 14800/ 52452 | consumed samples: 7577600 | consumed tokens: 15518924800 | elapsed time per iteration (s): 1.74 | learning rate: 1.685E-04 | global batch size: 512 | lm loss: 1.480775E+00 | grad norm: 0.191 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.071 | TFLOPs: 44.29 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14800 | lm loss value: 4.119904E+00 | lm loss PPL: 6.155331E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14800 | lm loss value: 4.099226E+00 | lm loss PPL: 6.029360E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14810/ 52452 | consumed samples: 7582720 | consumed tokens: 15529410560 | elapsed time per iteration (s): 9.19 | learning rate: 1.684E-04 | global batch size: 512 | lm loss: 1.465103E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.734 | TFLOPs: 8.37 | 31: iteration 14820/ 52452 | consumed samples: 7587840 | consumed tokens: 15539896320 | elapsed time per iteration (s): 1.79 | learning rate: 1.684E-04 | global batch size: 512 | lm loss: 1.474778E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.307 | TFLOPs: 42.82 | 31: iteration 14830/ 52452 | consumed samples: 7592960 | consumed tokens: 15550382080 | elapsed time per iteration (s): 1.78 | learning rate: 1.683E-04 | global batch size: 512 | lm loss: 1.465647E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.165 | TFLOPs: 43.25 | 31: iteration 14840/ 52452 | consumed samples: 7598080 | consumed tokens: 15560867840 | elapsed time per iteration (s): 1.77 | learning rate: 1.683E-04 | global batch size: 512 | lm loss: 1.471059E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.496 | TFLOPs: 43.45 | 31: iteration 14850/ 52452 | consumed samples: 7603200 | consumed tokens: 15571353600 | elapsed time per iteration (s): 1.76 | learning rate: 1.683E-04 | global batch size: 512 | lm loss: 1.473646E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.205 | TFLOPs: 43.71 | 31: iteration 14860/ 52452 | consumed samples: 7608320 | consumed tokens: 15581839360 | elapsed time per iteration (s): 1.75 | learning rate: 1.682E-04 | global batch size: 512 | lm loss: 1.487299E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.173 | TFLOPs: 44.00 | 31: iteration 14870/ 52452 | consumed samples: 7613440 | consumed tokens: 15592325120 | elapsed time per iteration (s): 1.76 | learning rate: 1.682E-04 | global batch size: 512 | lm loss: 1.471086E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.201 | TFLOPs: 43.56 | 31: iteration 14880/ 52452 | consumed samples: 7618560 | consumed tokens: 15602810880 | elapsed time per iteration (s): 1.80 | learning rate: 1.681E-04 | global batch size: 512 | lm loss: 1.459841E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.369 | TFLOPs: 42.68 | 31: iteration 14890/ 52452 | consumed samples: 7623680 | consumed tokens: 15613296640 | elapsed time per iteration (s): 1.79 | learning rate: 1.681E-04 | global batch size: 512 | lm loss: 1.457825E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.506 | TFLOPs: 43.00 | 31: iteration 14900/ 52452 | consumed samples: 7628800 | consumed tokens: 15623782400 | elapsed time per iteration (s): 1.74 | learning rate: 1.681E-04 | global batch size: 512 | lm loss: 1.470904E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.561 | TFLOPs: 44.21 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 14900 | lm loss value: 4.070697E+00 | lm loss PPL: 5.859778E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 14900 | lm loss value: 4.065826E+00 | lm loss PPL: 5.831308E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 14910/ 52452 | consumed samples: 7633920 | consumed tokens: 15634268160 | elapsed time per iteration (s): 9.23 | learning rate: 1.680E-04 | global batch size: 512 | lm loss: 1.469864E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.471 | TFLOPs: 8.33 | 31: iteration 14920/ 52452 | consumed samples: 7639040 | consumed tokens: 15644753920 | elapsed time per iteration (s): 1.76 | learning rate: 1.680E-04 | global batch size: 512 | lm loss: 1.459738E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.422 | TFLOPs: 43.74 | 31: iteration 14930/ 52452 | consumed samples: 7644160 | consumed tokens: 15655239680 | elapsed time per iteration (s): 1.79 | learning rate: 1.679E-04 | global batch size: 512 | lm loss: 1.465546E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.344 | TFLOPs: 42.83 | 31: iteration 14940/ 52452 | consumed samples: 7649280 | consumed tokens: 15665725440 | elapsed time per iteration (s): 1.72 | learning rate: 1.679E-04 | global batch size: 512 | lm loss: 1.468454E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.400 | TFLOPs: 44.64 | 31: iteration 14950/ 52452 | consumed samples: 7654400 | consumed tokens: 15676211200 | elapsed time per iteration (s): 1.73 | learning rate: 1.678E-04 | global batch size: 512 | lm loss: 1.468944E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.321 | TFLOPs: 44.48 | 31: iteration 14960/ 52452 | consumed samples: 7659520 | consumed tokens: 15686696960 | elapsed time per iteration (s): 1.74 | learning rate: 1.678E-04 | global batch size: 512 | lm loss: 1.464582E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.615 | TFLOPs: 44.07 | 31: iteration 14970/ 52452 | consumed samples: 7664640 | consumed tokens: 15697182720 | elapsed time per iteration (s): 1.75 | learning rate: 1.678E-04 | global batch size: 512 | lm loss: 1.471518E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.328 | TFLOPs: 43.88 | 31: iteration 14980/ 52452 | consumed samples: 7669760 | consumed tokens: 15707668480 | elapsed time per iteration (s): 1.75 | learning rate: 1.677E-04 | global batch size: 512 | lm loss: 1.460609E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.686 | TFLOPs: 43.93 | 31: iteration 14990/ 52452 | consumed samples: 7674880 | consumed tokens: 15718154240 | elapsed time per iteration (s): 1.74 | learning rate: 1.677E-04 | global batch size: 512 | lm loss: 1.466166E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.941 | TFLOPs: 44.12 | 31: iteration 15000/ 52452 | consumed samples: 7680000 | consumed tokens: 15728640000 | elapsed time per iteration (s): 1.75 | learning rate: 1.676E-04 | global batch size: 512 | lm loss: 1.457988E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.284 | TFLOPs: 44.02 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15000 | lm loss value: 4.233010E+00 | lm loss PPL: 6.892437E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15000 | lm loss value: 4.182567E+00 | lm loss PPL: 6.553387E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15010/ 52452 | consumed samples: 7685120 | consumed tokens: 15739125760 | elapsed time per iteration (s): 9.24 | learning rate: 1.676E-04 | global batch size: 512 | lm loss: 1.469504E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.390 | TFLOPs: 8.31 | 31: iteration 15020/ 52452 | consumed samples: 7690240 | consumed tokens: 15749611520 | elapsed time per iteration (s): 1.79 | learning rate: 1.676E-04 | global batch size: 512 | lm loss: 1.469702E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.416 | TFLOPs: 42.84 | 31: iteration 15030/ 52452 | consumed samples: 7695360 | consumed tokens: 15760097280 | elapsed time per iteration (s): 1.77 | learning rate: 1.675E-04 | global batch size: 512 | lm loss: 1.465596E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.786 | TFLOPs: 43.35 | 31: iteration 15040/ 52452 | consumed samples: 7700480 | consumed tokens: 15770583040 | elapsed time per iteration (s): 1.74 | learning rate: 1.675E-04 | global batch size: 512 | lm loss: 1.470287E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.715 | TFLOPs: 44.08 | 31: iteration 15050/ 52452 | consumed samples: 7705600 | consumed tokens: 15781068800 | elapsed time per iteration (s): 1.74 | learning rate: 1.674E-04 | global batch size: 512 | lm loss: 1.461515E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.749 | TFLOPs: 44.09 | 31: iteration 15060/ 52452 | consumed samples: 7710720 | consumed tokens: 15791554560 | elapsed time per iteration (s): 1.76 | learning rate: 1.674E-04 | global batch size: 512 | lm loss: 1.462459E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.596 | TFLOPs: 43.62 | 31: iteration 15070/ 52452 | consumed samples: 7715840 | consumed tokens: 15802040320 | elapsed time per iteration (s): 1.77 | learning rate: 1.673E-04 | global batch size: 512 | lm loss: 1.470205E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.546 | TFLOPs: 43.31 | 31: iteration 15080/ 52452 | consumed samples: 7720960 | consumed tokens: 15812526080 | elapsed time per iteration (s): 1.76 | learning rate: 1.673E-04 | global batch size: 512 | lm loss: 1.472779E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.178 | TFLOPs: 43.70 | 31: iteration 15090/ 52452 | consumed samples: 7726080 | consumed tokens: 15823011840 | elapsed time per iteration (s): 1.76 | learning rate: 1.673E-04 | global batch size: 512 | lm loss: 1.462220E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.696 | TFLOPs: 43.63 | 31: iteration 15100/ 52452 | consumed samples: 7731200 | consumed tokens: 15833497600 | elapsed time per iteration (s): 1.75 | learning rate: 1.672E-04 | global batch size: 512 | lm loss: 1.460681E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.600 | TFLOPs: 43.92 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15100 | lm loss value: 4.244308E+00 | lm loss PPL: 6.970754E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15100 | lm loss value: 4.210097E+00 | lm loss PPL: 6.736306E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15110/ 52452 | consumed samples: 7736320 | consumed tokens: 15843983360 | elapsed time per iteration (s): 9.27 | learning rate: 1.672E-04 | global batch size: 512 | lm loss: 1.469634E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.235 | TFLOPs: 8.29 | 31: iteration 15120/ 52452 | consumed samples: 7741440 | consumed tokens: 15854469120 | elapsed time per iteration (s): 1.77 | learning rate: 1.671E-04 | global batch size: 512 | lm loss: 1.467178E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.818 | TFLOPs: 43.50 | 31: iteration 15130/ 52452 | consumed samples: 7746560 | consumed tokens: 15864954880 | elapsed time per iteration (s): 1.77 | learning rate: 1.671E-04 | global batch size: 512 | lm loss: 1.453324E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.483 | TFLOPs: 43.30 | 31: iteration 15140/ 52452 | consumed samples: 7751680 | consumed tokens: 15875440640 | elapsed time per iteration (s): 1.75 | learning rate: 1.670E-04 | global batch size: 512 | lm loss: 1.463673E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.359 | TFLOPs: 44.03 | 31: iteration 15150/ 52452 | consumed samples: 7756800 | consumed tokens: 15885926400 | elapsed time per iteration (s): 1.79 | learning rate: 1.670E-04 | global batch size: 512 | lm loss: 1.463735E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.096 | TFLOPs: 42.94 | 31: iteration 15160/ 52452 | consumed samples: 7761920 | consumed tokens: 15896412160 | elapsed time per iteration (s): 1.78 | learning rate: 1.670E-04 | global batch size: 512 | lm loss: 1.473831E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.116 | TFLOPs: 43.24 | 31: iteration 15170/ 52452 | consumed samples: 7767040 | consumed tokens: 15906897920 | elapsed time per iteration (s): 1.77 | learning rate: 1.669E-04 | global batch size: 512 | lm loss: 1.453630E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.602 | TFLOPs: 43.32 | 31: iteration 15180/ 52452 | consumed samples: 7772160 | consumed tokens: 15917383680 | elapsed time per iteration (s): 1.75 | learning rate: 1.669E-04 | global batch size: 512 | lm loss: 1.445229E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.198 | TFLOPs: 44.01 | 31: iteration 15190/ 52452 | consumed samples: 7777280 | consumed tokens: 15927869440 | elapsed time per iteration (s): 1.78 | learning rate: 1.668E-04 | global batch size: 512 | lm loss: 1.459003E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.332 | TFLOPs: 43.28 | 31: iteration 15200/ 52452 | consumed samples: 7782400 | consumed tokens: 15938355200 | elapsed time per iteration (s): 1.86 | learning rate: 1.668E-04 | global batch size: 512 | lm loss: 1.464452E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.402 | TFLOPs: 41.34 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15200 | lm loss value: 4.119879E+00 | lm loss PPL: 6.155181E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15200 | lm loss value: 4.091331E+00 | lm loss PPL: 5.981946E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15210/ 52452 | consumed samples: 7787520 | consumed tokens: 15948840960 | elapsed time per iteration (s): 9.16 | learning rate: 1.668E-04 | global batch size: 512 | lm loss: 1.460998E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.920 | TFLOPs: 8.39 | 31: iteration 15220/ 52452 | consumed samples: 7792640 | consumed tokens: 15959326720 | elapsed time per iteration (s): 1.75 | learning rate: 1.667E-04 | global batch size: 512 | lm loss: 1.465632E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.632 | TFLOPs: 43.92 | 31: iteration 15230/ 52452 | consumed samples: 7797760 | consumed tokens: 15969812480 | elapsed time per iteration (s): 1.75 | learning rate: 1.667E-04 | global batch size: 512 | lm loss: 1.448283E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.157 | TFLOPs: 44.00 | 31: iteration 15240/ 52452 | consumed samples: 7802880 | consumed tokens: 15980298240 | elapsed time per iteration (s): 1.78 | learning rate: 1.666E-04 | global batch size: 512 | lm loss: 1.465103E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.921 | TFLOPs: 43.22 | 31: iteration 15250/ 52452 | consumed samples: 7808000 | consumed tokens: 15990784000 | elapsed time per iteration (s): 1.79 | learning rate: 1.666E-04 | global batch size: 512 | lm loss: 1.472130E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.308 | TFLOPs: 42.82 | 31: iteration 15260/ 52452 | consumed samples: 7813120 | consumed tokens: 16001269760 | elapsed time per iteration (s): 1.78 | learning rate: 1.665E-04 | global batch size: 512 | lm loss: 1.459353E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.709 | TFLOPs: 43.18 | 31: iteration 15270/ 52452 | consumed samples: 7818240 | consumed tokens: 16011755520 | elapsed time per iteration (s): 1.77 | learning rate: 1.665E-04 | global batch size: 512 | lm loss: 1.461416E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.663 | TFLOPs: 43.48 | 31: iteration 15280/ 52452 | consumed samples: 7823360 | consumed tokens: 16022241280 | elapsed time per iteration (s): 1.77 | learning rate: 1.665E-04 | global batch size: 512 | lm loss: 1.465184E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.873 | TFLOPs: 43.36 | 31: iteration 15290/ 52452 | consumed samples: 7828480 | consumed tokens: 16032727040 | elapsed time per iteration (s): 1.84 | learning rate: 1.664E-04 | global batch size: 512 | lm loss: 1.449066E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.907 | TFLOPs: 41.71 | 31: iteration 15300/ 52452 | consumed samples: 7833600 | consumed tokens: 16043212800 | elapsed time per iteration (s): 1.85 | learning rate: 1.664E-04 | global batch size: 512 | lm loss: 1.465953E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.998 | TFLOPs: 41.58 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15300 | lm loss value: 4.074057E+00 | lm loss PPL: 5.879499E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15300 | lm loss value: 4.055645E+00 | lm loss PPL: 5.772238E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15310/ 52452 | consumed samples: 7838720 | consumed tokens: 16053698560 | elapsed time per iteration (s): 9.25 | learning rate: 1.663E-04 | global batch size: 512 | lm loss: 1.444641E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.343 | TFLOPs: 8.31 | 31: iteration 15320/ 52452 | consumed samples: 7843840 | consumed tokens: 16064184320 | elapsed time per iteration (s): 1.79 | learning rate: 1.663E-04 | global batch size: 512 | lm loss: 1.450261E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.779 | TFLOPs: 42.89 | 31: iteration 15330/ 52452 | consumed samples: 7848960 | consumed tokens: 16074670080 | elapsed time per iteration (s): 1.75 | learning rate: 1.662E-04 | global batch size: 512 | lm loss: 1.461140E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.796 | TFLOPs: 43.80 | 31: iteration 15340/ 52452 | consumed samples: 7854080 | consumed tokens: 16085155840 | elapsed time per iteration (s): 1.74 | learning rate: 1.662E-04 | global batch size: 512 | lm loss: 1.440044E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.385 | TFLOPs: 44.19 | 31: iteration 15350/ 52452 | consumed samples: 7859200 | consumed tokens: 16095641600 | elapsed time per iteration (s): 1.76 | learning rate: 1.662E-04 | global batch size: 512 | lm loss: 1.468592E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.880 | TFLOPs: 43.66 | 31: iteration 15360/ 52452 | consumed samples: 7864320 | consumed tokens: 16106127360 | elapsed time per iteration (s): 1.75 | learning rate: 1.661E-04 | global batch size: 512 | lm loss: 1.470155E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.329 | TFLOPs: 43.88 | 31: iteration 15370/ 52452 | consumed samples: 7869440 | consumed tokens: 16116613120 | elapsed time per iteration (s): 1.75 | learning rate: 1.661E-04 | global batch size: 512 | lm loss: 1.460856E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.208 | TFLOPs: 44.01 | 31: iteration 15380/ 52452 | consumed samples: 7874560 | consumed tokens: 16127098880 | elapsed time per iteration (s): 1.76 | learning rate: 1.660E-04 | global batch size: 512 | lm loss: 1.465906E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.698 | TFLOPs: 43.78 | 31: iteration 15390/ 52452 | consumed samples: 7879680 | consumed tokens: 16137584640 | elapsed time per iteration (s): 1.75 | learning rate: 1.660E-04 | global batch size: 512 | lm loss: 1.473247E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.132 | TFLOPs: 43.85 | 31: iteration 15400/ 52452 | consumed samples: 7884800 | consumed tokens: 16148070400 | elapsed time per iteration (s): 1.76 | learning rate: 1.659E-04 | global batch size: 512 | lm loss: 1.464693E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.988 | TFLOPs: 43.68 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15400 | lm loss value: 4.134447E+00 | lm loss PPL: 6.245502E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15400 | lm loss value: 4.100525E+00 | lm loss PPL: 6.037197E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15410/ 52452 | consumed samples: 7889920 | consumed tokens: 16158556160 | elapsed time per iteration (s): 9.13 | learning rate: 1.659E-04 | global batch size: 512 | lm loss: 1.446470E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.085 | TFLOPs: 8.42 | 31: iteration 15420/ 52452 | consumed samples: 7895040 | consumed tokens: 16169041920 | elapsed time per iteration (s): 1.74 | learning rate: 1.659E-04 | global batch size: 512 | lm loss: 1.459667E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.720 | TFLOPs: 44.24 | 31: iteration 15430/ 52452 | consumed samples: 7900160 | consumed tokens: 16179527680 | elapsed time per iteration (s): 1.75 | learning rate: 1.658E-04 | global batch size: 512 | lm loss: 1.450788E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.538 | TFLOPs: 43.91 | 31: iteration 15440/ 52452 | consumed samples: 7905280 | consumed tokens: 16190013440 | elapsed time per iteration (s): 1.74 | learning rate: 1.658E-04 | global batch size: 512 | lm loss: 1.454302E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.936 | TFLOPs: 44.27 | 31: iteration 15450/ 52452 | consumed samples: 7910400 | consumed tokens: 16200499200 | elapsed time per iteration (s): 1.75 | learning rate: 1.657E-04 | global batch size: 512 | lm loss: 1.476477E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.080 | TFLOPs: 43.84 | 31: iteration 15460/ 52452 | consumed samples: 7915520 | consumed tokens: 16210984960 | elapsed time per iteration (s): 1.74 | learning rate: 1.657E-04 | global batch size: 512 | lm loss: 1.454346E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.519 | TFLOPs: 44.06 | 31: iteration 15470/ 52452 | consumed samples: 7920640 | consumed tokens: 16221470720 | elapsed time per iteration (s): 1.76 | learning rate: 1.656E-04 | global batch size: 512 | lm loss: 1.455177E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.904 | TFLOPs: 43.66 | 31: iteration 15480/ 52452 | consumed samples: 7925760 | consumed tokens: 16231956480 | elapsed time per iteration (s): 1.79 | learning rate: 1.656E-04 | global batch size: 512 | lm loss: 1.454885E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.813 | TFLOPs: 43.05 | 31: iteration 15490/ 52452 | consumed samples: 7930880 | consumed tokens: 16242442240 | elapsed time per iteration (s): 1.73 | learning rate: 1.656E-04 | global batch size: 512 | lm loss: 1.455378E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.971 | TFLOPs: 44.42 | 31: iteration 15500/ 52452 | consumed samples: 7936000 | consumed tokens: 16252928000 | elapsed time per iteration (s): 1.76 | learning rate: 1.655E-04 | global batch size: 512 | lm loss: 1.463378E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.323 | TFLOPs: 43.58 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15500 | lm loss value: 4.118518E+00 | lm loss PPL: 6.146810E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15500 | lm loss value: 4.071376E+00 | lm loss PPL: 5.863758E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15510/ 52452 | consumed samples: 7941120 | consumed tokens: 16263413760 | elapsed time per iteration (s): 9.17 | learning rate: 1.655E-04 | global batch size: 512 | lm loss: 1.455553E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.851 | TFLOPs: 8.38 | 31: iteration 15520/ 52452 | consumed samples: 7946240 | consumed tokens: 16273899520 | elapsed time per iteration (s): 1.72 | learning rate: 1.654E-04 | global batch size: 512 | lm loss: 1.451461E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.068 | TFLOPs: 44.59 | 31: iteration 15530/ 52452 | consumed samples: 7951360 | consumed tokens: 16284385280 | elapsed time per iteration (s): 1.73 | learning rate: 1.654E-04 | global batch size: 512 | lm loss: 1.441006E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.849 | TFLOPs: 44.41 | 31: iteration 15540/ 52452 | consumed samples: 7956480 | consumed tokens: 16294871040 | elapsed time per iteration (s): 1.74 | learning rate: 1.653E-04 | global batch size: 512 | lm loss: 1.454784E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.936 | TFLOPs: 44.12 | 31: iteration 15550/ 52452 | consumed samples: 7961600 | consumed tokens: 16305356800 | elapsed time per iteration (s): 1.80 | learning rate: 1.653E-04 | global batch size: 512 | lm loss: 1.459499E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.894 | TFLOPs: 42.76 | 31: iteration 15560/ 52452 | consumed samples: 7966720 | consumed tokens: 16315842560 | elapsed time per iteration (s): 1.75 | learning rate: 1.653E-04 | global batch size: 512 | lm loss: 1.460399E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.122 | TFLOPs: 43.85 | 31: iteration 15570/ 52452 | consumed samples: 7971840 | consumed tokens: 16326328320 | elapsed time per iteration (s): 1.72 | learning rate: 1.652E-04 | global batch size: 512 | lm loss: 1.475473E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.575 | TFLOPs: 44.66 | 31: iteration 15580/ 52452 | consumed samples: 7976960 | consumed tokens: 16336814080 | elapsed time per iteration (s): 1.80 | learning rate: 1.652E-04 | global batch size: 512 | lm loss: 1.453109E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.216 | TFLOPs: 42.66 | 31: iteration 15590/ 52452 | consumed samples: 7982080 | consumed tokens: 16347299840 | elapsed time per iteration (s): 1.78 | learning rate: 1.651E-04 | global batch size: 512 | lm loss: 1.463650E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.442 | TFLOPs: 43.29 | 31: iteration 15600/ 52452 | consumed samples: 7987200 | consumed tokens: 16357785600 | elapsed time per iteration (s): 1.73 | learning rate: 1.651E-04 | global batch size: 512 | lm loss: 1.453790E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.527 | TFLOPs: 44.51 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15600 | lm loss value: 4.317615E+00 | lm loss PPL: 7.500952E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15600 | lm loss value: 4.273258E+00 | lm loss PPL: 7.175501E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15610/ 52452 | consumed samples: 7992320 | consumed tokens: 16368271360 | elapsed time per iteration (s): 9.16 | learning rate: 1.650E-04 | global batch size: 512 | lm loss: 1.460106E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.913 | TFLOPs: 8.39 | 31: iteration 15620/ 52452 | consumed samples: 7997440 | consumed tokens: 16378757120 | elapsed time per iteration (s): 1.81 | learning rate: 1.650E-04 | global batch size: 512 | lm loss: 1.460808E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.786 | TFLOPs: 42.44 | 31: iteration 15630/ 52452 | consumed samples: 8002560 | consumed tokens: 16389242880 | elapsed time per iteration (s): 1.78 | learning rate: 1.650E-04 | global batch size: 512 | lm loss: 1.455444E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.086 | TFLOPs: 43.09 | 31: iteration 15640/ 52452 | consumed samples: 8007680 | consumed tokens: 16399728640 | elapsed time per iteration (s): 1.76 | learning rate: 1.649E-04 | global batch size: 512 | lm loss: 1.458048E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.582 | TFLOPs: 43.61 | 31: iteration 15650/ 52452 | consumed samples: 8012800 | consumed tokens: 16410214400 | elapsed time per iteration (s): 1.73 | learning rate: 1.649E-04 | global batch size: 512 | lm loss: 1.451667E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.114 | TFLOPs: 44.29 | 31: iteration 15660/ 52452 | consumed samples: 8017920 | consumed tokens: 16420700160 | elapsed time per iteration (s): 1.77 | learning rate: 1.648E-04 | global batch size: 512 | lm loss: 1.455862E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.613 | TFLOPs: 43.47 | 31: iteration 15670/ 52452 | consumed samples: 8023040 | consumed tokens: 16431185920 | elapsed time per iteration (s): 1.73 | learning rate: 1.648E-04 | global batch size: 512 | lm loss: 1.451292E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.320 | TFLOPs: 44.33 | 31: iteration 15680/ 52452 | consumed samples: 8028160 | consumed tokens: 16441671680 | elapsed time per iteration (s): 1.73 | learning rate: 1.647E-04 | global batch size: 512 | lm loss: 1.450881E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.118 | TFLOPs: 44.45 | 31: iteration 15690/ 52452 | consumed samples: 8033280 | consumed tokens: 16452157440 | elapsed time per iteration (s): 1.78 | learning rate: 1.647E-04 | global batch size: 512 | lm loss: 1.448362E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.373 | TFLOPs: 43.13 | 31: iteration 15700/ 52452 | consumed samples: 8038400 | consumed tokens: 16462643200 | elapsed time per iteration (s): 1.84 | learning rate: 1.647E-04 | global batch size: 512 | lm loss: 1.456008E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.217 | TFLOPs: 41.76 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15700 | lm loss value: 4.142117E+00 | lm loss PPL: 6.293589E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15700 | lm loss value: 4.104825E+00 | lm loss PPL: 6.063211E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15710/ 52452 | consumed samples: 8043520 | consumed tokens: 16473128960 | elapsed time per iteration (s): 9.15 | learning rate: 1.646E-04 | global batch size: 512 | lm loss: 1.453339E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.982 | TFLOPs: 8.40 | 31: iteration 15720/ 52452 | consumed samples: 8048640 | consumed tokens: 16483614720 | elapsed time per iteration (s): 1.74 | learning rate: 1.646E-04 | global batch size: 512 | lm loss: 1.451202E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.465 | TFLOPs: 44.05 | 31: iteration 15730/ 52452 | consumed samples: 8053760 | consumed tokens: 16494100480 | elapsed time per iteration (s): 1.79 | learning rate: 1.645E-04 | global batch size: 512 | lm loss: 1.446917E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.638 | TFLOPs: 43.02 | 31: iteration 15740/ 52452 | consumed samples: 8058880 | consumed tokens: 16504586240 | elapsed time per iteration (s): 1.77 | learning rate: 1.645E-04 | global batch size: 512 | lm loss: 1.440770E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.781 | TFLOPs: 43.49 | 31: iteration 15750/ 52452 | consumed samples: 8064000 | consumed tokens: 16515072000 | elapsed time per iteration (s): 1.74 | learning rate: 1.644E-04 | global batch size: 512 | lm loss: 1.462309E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.878 | TFLOPs: 44.26 | 31: iteration 15760/ 52452 | consumed samples: 8069120 | consumed tokens: 16525557760 | elapsed time per iteration (s): 1.74 | learning rate: 1.644E-04 | global batch size: 512 | lm loss: 1.451423E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.886 | TFLOPs: 44.11 | 31: iteration 15770/ 52452 | consumed samples: 8074240 | consumed tokens: 16536043520 | elapsed time per iteration (s): 1.76 | learning rate: 1.644E-04 | global batch size: 512 | lm loss: 1.465891E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.454 | TFLOPs: 43.75 | 31: iteration 15780/ 52452 | consumed samples: 8079360 | consumed tokens: 16546529280 | elapsed time per iteration (s): 1.73 | learning rate: 1.643E-04 | global batch size: 512 | lm loss: 1.439324E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.514 | TFLOPs: 44.51 | 31: iteration 15790/ 52452 | consumed samples: 8084480 | consumed tokens: 16557015040 | elapsed time per iteration (s): 1.73 | learning rate: 1.643E-04 | global batch size: 512 | lm loss: 1.446002E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.503 | TFLOPs: 44.35 | 31: iteration 15800/ 52452 | consumed samples: 8089600 | consumed tokens: 16567500800 | elapsed time per iteration (s): 1.76 | learning rate: 1.642E-04 | global batch size: 512 | lm loss: 1.450181E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.583 | TFLOPs: 43.61 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15800 | lm loss value: 3.996189E+00 | lm loss PPL: 5.439047E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15800 | lm loss value: 4.001107E+00 | lm loss PPL: 5.465864E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15810/ 52452 | consumed samples: 8094720 | consumed tokens: 16577986560 | elapsed time per iteration (s): 9.16 | learning rate: 1.642E-04 | global batch size: 512 | lm loss: 1.458546E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.902 | TFLOPs: 8.39 | 31: iteration 15820/ 52452 | consumed samples: 8099840 | consumed tokens: 16588472320 | elapsed time per iteration (s): 1.75 | learning rate: 1.641E-04 | global batch size: 512 | lm loss: 1.451569E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.133 | TFLOPs: 44.00 | 31: iteration 15830/ 52452 | consumed samples: 8104960 | consumed tokens: 16598958080 | elapsed time per iteration (s): 1.74 | learning rate: 1.641E-04 | global batch size: 512 | lm loss: 1.449044E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.450 | TFLOPs: 44.05 | 31: iteration 15840/ 52452 | consumed samples: 8110080 | consumed tokens: 16609443840 | elapsed time per iteration (s): 1.80 | learning rate: 1.641E-04 | global batch size: 512 | lm loss: 1.454284E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.167 | TFLOPs: 42.80 | 31: iteration 15850/ 52452 | consumed samples: 8115200 | consumed tokens: 16619929600 | elapsed time per iteration (s): 1.77 | learning rate: 1.640E-04 | global batch size: 512 | lm loss: 1.457326E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.868 | TFLOPs: 43.36 | 31: iteration 15860/ 52452 | consumed samples: 8120320 | consumed tokens: 16630415360 | elapsed time per iteration (s): 1.75 | learning rate: 1.640E-04 | global batch size: 512 | lm loss: 1.457286E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.104 | TFLOPs: 43.99 | 31: iteration 15870/ 52452 | consumed samples: 8125440 | consumed tokens: 16640901120 | elapsed time per iteration (s): 1.73 | learning rate: 1.639E-04 | global batch size: 512 | lm loss: 1.455557E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.180 | TFLOPs: 44.30 | 31: iteration 15880/ 52452 | consumed samples: 8130560 | consumed tokens: 16651386880 | elapsed time per iteration (s): 1.77 | learning rate: 1.639E-04 | global batch size: 512 | lm loss: 1.446708E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.275 | TFLOPs: 43.42 | 31: iteration 15890/ 52452 | consumed samples: 8135680 | consumed tokens: 16661872640 | elapsed time per iteration (s): 1.81 | learning rate: 1.638E-04 | global batch size: 512 | lm loss: 1.446684E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.112 | TFLOPs: 42.34 | 31: iteration 15900/ 52452 | consumed samples: 8140800 | consumed tokens: 16672358400 | elapsed time per iteration (s): 1.75 | learning rate: 1.638E-04 | global batch size: 512 | lm loss: 1.448850E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.296 | TFLOPs: 43.87 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 15900 | lm loss value: 4.108982E+00 | lm loss PPL: 6.088468E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 15900 | lm loss value: 4.100289E+00 | lm loss PPL: 6.035775E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 15910/ 52452 | consumed samples: 8145920 | consumed tokens: 16682844160 | elapsed time per iteration (s): 9.14 | learning rate: 1.637E-04 | global batch size: 512 | lm loss: 1.452627E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.040 | TFLOPs: 8.41 | 31: iteration 15920/ 52452 | consumed samples: 8151040 | consumed tokens: 16693329920 | elapsed time per iteration (s): 1.79 | learning rate: 1.637E-04 | global batch size: 512 | lm loss: 1.446820E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.795 | TFLOPs: 43.05 | 31: iteration 15930/ 52452 | consumed samples: 8156160 | consumed tokens: 16703815680 | elapsed time per iteration (s): 1.77 | learning rate: 1.637E-04 | global batch size: 512 | lm loss: 1.451171E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.423 | TFLOPs: 43.44 | 31: iteration 15940/ 52452 | consumed samples: 8161280 | consumed tokens: 16714301440 | elapsed time per iteration (s): 1.73 | learning rate: 1.636E-04 | global batch size: 512 | lm loss: 1.450404E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.185 | TFLOPs: 44.31 | 31: iteration 15950/ 52452 | consumed samples: 8166400 | consumed tokens: 16724787200 | elapsed time per iteration (s): 1.84 | learning rate: 1.636E-04 | global batch size: 512 | lm loss: 1.444394E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.507 | TFLOPs: 41.80 | 31: iteration 15960/ 52452 | consumed samples: 8171520 | consumed tokens: 16735272960 | elapsed time per iteration (s): 1.77 | learning rate: 1.635E-04 | global batch size: 512 | lm loss: 1.440683E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.708 | TFLOPs: 43.33 | 31: iteration 15970/ 52452 | consumed samples: 8176640 | consumed tokens: 16745758720 | elapsed time per iteration (s): 1.81 | learning rate: 1.635E-04 | global batch size: 512 | lm loss: 1.445889E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.933 | TFLOPs: 42.47 | 31: iteration 15980/ 52452 | consumed samples: 8181760 | consumed tokens: 16756244480 | elapsed time per iteration (s): 1.75 | learning rate: 1.634E-04 | global batch size: 512 | lm loss: 1.449318E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.203 | TFLOPs: 43.86 | 31: iteration 15990/ 52452 | consumed samples: 8186880 | consumed tokens: 16766730240 | elapsed time per iteration (s): 1.79 | learning rate: 1.634E-04 | global batch size: 512 | lm loss: 1.456769E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.277 | TFLOPs: 42.97 | 0: [2022-12-25 20:16:52,353] [INFO] [logging.py:68:log_dist] [Rank 0] step=16000, skipped=0, lr=[0.00016335244940516108, 0.00016335244940516108, 0.00016335244940516108], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 16000/ 52452 | consumed samples: 8192000 | consumed tokens: 16777216000 | elapsed time per iteration (s): 1.77 | learning rate: 1.634E-04 | global batch size: 512 | lm loss: 1.459315E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.586 | TFLOPs: 43.47 | 0: steps: 16000 loss: 1.4667 iter time (s): 1.773 samples/sec: 288.743 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16000 | lm loss value: 4.143299E+00 | lm loss PPL: 6.301033E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16000 | lm loss value: 4.104615E+00 | lm loss PPL: 6.061941E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16010/ 52452 | consumed samples: 8197120 | consumed tokens: 16787701760 | elapsed time per iteration (s): 9.17 | learning rate: 1.633E-04 | global batch size: 512 | lm loss: 1.446958E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.827 | TFLOPs: 8.38 | 31: iteration 16020/ 52452 | consumed samples: 8202240 | consumed tokens: 16798187520 | elapsed time per iteration (s): 1.76 | learning rate: 1.633E-04 | global batch size: 512 | lm loss: 1.466588E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.607 | TFLOPs: 43.62 | 31: iteration 16030/ 52452 | consumed samples: 8207360 | consumed tokens: 16808673280 | elapsed time per iteration (s): 1.72 | learning rate: 1.632E-04 | global batch size: 512 | lm loss: 1.448244E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.370 | TFLOPs: 44.63 | 31: iteration 16040/ 52452 | consumed samples: 8212480 | consumed tokens: 16819159040 | elapsed time per iteration (s): 1.72 | learning rate: 1.632E-04 | global batch size: 512 | lm loss: 1.457523E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.288 | TFLOPs: 44.62 | 31: iteration 16050/ 52452 | consumed samples: 8217600 | consumed tokens: 16829644800 | elapsed time per iteration (s): 1.72 | learning rate: 1.631E-04 | global batch size: 512 | lm loss: 1.443529E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.437 | TFLOPs: 44.64 | 31: iteration 16060/ 52452 | consumed samples: 8222720 | consumed tokens: 16840130560 | elapsed time per iteration (s): 1.76 | learning rate: 1.631E-04 | global batch size: 512 | lm loss: 1.430574E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.511 | TFLOPs: 43.60 | 31: iteration 16070/ 52452 | consumed samples: 8227840 | consumed tokens: 16850616320 | elapsed time per iteration (s): 1.80 | learning rate: 1.630E-04 | global batch size: 512 | lm loss: 1.448138E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.686 | TFLOPs: 42.58 | 31: iteration 16080/ 52452 | consumed samples: 8232960 | consumed tokens: 16861102080 | elapsed time per iteration (s): 1.77 | learning rate: 1.630E-04 | global batch size: 512 | lm loss: 1.454474E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.323 | TFLOPs: 43.43 | 31: iteration 16090/ 52452 | consumed samples: 8238080 | consumed tokens: 16871587840 | elapsed time per iteration (s): 1.77 | learning rate: 1.630E-04 | global batch size: 512 | lm loss: 1.447878E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.679 | TFLOPs: 43.48 | 31: iteration 16100/ 52452 | consumed samples: 8243200 | consumed tokens: 16882073600 | elapsed time per iteration (s): 1.78 | learning rate: 1.629E-04 | global batch size: 512 | lm loss: 1.452517E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.228 | TFLOPs: 43.26 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16100 | lm loss value: 4.149067E+00 | lm loss PPL: 6.337487E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16100 | lm loss value: 4.131379E+00 | lm loss PPL: 6.226370E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16110/ 52452 | consumed samples: 8248320 | consumed tokens: 16892559360 | elapsed time per iteration (s): 9.12 | learning rate: 1.629E-04 | global batch size: 512 | lm loss: 1.449178E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.148 | TFLOPs: 8.43 | 31: iteration 16120/ 52452 | consumed samples: 8253440 | consumed tokens: 16903045120 | elapsed time per iteration (s): 1.82 | learning rate: 1.628E-04 | global batch size: 512 | lm loss: 1.457030E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.595 | TFLOPs: 42.12 | 31: iteration 16130/ 52452 | consumed samples: 8258560 | consumed tokens: 16913530880 | elapsed time per iteration (s): 1.74 | learning rate: 1.628E-04 | global batch size: 512 | lm loss: 1.449926E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.191 | TFLOPs: 44.16 | 31: iteration 16140/ 52452 | consumed samples: 8263680 | consumed tokens: 16924016640 | elapsed time per iteration (s): 1.87 | learning rate: 1.627E-04 | global batch size: 512 | lm loss: 1.466140E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.158 | TFLOPs: 41.15 | 31: iteration 16150/ 52452 | consumed samples: 8268800 | consumed tokens: 16934502400 | elapsed time per iteration (s): 1.80 | learning rate: 1.627E-04 | global batch size: 512 | lm loss: 1.446908E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.754 | TFLOPs: 42.74 | 31: iteration 16160/ 52452 | consumed samples: 8273920 | consumed tokens: 16944988160 | elapsed time per iteration (s): 1.84 | learning rate: 1.626E-04 | global batch size: 512 | lm loss: 1.439904E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.360 | TFLOPs: 41.78 | 31: iteration 16170/ 52452 | consumed samples: 8279040 | consumed tokens: 16955473920 | elapsed time per iteration (s): 1.77 | learning rate: 1.626E-04 | global batch size: 512 | lm loss: 1.446747E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.879 | TFLOPs: 43.51 | 31: iteration 16180/ 52452 | consumed samples: 8284160 | consumed tokens: 16965959680 | elapsed time per iteration (s): 1.75 | learning rate: 1.626E-04 | global batch size: 512 | lm loss: 1.436689E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.146 | TFLOPs: 44.00 | 31: iteration 16190/ 52452 | consumed samples: 8289280 | consumed tokens: 16976445440 | elapsed time per iteration (s): 1.83 | learning rate: 1.625E-04 | global batch size: 512 | lm loss: 1.448332E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.603 | TFLOPs: 41.97 | 31: iteration 16200/ 52452 | consumed samples: 8294400 | consumed tokens: 16986931200 | elapsed time per iteration (s): 1.84 | learning rate: 1.625E-04 | global batch size: 512 | lm loss: 1.444310E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.736 | TFLOPs: 41.84 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16200 | lm loss value: 3.999865E+00 | lm loss PPL: 5.459077E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16200 | lm loss value: 4.018014E+00 | lm loss PPL: 5.559059E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16210/ 52452 | consumed samples: 8299520 | consumed tokens: 16997416960 | elapsed time per iteration (s): 9.34 | learning rate: 1.624E-04 | global batch size: 512 | lm loss: 1.448816E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.839 | TFLOPs: 8.23 | 31: iteration 16220/ 52452 | consumed samples: 8304640 | consumed tokens: 17007902720 | elapsed time per iteration (s): 1.78 | learning rate: 1.624E-04 | global batch size: 512 | lm loss: 1.450028E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.404 | TFLOPs: 43.14 | 31: iteration 16230/ 52452 | consumed samples: 8309760 | consumed tokens: 17018388480 | elapsed time per iteration (s): 1.79 | learning rate: 1.623E-04 | global batch size: 512 | lm loss: 1.451617E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.750 | TFLOPs: 42.89 | 31: iteration 16240/ 52452 | consumed samples: 8314880 | consumed tokens: 17028874240 | elapsed time per iteration (s): 1.73 | learning rate: 1.623E-04 | global batch size: 512 | lm loss: 1.450784E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.823 | TFLOPs: 44.40 | 31: iteration 16250/ 52452 | consumed samples: 8320000 | consumed tokens: 17039360000 | elapsed time per iteration (s): 1.78 | learning rate: 1.623E-04 | global batch size: 512 | lm loss: 1.439707E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.097 | TFLOPs: 43.24 | 31: iteration 16260/ 52452 | consumed samples: 8325120 | consumed tokens: 17049845760 | elapsed time per iteration (s): 1.78 | learning rate: 1.622E-04 | global batch size: 512 | lm loss: 1.451455E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.189 | TFLOPs: 43.26 | 31: iteration 16270/ 52452 | consumed samples: 8330240 | consumed tokens: 17060331520 | elapsed time per iteration (s): 1.76 | learning rate: 1.622E-04 | global batch size: 512 | lm loss: 1.447876E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.559 | TFLOPs: 43.61 | 31: iteration 16280/ 52452 | consumed samples: 8335360 | consumed tokens: 17070817280 | elapsed time per iteration (s): 1.77 | learning rate: 1.621E-04 | global batch size: 512 | lm loss: 1.444216E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.618 | TFLOPs: 43.47 | 31: iteration 16290/ 52452 | consumed samples: 8340480 | consumed tokens: 17081303040 | elapsed time per iteration (s): 1.74 | learning rate: 1.621E-04 | global batch size: 512 | lm loss: 1.446660E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.770 | TFLOPs: 44.09 | 31: iteration 16300/ 52452 | consumed samples: 8345600 | consumed tokens: 17091788800 | elapsed time per iteration (s): 1.79 | learning rate: 1.620E-04 | global batch size: 512 | lm loss: 1.450252E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.645 | TFLOPs: 43.02 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16300 | lm loss value: 4.047142E+00 | lm loss PPL: 5.723365E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16300 | lm loss value: 4.036676E+00 | lm loss PPL: 5.663779E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16310/ 52452 | consumed samples: 8350720 | consumed tokens: 17102274560 | elapsed time per iteration (s): 9.15 | learning rate: 1.620E-04 | global batch size: 512 | lm loss: 1.449954E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.959 | TFLOPs: 8.40 | 31: iteration 16320/ 52452 | consumed samples: 8355840 | consumed tokens: 17112760320 | elapsed time per iteration (s): 1.74 | learning rate: 1.619E-04 | global batch size: 512 | lm loss: 1.430411E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.761 | TFLOPs: 44.24 | 31: iteration 16330/ 52452 | consumed samples: 8360960 | consumed tokens: 17123246080 | elapsed time per iteration (s): 1.76 | learning rate: 1.619E-04 | global batch size: 512 | lm loss: 1.450155E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.216 | TFLOPs: 43.71 | 31: iteration 16340/ 52452 | consumed samples: 8366080 | consumed tokens: 17133731840 | elapsed time per iteration (s): 1.74 | learning rate: 1.619E-04 | global batch size: 512 | lm loss: 1.440386E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.537 | TFLOPs: 44.06 | 31: iteration 16350/ 52452 | consumed samples: 8371200 | consumed tokens: 17144217600 | elapsed time per iteration (s): 1.74 | learning rate: 1.618E-04 | global batch size: 512 | lm loss: 1.443385E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.605 | TFLOPs: 44.07 | 31: iteration 16360/ 52452 | consumed samples: 8376320 | consumed tokens: 17154703360 | elapsed time per iteration (s): 1.75 | learning rate: 1.618E-04 | global batch size: 512 | lm loss: 1.455590E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.903 | TFLOPs: 43.96 | 31: iteration 16370/ 52452 | consumed samples: 8381440 | consumed tokens: 17165189120 | elapsed time per iteration (s): 1.72 | learning rate: 1.617E-04 | global batch size: 512 | lm loss: 1.441203E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.470 | TFLOPs: 44.65 | 31: iteration 16380/ 52452 | consumed samples: 8386560 | consumed tokens: 17175674880 | elapsed time per iteration (s): 1.76 | learning rate: 1.617E-04 | global batch size: 512 | lm loss: 1.444335E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.887 | TFLOPs: 43.66 | 31: iteration 16390/ 52452 | consumed samples: 8391680 | consumed tokens: 17186160640 | elapsed time per iteration (s): 1.74 | learning rate: 1.616E-04 | global batch size: 512 | lm loss: 1.439448E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.129 | TFLOPs: 44.15 | 31: iteration 16400/ 52452 | consumed samples: 8396800 | consumed tokens: 17196646400 | elapsed time per iteration (s): 1.75 | learning rate: 1.616E-04 | global batch size: 512 | lm loss: 1.430300E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.347 | TFLOPs: 44.03 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16400 | lm loss value: 4.121159E+00 | lm loss PPL: 6.163061E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16400 | lm loss value: 4.097676E+00 | lm loss PPL: 6.020021E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16410/ 52452 | consumed samples: 8401920 | consumed tokens: 17207132160 | elapsed time per iteration (s): 9.18 | learning rate: 1.615E-04 | global batch size: 512 | lm loss: 1.447849E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.771 | TFLOPs: 8.37 | 31: iteration 16420/ 52452 | consumed samples: 8407040 | consumed tokens: 17217617920 | elapsed time per iteration (s): 1.77 | learning rate: 1.615E-04 | global batch size: 512 | lm loss: 1.437035E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.791 | TFLOPs: 43.35 | 31: iteration 16430/ 52452 | consumed samples: 8412160 | consumed tokens: 17228103680 | elapsed time per iteration (s): 1.74 | learning rate: 1.614E-04 | global batch size: 512 | lm loss: 1.440694E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.940 | TFLOPs: 44.12 | 31: iteration 16440/ 52452 | consumed samples: 8417280 | consumed tokens: 17238589440 | elapsed time per iteration (s): 1.74 | learning rate: 1.614E-04 | global batch size: 512 | lm loss: 1.442093E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.457 | TFLOPs: 44.05 | 31: iteration 16450/ 52452 | consumed samples: 8422400 | consumed tokens: 17249075200 | elapsed time per iteration (s): 1.74 | learning rate: 1.614E-04 | global batch size: 512 | lm loss: 1.434723E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.957 | TFLOPs: 44.12 | 31: iteration 16460/ 52452 | consumed samples: 8427520 | consumed tokens: 17259560960 | elapsed time per iteration (s): 1.72 | learning rate: 1.613E-04 | global batch size: 512 | lm loss: 1.438030E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.920 | TFLOPs: 44.57 | 31: iteration 16470/ 52452 | consumed samples: 8432640 | consumed tokens: 17270046720 | elapsed time per iteration (s): 1.74 | learning rate: 1.613E-04 | global batch size: 512 | lm loss: 1.452391E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.976 | TFLOPs: 44.27 | 31: iteration 16480/ 52452 | consumed samples: 8437760 | consumed tokens: 17280532480 | elapsed time per iteration (s): 1.78 | learning rate: 1.612E-04 | global batch size: 512 | lm loss: 1.435664E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.043 | TFLOPs: 43.23 | 31: iteration 16490/ 52452 | consumed samples: 8442880 | consumed tokens: 17291018240 | elapsed time per iteration (s): 1.77 | learning rate: 1.612E-04 | global batch size: 512 | lm loss: 1.446190E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.832 | TFLOPs: 43.35 | 31: iteration 16500/ 52452 | consumed samples: 8448000 | consumed tokens: 17301504000 | elapsed time per iteration (s): 1.83 | learning rate: 1.611E-04 | global batch size: 512 | lm loss: 1.442888E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.055 | TFLOPs: 42.03 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16500 | lm loss value: 4.082606E+00 | lm loss PPL: 5.929982E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16500 | lm loss value: 4.062874E+00 | lm loss PPL: 5.814116E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16510/ 52452 | consumed samples: 8453120 | consumed tokens: 17311989760 | elapsed time per iteration (s): 9.13 | learning rate: 1.611E-04 | global batch size: 512 | lm loss: 1.440728E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.109 | TFLOPs: 8.42 | 31: iteration 16520/ 52452 | consumed samples: 8458240 | consumed tokens: 17322475520 | elapsed time per iteration (s): 1.75 | learning rate: 1.610E-04 | global batch size: 512 | lm loss: 1.449863E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.146 | TFLOPs: 44.00 | 31: iteration 16530/ 52452 | consumed samples: 8463360 | consumed tokens: 17332961280 | elapsed time per iteration (s): 1.75 | learning rate: 1.610E-04 | global batch size: 512 | lm loss: 1.444049E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.237 | TFLOPs: 44.01 | 31: iteration 16540/ 52452 | consumed samples: 8468480 | consumed tokens: 17343447040 | elapsed time per iteration (s): 1.73 | learning rate: 1.610E-04 | global batch size: 512 | lm loss: 1.431900E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.239 | TFLOPs: 44.31 | 31: iteration 16550/ 52452 | consumed samples: 8473600 | consumed tokens: 17353932800 | elapsed time per iteration (s): 1.78 | learning rate: 1.609E-04 | global batch size: 512 | lm loss: 1.439795E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.259 | TFLOPs: 43.12 | 31: iteration 16560/ 52452 | consumed samples: 8478720 | consumed tokens: 17364418560 | elapsed time per iteration (s): 1.83 | learning rate: 1.609E-04 | global batch size: 512 | lm loss: 1.442565E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.479 | TFLOPs: 42.10 | 31: iteration 16570/ 52452 | consumed samples: 8483840 | consumed tokens: 17374904320 | elapsed time per iteration (s): 1.86 | learning rate: 1.608E-04 | global batch size: 512 | lm loss: 1.447389E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.140 | TFLOPs: 41.30 | 31: iteration 16580/ 52452 | consumed samples: 8488960 | consumed tokens: 17385390080 | elapsed time per iteration (s): 1.74 | learning rate: 1.608E-04 | global batch size: 512 | lm loss: 1.448348E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.866 | TFLOPs: 44.11 | 31: iteration 16590/ 52452 | consumed samples: 8494080 | consumed tokens: 17395875840 | elapsed time per iteration (s): 1.77 | learning rate: 1.607E-04 | global batch size: 512 | lm loss: 1.440606E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.763 | TFLOPs: 43.49 | 31: iteration 16600/ 52452 | consumed samples: 8499200 | consumed tokens: 17406361600 | elapsed time per iteration (s): 1.74 | learning rate: 1.607E-04 | global batch size: 512 | lm loss: 1.437860E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.346 | TFLOPs: 44.18 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16600 | lm loss value: 4.086290E+00 | lm loss PPL: 5.951869E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16600 | lm loss value: 4.067980E+00 | lm loss PPL: 5.843881E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16610/ 52452 | consumed samples: 8504320 | consumed tokens: 17416847360 | elapsed time per iteration (s): 9.14 | learning rate: 1.606E-04 | global batch size: 512 | lm loss: 1.444221E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.033 | TFLOPs: 8.41 | 31: iteration 16620/ 52452 | consumed samples: 8509440 | consumed tokens: 17427333120 | elapsed time per iteration (s): 1.78 | learning rate: 1.606E-04 | global batch size: 512 | lm loss: 1.443690E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.012 | TFLOPs: 43.23 | 31: iteration 16630/ 52452 | consumed samples: 8514560 | consumed tokens: 17437818880 | elapsed time per iteration (s): 1.75 | learning rate: 1.606E-04 | global batch size: 512 | lm loss: 1.437540E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.381 | TFLOPs: 43.88 | 31: iteration 16640/ 52452 | consumed samples: 8519680 | consumed tokens: 17448304640 | elapsed time per iteration (s): 1.77 | learning rate: 1.605E-04 | global batch size: 512 | lm loss: 1.441325E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.153 | TFLOPs: 43.40 | 31: iteration 16650/ 52452 | consumed samples: 8524800 | consumed tokens: 17458790400 | elapsed time per iteration (s): 1.72 | learning rate: 1.605E-04 | global batch size: 512 | lm loss: 1.439964E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.028 | TFLOPs: 44.58 | 31: iteration 16660/ 52452 | consumed samples: 8529920 | consumed tokens: 17469276160 | elapsed time per iteration (s): 1.79 | learning rate: 1.604E-04 | global batch size: 512 | lm loss: 1.452754E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.291 | TFLOPs: 42.97 | 31: iteration 16670/ 52452 | consumed samples: 8535040 | consumed tokens: 17479761920 | elapsed time per iteration (s): 1.75 | learning rate: 1.604E-04 | global batch size: 512 | lm loss: 1.446447E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.298 | TFLOPs: 44.02 | 31: iteration 16680/ 52452 | consumed samples: 8540160 | consumed tokens: 17490247680 | elapsed time per iteration (s): 1.78 | learning rate: 1.603E-04 | global batch size: 512 | lm loss: 1.432845E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.954 | TFLOPs: 43.07 | 31: iteration 16690/ 52452 | consumed samples: 8545280 | consumed tokens: 17500733440 | elapsed time per iteration (s): 1.73 | learning rate: 1.603E-04 | global batch size: 512 | lm loss: 1.429130E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.224 | TFLOPs: 44.46 | 31: iteration 16700/ 52452 | consumed samples: 8550400 | consumed tokens: 17511219200 | elapsed time per iteration (s): 1.82 | learning rate: 1.602E-04 | global batch size: 512 | lm loss: 1.425914E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.614 | TFLOPs: 42.27 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16700 | lm loss value: 4.236827E+00 | lm loss PPL: 6.918800E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16700 | lm loss value: 4.168085E+00 | lm loss PPL: 6.459165E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16710/ 52452 | consumed samples: 8555520 | consumed tokens: 17521704960 | elapsed time per iteration (s): 9.15 | learning rate: 1.602E-04 | global batch size: 512 | lm loss: 1.444135E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.943 | TFLOPs: 8.40 | 31: iteration 16720/ 52452 | consumed samples: 8560640 | consumed tokens: 17532190720 | elapsed time per iteration (s): 1.80 | learning rate: 1.601E-04 | global batch size: 512 | lm loss: 1.431329E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.493 | TFLOPs: 42.70 | 31: iteration 16730/ 52452 | consumed samples: 8565760 | consumed tokens: 17542676480 | elapsed time per iteration (s): 1.81 | learning rate: 1.601E-04 | global batch size: 512 | lm loss: 1.429991E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.654 | TFLOPs: 42.42 | 31: iteration 16740/ 52452 | consumed samples: 8570880 | consumed tokens: 17553162240 | elapsed time per iteration (s): 1.79 | learning rate: 1.601E-04 | global batch size: 512 | lm loss: 1.444002E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.388 | TFLOPs: 42.99 | 31: iteration 16750/ 52452 | consumed samples: 8576000 | consumed tokens: 17563648000 | elapsed time per iteration (s): 1.81 | learning rate: 1.600E-04 | global batch size: 512 | lm loss: 1.434151E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.273 | TFLOPs: 42.37 | 31: iteration 16760/ 52452 | consumed samples: 8581120 | consumed tokens: 17574133760 | elapsed time per iteration (s): 1.74 | learning rate: 1.600E-04 | global batch size: 512 | lm loss: 1.429906E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.267 | TFLOPs: 44.17 | 31: iteration 16770/ 52452 | consumed samples: 8586240 | consumed tokens: 17584619520 | elapsed time per iteration (s): 1.78 | learning rate: 1.599E-04 | global batch size: 512 | lm loss: 1.438233E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.256 | TFLOPs: 43.27 | 31: iteration 16780/ 52452 | consumed samples: 8591360 | consumed tokens: 17595105280 | elapsed time per iteration (s): 1.83 | learning rate: 1.599E-04 | global batch size: 512 | lm loss: 1.437983E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.442 | TFLOPs: 41.94 | 31: iteration 16790/ 52452 | consumed samples: 8596480 | consumed tokens: 17605591040 | elapsed time per iteration (s): 1.75 | learning rate: 1.598E-04 | global batch size: 512 | lm loss: 1.435527E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.331 | TFLOPs: 43.88 | 31: iteration 16800/ 52452 | consumed samples: 8601600 | consumed tokens: 17616076800 | elapsed time per iteration (s): 1.74 | learning rate: 1.598E-04 | global batch size: 512 | lm loss: 1.446138E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.983 | TFLOPs: 44.28 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16800 | lm loss value: 4.162879E+00 | lm loss PPL: 6.425628E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16800 | lm loss value: 4.099966E+00 | lm loss PPL: 6.033821E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16810/ 52452 | consumed samples: 8606720 | consumed tokens: 17626562560 | elapsed time per iteration (s): 9.19 | learning rate: 1.597E-04 | global batch size: 512 | lm loss: 1.446093E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.698 | TFLOPs: 8.36 | 31: iteration 16820/ 52452 | consumed samples: 8611840 | consumed tokens: 17637048320 | elapsed time per iteration (s): 1.72 | learning rate: 1.597E-04 | global batch size: 512 | lm loss: 1.433344E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.849 | TFLOPs: 44.56 | 31: iteration 16830/ 52452 | consumed samples: 8616960 | consumed tokens: 17647534080 | elapsed time per iteration (s): 1.75 | learning rate: 1.596E-04 | global batch size: 512 | lm loss: 1.449375E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.968 | TFLOPs: 43.82 | 31: iteration 16840/ 52452 | consumed samples: 8622080 | consumed tokens: 17658019840 | elapsed time per iteration (s): 1.73 | learning rate: 1.596E-04 | global batch size: 512 | lm loss: 1.437881E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.686 | TFLOPs: 44.38 | 31: iteration 16850/ 52452 | consumed samples: 8627200 | consumed tokens: 17668505600 | elapsed time per iteration (s): 1.72 | learning rate: 1.596E-04 | global batch size: 512 | lm loss: 1.433016E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.174 | TFLOPs: 44.60 | 31: iteration 16860/ 52452 | consumed samples: 8632320 | consumed tokens: 17678991360 | elapsed time per iteration (s): 1.81 | learning rate: 1.595E-04 | global batch size: 512 | lm loss: 1.432261E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.564 | TFLOPs: 42.41 | 31: iteration 16870/ 52452 | consumed samples: 8637440 | consumed tokens: 17689477120 | elapsed time per iteration (s): 1.73 | learning rate: 1.595E-04 | global batch size: 512 | lm loss: 1.447545E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.494 | TFLOPs: 44.35 | 31: iteration 16880/ 52452 | consumed samples: 8642560 | consumed tokens: 17699962880 | elapsed time per iteration (s): 1.76 | learning rate: 1.594E-04 | global batch size: 512 | lm loss: 1.425579E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.159 | TFLOPs: 43.70 | 31: iteration 16890/ 52452 | consumed samples: 8647680 | consumed tokens: 17710448640 | elapsed time per iteration (s): 1.73 | learning rate: 1.594E-04 | global batch size: 512 | lm loss: 1.440385E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.169 | TFLOPs: 44.30 | 31: iteration 16900/ 52452 | consumed samples: 8652800 | consumed tokens: 17720934400 | elapsed time per iteration (s): 1.73 | learning rate: 1.593E-04 | global batch size: 512 | lm loss: 1.427532E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.924 | TFLOPs: 44.42 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 16900 | lm loss value: 4.032022E+00 | lm loss PPL: 5.637481E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 16900 | lm loss value: 4.029409E+00 | lm loss PPL: 5.622767E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 16910/ 52452 | consumed samples: 8657920 | consumed tokens: 17731420160 | elapsed time per iteration (s): 9.15 | learning rate: 1.593E-04 | global batch size: 512 | lm loss: 1.429138E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.929 | TFLOPs: 8.39 | 31: iteration 16920/ 52452 | consumed samples: 8663040 | consumed tokens: 17741905920 | elapsed time per iteration (s): 1.75 | learning rate: 1.592E-04 | global batch size: 512 | lm loss: 1.433102E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.465 | TFLOPs: 43.90 | 31: iteration 16930/ 52452 | consumed samples: 8668160 | consumed tokens: 17752391680 | elapsed time per iteration (s): 1.77 | learning rate: 1.592E-04 | global batch size: 512 | lm loss: 1.440454E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.927 | TFLOPs: 43.52 | 31: iteration 16940/ 52452 | consumed samples: 8673280 | consumed tokens: 17762877440 | elapsed time per iteration (s): 1.76 | learning rate: 1.591E-04 | global batch size: 512 | lm loss: 1.432045E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.173 | TFLOPs: 43.70 | 31: iteration 16950/ 52452 | consumed samples: 8678400 | consumed tokens: 17773363200 | elapsed time per iteration (s): 1.74 | learning rate: 1.591E-04 | global batch size: 512 | lm loss: 1.430317E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.505 | TFLOPs: 44.20 | 31: iteration 16960/ 52452 | consumed samples: 8683520 | consumed tokens: 17783848960 | elapsed time per iteration (s): 1.74 | learning rate: 1.591E-04 | global batch size: 512 | lm loss: 1.433448E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.453 | TFLOPs: 44.20 | 31: iteration 16970/ 52452 | consumed samples: 8688640 | consumed tokens: 17794334720 | elapsed time per iteration (s): 1.75 | learning rate: 1.590E-04 | global batch size: 512 | lm loss: 1.436269E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.124 | TFLOPs: 44.00 | 31: iteration 16980/ 52452 | consumed samples: 8693760 | consumed tokens: 17804820480 | elapsed time per iteration (s): 1.72 | learning rate: 1.590E-04 | global batch size: 512 | lm loss: 1.428595E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.341 | TFLOPs: 44.63 | 31: iteration 16990/ 52452 | consumed samples: 8698880 | consumed tokens: 17815306240 | elapsed time per iteration (s): 1.77 | learning rate: 1.589E-04 | global batch size: 512 | lm loss: 1.438227E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.295 | TFLOPs: 43.42 | 31: iteration 17000/ 52452 | consumed samples: 8704000 | consumed tokens: 17825792000 | elapsed time per iteration (s): 1.79 | learning rate: 1.589E-04 | global batch size: 512 | lm loss: 1.432551E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.482 | TFLOPs: 43.00 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17000 | lm loss value: 4.186368E+00 | lm loss PPL: 6.578346E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17000 | lm loss value: 4.161473E+00 | lm loss PPL: 6.416596E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17010/ 52452 | consumed samples: 8709120 | consumed tokens: 17836277760 | elapsed time per iteration (s): 9.17 | learning rate: 1.588E-04 | global batch size: 512 | lm loss: 1.424689E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.828 | TFLOPs: 8.38 | 31: iteration 17020/ 52452 | consumed samples: 8714240 | consumed tokens: 17846763520 | elapsed time per iteration (s): 1.73 | learning rate: 1.588E-04 | global batch size: 512 | lm loss: 1.442001E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.766 | TFLOPs: 44.39 | 31: iteration 17030/ 52452 | consumed samples: 8719360 | consumed tokens: 17857249280 | elapsed time per iteration (s): 1.74 | learning rate: 1.587E-04 | global batch size: 512 | lm loss: 1.419176E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.849 | TFLOPs: 44.26 | 31: iteration 17040/ 52452 | consumed samples: 8724480 | consumed tokens: 17867735040 | elapsed time per iteration (s): 1.76 | learning rate: 1.587E-04 | global batch size: 512 | lm loss: 1.423585E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.216 | TFLOPs: 43.71 | 31: iteration 17050/ 52452 | consumed samples: 8729600 | consumed tokens: 17878220800 | elapsed time per iteration (s): 1.75 | learning rate: 1.586E-04 | global batch size: 512 | lm loss: 1.424553E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.803 | TFLOPs: 43.80 | 31: iteration 17060/ 52452 | consumed samples: 8734720 | consumed tokens: 17888706560 | elapsed time per iteration (s): 1.74 | learning rate: 1.586E-04 | global batch size: 512 | lm loss: 1.420312E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.902 | TFLOPs: 44.11 | 31: iteration 17070/ 52452 | consumed samples: 8739840 | consumed tokens: 17899192320 | elapsed time per iteration (s): 1.73 | learning rate: 1.586E-04 | global batch size: 512 | lm loss: 1.424627E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.250 | TFLOPs: 44.47 | 31: iteration 17080/ 52452 | consumed samples: 8744960 | consumed tokens: 17909678080 | elapsed time per iteration (s): 1.78 | learning rate: 1.585E-04 | global batch size: 512 | lm loss: 1.443261E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.388 | TFLOPs: 43.29 | 31: iteration 17090/ 52452 | consumed samples: 8750080 | consumed tokens: 17920163840 | elapsed time per iteration (s): 1.75 | learning rate: 1.585E-04 | global batch size: 512 | lm loss: 1.437884E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.266 | TFLOPs: 43.87 | 31: iteration 17100/ 52452 | consumed samples: 8755200 | consumed tokens: 17930649600 | elapsed time per iteration (s): 1.79 | learning rate: 1.584E-04 | global batch size: 512 | lm loss: 1.431327E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.754 | TFLOPs: 42.89 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17100 | lm loss value: 4.197147E+00 | lm loss PPL: 6.649634E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17100 | lm loss value: 4.163978E+00 | lm loss PPL: 6.432691E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17110/ 52452 | consumed samples: 8760320 | consumed tokens: 17941135360 | elapsed time per iteration (s): 9.16 | learning rate: 1.584E-04 | global batch size: 512 | lm loss: 1.436131E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.868 | TFLOPs: 8.39 | 31: iteration 17120/ 52452 | consumed samples: 8765440 | consumed tokens: 17951621120 | elapsed time per iteration (s): 1.75 | learning rate: 1.583E-04 | global batch size: 512 | lm loss: 1.421050E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.792 | TFLOPs: 43.95 | 31: iteration 17130/ 52452 | consumed samples: 8770560 | consumed tokens: 17962106880 | elapsed time per iteration (s): 1.78 | learning rate: 1.583E-04 | global batch size: 512 | lm loss: 1.424287E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.964 | TFLOPs: 43.22 | 31: iteration 17140/ 52452 | consumed samples: 8775680 | consumed tokens: 17972592640 | elapsed time per iteration (s): 1.76 | learning rate: 1.582E-04 | global batch size: 512 | lm loss: 1.428153E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.476 | TFLOPs: 43.75 | 31: iteration 17150/ 52452 | consumed samples: 8780800 | consumed tokens: 17983078400 | elapsed time per iteration (s): 1.80 | learning rate: 1.582E-04 | global batch size: 512 | lm loss: 1.427654E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.720 | TFLOPs: 42.58 | 31: iteration 17160/ 52452 | consumed samples: 8785920 | consumed tokens: 17993564160 | elapsed time per iteration (s): 1.72 | learning rate: 1.581E-04 | global batch size: 512 | lm loss: 1.422362E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.538 | TFLOPs: 44.66 | 31: iteration 17170/ 52452 | consumed samples: 8791040 | consumed tokens: 18004049920 | elapsed time per iteration (s): 1.76 | learning rate: 1.581E-04 | global batch size: 512 | lm loss: 1.431531E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.421 | TFLOPs: 43.59 | 31: iteration 17180/ 52452 | consumed samples: 8796160 | consumed tokens: 18014535680 | elapsed time per iteration (s): 1.90 | learning rate: 1.580E-04 | global batch size: 512 | lm loss: 1.450033E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.900 | TFLOPs: 40.36 | 31: iteration 17190/ 52452 | consumed samples: 8801280 | consumed tokens: 18025021440 | elapsed time per iteration (s): 1.78 | learning rate: 1.580E-04 | global batch size: 512 | lm loss: 1.431555E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.081 | TFLOPs: 43.09 | 31: iteration 17200/ 52452 | consumed samples: 8806400 | consumed tokens: 18035507200 | elapsed time per iteration (s): 1.74 | learning rate: 1.580E-04 | global batch size: 512 | lm loss: 1.431864E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.926 | TFLOPs: 44.12 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17200 | lm loss value: 4.183019E+00 | lm loss PPL: 6.556347E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17200 | lm loss value: 4.121400E+00 | lm loss PPL: 6.164548E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17210/ 52452 | consumed samples: 8811520 | consumed tokens: 18045992960 | elapsed time per iteration (s): 9.18 | learning rate: 1.579E-04 | global batch size: 512 | lm loss: 1.424328E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.750 | TFLOPs: 8.37 | 31: iteration 17220/ 52452 | consumed samples: 8816640 | consumed tokens: 18056478720 | elapsed time per iteration (s): 1.75 | learning rate: 1.579E-04 | global batch size: 512 | lm loss: 1.427783E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.720 | TFLOPs: 43.94 | 31: iteration 17230/ 52452 | consumed samples: 8821760 | consumed tokens: 18066964480 | elapsed time per iteration (s): 1.76 | learning rate: 1.578E-04 | global batch size: 512 | lm loss: 1.436975E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.461 | TFLOPs: 43.75 | 31: iteration 17240/ 52452 | consumed samples: 8826880 | consumed tokens: 18077450240 | elapsed time per iteration (s): 1.73 | learning rate: 1.578E-04 | global batch size: 512 | lm loss: 1.426126E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.443 | TFLOPs: 44.49 | 31: iteration 17250/ 52452 | consumed samples: 8832000 | consumed tokens: 18087936000 | elapsed time per iteration (s): 1.74 | learning rate: 1.577E-04 | global batch size: 512 | lm loss: 1.432123E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.237 | TFLOPs: 44.16 | 31: iteration 17260/ 52452 | consumed samples: 8837120 | consumed tokens: 18098421760 | elapsed time per iteration (s): 1.76 | learning rate: 1.577E-04 | global batch size: 512 | lm loss: 1.432630E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.288 | TFLOPs: 43.72 | 31: iteration 17270/ 52452 | consumed samples: 8842240 | consumed tokens: 18108907520 | elapsed time per iteration (s): 1.74 | learning rate: 1.576E-04 | global batch size: 512 | lm loss: 1.428843E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.047 | TFLOPs: 44.13 | 31: iteration 17280/ 52452 | consumed samples: 8847360 | consumed tokens: 18119393280 | elapsed time per iteration (s): 1.75 | learning rate: 1.576E-04 | global batch size: 512 | lm loss: 1.433832E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.558 | TFLOPs: 43.91 | 31: iteration 17290/ 52452 | consumed samples: 8852480 | consumed tokens: 18129879040 | elapsed time per iteration (s): 1.78 | learning rate: 1.575E-04 | global batch size: 512 | lm loss: 1.419589E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.259 | TFLOPs: 43.12 | 31: iteration 17300/ 52452 | consumed samples: 8857600 | consumed tokens: 18140364800 | elapsed time per iteration (s): 1.74 | learning rate: 1.575E-04 | global batch size: 512 | lm loss: 1.429327E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.882 | TFLOPs: 44.11 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17300 | lm loss value: 4.066175E+00 | lm loss PPL: 5.833341E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17300 | lm loss value: 4.030128E+00 | lm loss PPL: 5.626814E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17310/ 52452 | consumed samples: 8862720 | consumed tokens: 18150850560 | elapsed time per iteration (s): 9.15 | learning rate: 1.574E-04 | global batch size: 512 | lm loss: 1.422076E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.960 | TFLOPs: 8.40 | 31: iteration 17320/ 52452 | consumed samples: 8867840 | consumed tokens: 18161336320 | elapsed time per iteration (s): 1.75 | learning rate: 1.574E-04 | global batch size: 512 | lm loss: 1.418858E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.457 | TFLOPs: 43.90 | 31: iteration 17330/ 52452 | consumed samples: 8872960 | consumed tokens: 18171822080 | elapsed time per iteration (s): 1.72 | learning rate: 1.574E-04 | global batch size: 512 | lm loss: 1.438241E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.464 | TFLOPs: 44.65 | 31: iteration 17340/ 52452 | consumed samples: 8878080 | consumed tokens: 18182307840 | elapsed time per iteration (s): 1.74 | learning rate: 1.573E-04 | global batch size: 512 | lm loss: 1.423738E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.659 | TFLOPs: 44.08 | 31: iteration 17350/ 52452 | consumed samples: 8883200 | consumed tokens: 18192793600 | elapsed time per iteration (s): 1.76 | learning rate: 1.573E-04 | global batch size: 512 | lm loss: 1.437401E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.308 | TFLOPs: 43.57 | 31: iteration 17360/ 52452 | consumed samples: 8888320 | consumed tokens: 18203279360 | elapsed time per iteration (s): 1.74 | learning rate: 1.572E-04 | global batch size: 512 | lm loss: 1.427028E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.185 | TFLOPs: 44.16 | 31: iteration 17370/ 52452 | consumed samples: 8893440 | consumed tokens: 18213765120 | elapsed time per iteration (s): 1.72 | learning rate: 1.572E-04 | global batch size: 512 | lm loss: 1.425085E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.483 | TFLOPs: 44.65 | 31: iteration 17380/ 52452 | consumed samples: 8898560 | consumed tokens: 18224250880 | elapsed time per iteration (s): 1.74 | learning rate: 1.571E-04 | global batch size: 512 | lm loss: 1.428527E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.993 | TFLOPs: 44.28 | 31: iteration 17390/ 52452 | consumed samples: 8903680 | consumed tokens: 18234736640 | elapsed time per iteration (s): 1.73 | learning rate: 1.571E-04 | global batch size: 512 | lm loss: 1.425081E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.546 | TFLOPs: 44.36 | 31: iteration 17400/ 52452 | consumed samples: 8908800 | consumed tokens: 18245222400 | elapsed time per iteration (s): 1.77 | learning rate: 1.570E-04 | global batch size: 512 | lm loss: 1.424970E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.856 | TFLOPs: 43.36 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17400 | lm loss value: 4.116670E+00 | lm loss PPL: 6.135460E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17400 | lm loss value: 4.081273E+00 | lm loss PPL: 5.922079E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17410/ 52452 | consumed samples: 8913920 | consumed tokens: 18255708160 | elapsed time per iteration (s): 14.71 | learning rate: 1.570E-04 | global batch size: 512 | lm loss: 1.426567E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 34.811 | TFLOPs: 5.22 | 31: iteration 17420/ 52452 | consumed samples: 8919040 | consumed tokens: 18266193920 | elapsed time per iteration (s): 1.75 | learning rate: 1.569E-04 | global batch size: 512 | lm loss: 1.429378E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.989 | TFLOPs: 43.83 | 31: iteration 17430/ 52452 | consumed samples: 8924160 | consumed tokens: 18276679680 | elapsed time per iteration (s): 1.83 | learning rate: 1.569E-04 | global batch size: 512 | lm loss: 1.435760E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.382 | TFLOPs: 41.93 | 31: iteration 17440/ 52452 | consumed samples: 8929280 | consumed tokens: 18287165440 | elapsed time per iteration (s): 1.75 | learning rate: 1.568E-04 | global batch size: 512 | lm loss: 1.425500E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.730 | TFLOPs: 43.94 | 31: iteration 17450/ 52452 | consumed samples: 8934400 | consumed tokens: 18297651200 | elapsed time per iteration (s): 1.78 | learning rate: 1.568E-04 | global batch size: 512 | lm loss: 1.432006E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.211 | TFLOPs: 43.11 | 31: iteration 17460/ 52452 | consumed samples: 8939520 | consumed tokens: 18308136960 | elapsed time per iteration (s): 1.80 | learning rate: 1.568E-04 | global batch size: 512 | lm loss: 1.424884E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.035 | TFLOPs: 42.78 | 31: iteration 17470/ 52452 | consumed samples: 8944640 | consumed tokens: 18318622720 | elapsed time per iteration (s): 1.73 | learning rate: 1.567E-04 | global batch size: 512 | lm loss: 1.416986E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.281 | TFLOPs: 44.47 | 31: iteration 17480/ 52452 | consumed samples: 8949760 | consumed tokens: 18329108480 | elapsed time per iteration (s): 1.79 | learning rate: 1.567E-04 | global batch size: 512 | lm loss: 1.427349E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.282 | TFLOPs: 42.97 | 31: iteration 17490/ 52452 | consumed samples: 8954880 | consumed tokens: 18339594240 | elapsed time per iteration (s): 1.72 | learning rate: 1.566E-04 | global batch size: 512 | lm loss: 1.418384E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.488 | TFLOPs: 44.65 | 31: iteration 17500/ 52452 | consumed samples: 8960000 | consumed tokens: 18350080000 | elapsed time per iteration (s): 1.77 | learning rate: 1.566E-04 | global batch size: 512 | lm loss: 1.430958E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.884 | TFLOPs: 43.51 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17500 | lm loss value: 4.099281E+00 | lm loss PPL: 6.029691E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17500 | lm loss value: 4.052873E+00 | lm loss PPL: 5.756258E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17510/ 52452 | consumed samples: 8965120 | consumed tokens: 18360565760 | elapsed time per iteration (s): 9.13 | learning rate: 1.565E-04 | global batch size: 512 | lm loss: 1.434084E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.083 | TFLOPs: 8.42 | 31: iteration 17520/ 52452 | consumed samples: 8970240 | consumed tokens: 18371051520 | elapsed time per iteration (s): 1.73 | learning rate: 1.565E-04 | global batch size: 512 | lm loss: 1.419703E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.681 | TFLOPs: 44.38 | 31: iteration 17530/ 52452 | consumed samples: 8975360 | consumed tokens: 18381537280 | elapsed time per iteration (s): 1.75 | learning rate: 1.564E-04 | global batch size: 512 | lm loss: 1.427917E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.318 | TFLOPs: 44.03 | 31: iteration 17540/ 52452 | consumed samples: 8980480 | consumed tokens: 18392023040 | elapsed time per iteration (s): 1.74 | learning rate: 1.564E-04 | global batch size: 512 | lm loss: 1.428101E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.911 | TFLOPs: 44.11 | 31: iteration 17550/ 52452 | consumed samples: 8985600 | consumed tokens: 18402508800 | elapsed time per iteration (s): 1.80 | learning rate: 1.563E-04 | global batch size: 512 | lm loss: 1.436058E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.157 | TFLOPs: 42.80 | 31: iteration 17560/ 52452 | consumed samples: 8990720 | consumed tokens: 18412994560 | elapsed time per iteration (s): 1.85 | learning rate: 1.563E-04 | global batch size: 512 | lm loss: 1.425772E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.358 | TFLOPs: 41.63 | 31: iteration 17570/ 52452 | consumed samples: 8995840 | consumed tokens: 18423480320 | elapsed time per iteration (s): 1.73 | learning rate: 1.562E-04 | global batch size: 512 | lm loss: 1.429882E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.632 | TFLOPs: 44.37 | 31: iteration 17580/ 52452 | consumed samples: 9000960 | consumed tokens: 18433966080 | elapsed time per iteration (s): 1.74 | learning rate: 1.562E-04 | global batch size: 512 | lm loss: 1.434446E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.636 | TFLOPs: 44.07 | 31: iteration 17590/ 52452 | consumed samples: 9006080 | consumed tokens: 18444451840 | elapsed time per iteration (s): 1.72 | learning rate: 1.561E-04 | global batch size: 512 | lm loss: 1.426706E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.452 | TFLOPs: 44.65 | 31: iteration 17600/ 52452 | consumed samples: 9011200 | consumed tokens: 18454937600 | elapsed time per iteration (s): 1.75 | learning rate: 1.561E-04 | global batch size: 512 | lm loss: 1.426696E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.025 | TFLOPs: 43.83 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17600 | lm loss value: 4.083965E+00 | lm loss PPL: 5.938047E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17600 | lm loss value: 4.041841E+00 | lm loss PPL: 5.693103E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17610/ 52452 | consumed samples: 9016320 | consumed tokens: 18465423360 | elapsed time per iteration (s): 9.14 | learning rate: 1.561E-04 | global batch size: 512 | lm loss: 1.423451E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.038 | TFLOPs: 8.41 | 31: iteration 17620/ 52452 | consumed samples: 9021440 | consumed tokens: 18475909120 | elapsed time per iteration (s): 1.74 | learning rate: 1.560E-04 | global batch size: 512 | lm loss: 1.431971E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.771 | TFLOPs: 44.09 | 31: iteration 17630/ 52452 | consumed samples: 9026560 | consumed tokens: 18486394880 | elapsed time per iteration (s): 1.73 | learning rate: 1.560E-04 | global batch size: 512 | lm loss: 1.417766E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.708 | TFLOPs: 44.38 | 31: iteration 17640/ 52452 | consumed samples: 9031680 | consumed tokens: 18496880640 | elapsed time per iteration (s): 1.73 | learning rate: 1.559E-04 | global batch size: 512 | lm loss: 1.431710E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.529 | TFLOPs: 44.36 | 31: iteration 17650/ 52452 | consumed samples: 9036800 | consumed tokens: 18507366400 | elapsed time per iteration (s): 1.74 | learning rate: 1.559E-04 | global batch size: 512 | lm loss: 1.428771E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.649 | TFLOPs: 44.08 | 31: iteration 17660/ 52452 | consumed samples: 9041920 | consumed tokens: 18517852160 | elapsed time per iteration (s): 1.75 | learning rate: 1.558E-04 | global batch size: 512 | lm loss: 1.417661E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.250 | TFLOPs: 44.02 | 31: iteration 17670/ 52452 | consumed samples: 9047040 | consumed tokens: 18528337920 | elapsed time per iteration (s): 1.73 | learning rate: 1.558E-04 | global batch size: 512 | lm loss: 1.432771E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.970 | TFLOPs: 44.42 | 31: iteration 17680/ 52452 | consumed samples: 9052160 | consumed tokens: 18538823680 | elapsed time per iteration (s): 1.77 | learning rate: 1.557E-04 | global batch size: 512 | lm loss: 1.423900E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.469 | TFLOPs: 43.45 | 31: iteration 17690/ 52452 | consumed samples: 9057280 | consumed tokens: 18549309440 | elapsed time per iteration (s): 1.75 | learning rate: 1.557E-04 | global batch size: 512 | lm loss: 1.418807E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.703 | TFLOPs: 43.93 | 31: iteration 17700/ 52452 | consumed samples: 9062400 | consumed tokens: 18559795200 | elapsed time per iteration (s): 1.76 | learning rate: 1.556E-04 | global batch size: 512 | lm loss: 1.418473E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.058 | TFLOPs: 43.69 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17700 | lm loss value: 4.128532E+00 | lm loss PPL: 6.208671E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17700 | lm loss value: 4.066656E+00 | lm loss PPL: 5.836145E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17710/ 52452 | consumed samples: 9067520 | consumed tokens: 18570280960 | elapsed time per iteration (s): 9.15 | learning rate: 1.556E-04 | global batch size: 512 | lm loss: 1.424467E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.936 | TFLOPs: 8.40 | 31: iteration 17720/ 52452 | consumed samples: 9072640 | consumed tokens: 18580766720 | elapsed time per iteration (s): 1.77 | learning rate: 1.555E-04 | global batch size: 512 | lm loss: 1.428455E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.360 | TFLOPs: 43.43 | 31: iteration 17730/ 52452 | consumed samples: 9077760 | consumed tokens: 18591252480 | elapsed time per iteration (s): 1.77 | learning rate: 1.555E-04 | global batch size: 512 | lm loss: 1.407450E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.533 | TFLOPs: 43.46 | 31: iteration 17740/ 52452 | consumed samples: 9082880 | consumed tokens: 18601738240 | elapsed time per iteration (s): 1.76 | learning rate: 1.554E-04 | global batch size: 512 | lm loss: 1.415409E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.649 | TFLOPs: 43.62 | 31: iteration 17750/ 52452 | consumed samples: 9088000 | consumed tokens: 18612224000 | elapsed time per iteration (s): 1.81 | learning rate: 1.554E-04 | global batch size: 512 | lm loss: 1.417235E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.245 | TFLOPs: 42.36 | 31: iteration 17760/ 52452 | consumed samples: 9093120 | consumed tokens: 18622709760 | elapsed time per iteration (s): 1.81 | learning rate: 1.553E-04 | global batch size: 512 | lm loss: 1.417114E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.102 | TFLOPs: 42.49 | 31: iteration 17770/ 52452 | consumed samples: 9098240 | consumed tokens: 18633195520 | elapsed time per iteration (s): 1.79 | learning rate: 1.553E-04 | global batch size: 512 | lm loss: 1.439900E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.243 | TFLOPs: 42.81 | 31: iteration 17780/ 52452 | consumed samples: 9103360 | consumed tokens: 18643681280 | elapsed time per iteration (s): 1.77 | learning rate: 1.553E-04 | global batch size: 512 | lm loss: 1.418379E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.674 | TFLOPs: 43.33 | 31: iteration 17790/ 52452 | consumed samples: 9108480 | consumed tokens: 18654167040 | elapsed time per iteration (s): 1.77 | learning rate: 1.552E-04 | global batch size: 512 | lm loss: 1.422511E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.958 | TFLOPs: 43.52 | 31: iteration 17800/ 52452 | consumed samples: 9113600 | consumed tokens: 18664652800 | elapsed time per iteration (s): 1.79 | learning rate: 1.552E-04 | global batch size: 512 | lm loss: 1.415382E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.097 | TFLOPs: 42.94 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17800 | lm loss value: 4.090397E+00 | lm loss PPL: 5.976363E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17800 | lm loss value: 4.061297E+00 | lm loss PPL: 5.804958E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17810/ 52452 | consumed samples: 9118720 | consumed tokens: 18675138560 | elapsed time per iteration (s): 9.16 | learning rate: 1.551E-04 | global batch size: 512 | lm loss: 1.431649E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.889 | TFLOPs: 8.39 | 31: iteration 17820/ 52452 | consumed samples: 9123840 | consumed tokens: 18685624320 | elapsed time per iteration (s): 1.77 | learning rate: 1.551E-04 | global batch size: 512 | lm loss: 1.431861E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.020 | TFLOPs: 43.38 | 31: iteration 17830/ 52452 | consumed samples: 9128960 | consumed tokens: 18696110080 | elapsed time per iteration (s): 1.77 | learning rate: 1.550E-04 | global batch size: 512 | lm loss: 1.406370E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.794 | TFLOPs: 43.50 | 31: iteration 17840/ 52452 | consumed samples: 9134080 | consumed tokens: 18706595840 | elapsed time per iteration (s): 1.77 | learning rate: 1.550E-04 | global batch size: 512 | lm loss: 1.419684E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.223 | TFLOPs: 43.41 | 31: iteration 17850/ 52452 | consumed samples: 9139200 | consumed tokens: 18717081600 | elapsed time per iteration (s): 1.79 | learning rate: 1.549E-04 | global batch size: 512 | lm loss: 1.422120E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.544 | TFLOPs: 43.01 | 31: iteration 17860/ 52452 | consumed samples: 9144320 | consumed tokens: 18727567360 | elapsed time per iteration (s): 1.77 | learning rate: 1.549E-04 | global batch size: 512 | lm loss: 1.413758E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.836 | TFLOPs: 43.35 | 31: iteration 17870/ 52452 | consumed samples: 9149440 | consumed tokens: 18738053120 | elapsed time per iteration (s): 1.82 | learning rate: 1.548E-04 | global batch size: 512 | lm loss: 1.416309E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.695 | TFLOPs: 42.13 | 31: iteration 17880/ 52452 | consumed samples: 9154560 | consumed tokens: 18748538880 | elapsed time per iteration (s): 1.76 | learning rate: 1.548E-04 | global batch size: 512 | lm loss: 1.410138E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.151 | TFLOPs: 43.55 | 31: iteration 17890/ 52452 | consumed samples: 9159680 | consumed tokens: 18759024640 | elapsed time per iteration (s): 1.79 | learning rate: 1.547E-04 | global batch size: 512 | lm loss: 1.419265E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.617 | TFLOPs: 42.87 | 31: iteration 17900/ 52452 | consumed samples: 9164800 | consumed tokens: 18769510400 | elapsed time per iteration (s): 1.73 | learning rate: 1.547E-04 | global batch size: 512 | lm loss: 1.424171E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.297 | TFLOPs: 44.32 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 17900 | lm loss value: 4.162196E+00 | lm loss PPL: 6.421236E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 17900 | lm loss value: 4.097885E+00 | lm loss PPL: 6.021281E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 17910/ 52452 | consumed samples: 9169920 | consumed tokens: 18779996160 | elapsed time per iteration (s): 9.22 | learning rate: 1.546E-04 | global batch size: 512 | lm loss: 1.424205E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.508 | TFLOPs: 8.33 | 31: iteration 17920/ 52452 | consumed samples: 9175040 | consumed tokens: 18790481920 | elapsed time per iteration (s): 1.77 | learning rate: 1.546E-04 | global batch size: 512 | lm loss: 1.432646E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.973 | TFLOPs: 43.37 | 31: iteration 17930/ 52452 | consumed samples: 9180160 | consumed tokens: 18800967680 | elapsed time per iteration (s): 1.89 | learning rate: 1.545E-04 | global batch size: 512 | lm loss: 1.414462E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.826 | TFLOPs: 40.65 | 31: iteration 17940/ 52452 | consumed samples: 9185280 | consumed tokens: 18811453440 | elapsed time per iteration (s): 1.80 | learning rate: 1.545E-04 | global batch size: 512 | lm loss: 1.408618E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.576 | TFLOPs: 42.71 | 31: iteration 17950/ 52452 | consumed samples: 9190400 | consumed tokens: 18821939200 | elapsed time per iteration (s): 1.80 | learning rate: 1.545E-04 | global batch size: 512 | lm loss: 1.418890E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.072 | TFLOPs: 42.64 | 31: iteration 17960/ 52452 | consumed samples: 9195520 | consumed tokens: 18832424960 | elapsed time per iteration (s): 1.82 | learning rate: 1.544E-04 | global batch size: 512 | lm loss: 1.420378E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.031 | TFLOPs: 42.33 | 31: iteration 17970/ 52452 | consumed samples: 9200640 | consumed tokens: 18842910720 | elapsed time per iteration (s): 1.86 | learning rate: 1.544E-04 | global batch size: 512 | lm loss: 1.424062E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.738 | TFLOPs: 41.24 | 31: iteration 17980/ 52452 | consumed samples: 9205760 | consumed tokens: 18853396480 | elapsed time per iteration (s): 1.73 | learning rate: 1.543E-04 | global batch size: 512 | lm loss: 1.427363E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.751 | TFLOPs: 44.54 | 31: iteration 17990/ 52452 | consumed samples: 9210880 | consumed tokens: 18863882240 | elapsed time per iteration (s): 1.76 | learning rate: 1.543E-04 | global batch size: 512 | lm loss: 1.427856E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.991 | TFLOPs: 43.68 | 0: [2022-12-25 21:41:18,349] [INFO] [logging.py:68:log_dist] [Rank 0] step=18000, skipped=0, lr=[0.0001542136017031031, 0.0001542136017031031, 0.0001542136017031031], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 18000/ 52452 | consumed samples: 9216000 | consumed tokens: 18874368000 | elapsed time per iteration (s): 1.74 | learning rate: 1.542E-04 | global batch size: 512 | lm loss: 1.417677E+00 | grad norm: 0.187 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.203 | TFLOPs: 44.16 | 0: steps: 18000 loss: 1.4449 iter time (s): 1.765 samples/sec: 290.025 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18000 | lm loss value: 3.989530E+00 | lm loss PPL: 5.402948E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18000 | lm loss value: 3.981180E+00 | lm loss PPL: 5.358021E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18010/ 52452 | consumed samples: 9221120 | consumed tokens: 18884853760 | elapsed time per iteration (s): 9.19 | learning rate: 1.542E-04 | global batch size: 512 | lm loss: 1.409434E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.725 | TFLOPs: 8.36 | 31: iteration 18020/ 52452 | consumed samples: 9226240 | consumed tokens: 18895339520 | elapsed time per iteration (s): 1.76 | learning rate: 1.541E-04 | global batch size: 512 | lm loss: 1.415872E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.846 | TFLOPs: 43.65 | 31: iteration 18030/ 52452 | consumed samples: 9231360 | consumed tokens: 18905825280 | elapsed time per iteration (s): 1.81 | learning rate: 1.541E-04 | global batch size: 512 | lm loss: 1.414206E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.778 | TFLOPs: 42.44 | 31: iteration 18040/ 52452 | consumed samples: 9236480 | consumed tokens: 18916311040 | elapsed time per iteration (s): 1.79 | learning rate: 1.540E-04 | global batch size: 512 | lm loss: 1.405147E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.304 | TFLOPs: 42.82 | 31: iteration 18050/ 52452 | consumed samples: 9241600 | consumed tokens: 18926796800 | elapsed time per iteration (s): 1.91 | learning rate: 1.540E-04 | global batch size: 512 | lm loss: 1.410479E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.434 | TFLOPs: 40.29 | 31: iteration 18060/ 52452 | consumed samples: 9246720 | consumed tokens: 18937282560 | elapsed time per iteration (s): 1.86 | learning rate: 1.539E-04 | global batch size: 512 | lm loss: 1.413778E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.061 | TFLOPs: 41.29 | 31: iteration 18070/ 52452 | consumed samples: 9251840 | consumed tokens: 18947768320 | elapsed time per iteration (s): 1.80 | learning rate: 1.539E-04 | global batch size: 512 | lm loss: 1.408221E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.639 | TFLOPs: 42.72 | 31: iteration 18080/ 52452 | consumed samples: 9256960 | consumed tokens: 18958254080 | elapsed time per iteration (s): 1.77 | learning rate: 1.538E-04 | global batch size: 512 | lm loss: 1.402939E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.455 | TFLOPs: 43.30 | 31: iteration 18090/ 52452 | consumed samples: 9262080 | consumed tokens: 18968739840 | elapsed time per iteration (s): 1.77 | learning rate: 1.538E-04 | global batch size: 512 | lm loss: 1.410137E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.755 | TFLOPs: 43.34 | 31: iteration 18100/ 52452 | consumed samples: 9267200 | consumed tokens: 18979225600 | elapsed time per iteration (s): 1.90 | learning rate: 1.537E-04 | global batch size: 512 | lm loss: 1.418493E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.976 | TFLOPs: 40.52 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18100 | lm loss value: 4.055499E+00 | lm loss PPL: 5.771396E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18100 | lm loss value: 4.033553E+00 | lm loss PPL: 5.646117E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18110/ 52452 | consumed samples: 9272320 | consumed tokens: 18989711360 | elapsed time per iteration (s): 9.18 | learning rate: 1.537E-04 | global batch size: 512 | lm loss: 1.428169E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.768 | TFLOPs: 8.37 | 31: iteration 18120/ 52452 | consumed samples: 9277440 | consumed tokens: 19000197120 | elapsed time per iteration (s): 1.77 | learning rate: 1.536E-04 | global batch size: 512 | lm loss: 1.417270E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.669 | TFLOPs: 43.48 | 31: iteration 18130/ 52452 | consumed samples: 9282560 | consumed tokens: 19010682880 | elapsed time per iteration (s): 1.76 | learning rate: 1.536E-04 | global batch size: 512 | lm loss: 1.417421E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.309 | TFLOPs: 43.57 | 31: iteration 18140/ 52452 | consumed samples: 9287680 | consumed tokens: 19021168640 | elapsed time per iteration (s): 1.74 | learning rate: 1.535E-04 | global batch size: 512 | lm loss: 1.411143E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.813 | TFLOPs: 44.25 | 31: iteration 18150/ 52452 | consumed samples: 9292800 | consumed tokens: 19031654400 | elapsed time per iteration (s): 1.78 | learning rate: 1.535E-04 | global batch size: 512 | lm loss: 1.412159E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.141 | TFLOPs: 43.25 | 31: iteration 18160/ 52452 | consumed samples: 9297920 | consumed tokens: 19042140160 | elapsed time per iteration (s): 1.78 | learning rate: 1.535E-04 | global batch size: 512 | lm loss: 1.416230E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.162 | TFLOPs: 43.25 | 31: iteration 18170/ 52452 | consumed samples: 9303040 | consumed tokens: 19052625920 | elapsed time per iteration (s): 1.78 | learning rate: 1.534E-04 | global batch size: 512 | lm loss: 1.428273E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.424 | TFLOPs: 43.29 | 31: iteration 18180/ 52452 | consumed samples: 9308160 | consumed tokens: 19063111680 | elapsed time per iteration (s): 1.77 | learning rate: 1.534E-04 | global batch size: 512 | lm loss: 1.410768E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.013 | TFLOPs: 43.53 | 31: iteration 18190/ 52452 | consumed samples: 9313280 | consumed tokens: 19073597440 | elapsed time per iteration (s): 1.76 | learning rate: 1.533E-04 | global batch size: 512 | lm loss: 1.410868E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.573 | TFLOPs: 43.61 | 31: iteration 18200/ 52452 | consumed samples: 9318400 | consumed tokens: 19084083200 | elapsed time per iteration (s): 1.80 | learning rate: 1.533E-04 | global batch size: 512 | lm loss: 1.419399E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.032 | TFLOPs: 42.63 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18200 | lm loss value: 4.075001E+00 | lm loss PPL: 5.885055E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18200 | lm loss value: 4.073068E+00 | lm loss PPL: 5.873690E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18210/ 52452 | consumed samples: 9323520 | consumed tokens: 19094568960 | elapsed time per iteration (s): 9.18 | learning rate: 1.532E-04 | global batch size: 512 | lm loss: 1.416918E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.774 | TFLOPs: 8.37 | 31: iteration 18220/ 52452 | consumed samples: 9328640 | consumed tokens: 19105054720 | elapsed time per iteration (s): 1.79 | learning rate: 1.532E-04 | global batch size: 512 | lm loss: 1.401158E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.462 | TFLOPs: 42.85 | 31: iteration 18230/ 52452 | consumed samples: 9333760 | consumed tokens: 19115540480 | elapsed time per iteration (s): 1.80 | learning rate: 1.531E-04 | global batch size: 512 | lm loss: 1.409551E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.388 | TFLOPs: 42.69 | 31: iteration 18240/ 52452 | consumed samples: 9338880 | consumed tokens: 19126026240 | elapsed time per iteration (s): 1.76 | learning rate: 1.531E-04 | global batch size: 512 | lm loss: 1.422949E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.708 | TFLOPs: 43.78 | 31: iteration 18250/ 52452 | consumed samples: 9344000 | consumed tokens: 19136512000 | elapsed time per iteration (s): 1.75 | learning rate: 1.530E-04 | global batch size: 512 | lm loss: 1.402249E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.319 | TFLOPs: 44.03 | 31: iteration 18260/ 52452 | consumed samples: 9349120 | consumed tokens: 19146997760 | elapsed time per iteration (s): 1.75 | learning rate: 1.530E-04 | global batch size: 512 | lm loss: 1.427785E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.859 | TFLOPs: 43.81 | 31: iteration 18270/ 52452 | consumed samples: 9354240 | consumed tokens: 19157483520 | elapsed time per iteration (s): 1.77 | learning rate: 1.529E-04 | global batch size: 512 | lm loss: 1.422414E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.248 | TFLOPs: 43.41 | 31: iteration 18280/ 52452 | consumed samples: 9359360 | consumed tokens: 19167969280 | elapsed time per iteration (s): 1.78 | learning rate: 1.529E-04 | global batch size: 512 | lm loss: 1.413330E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.318 | TFLOPs: 43.12 | 31: iteration 18290/ 52452 | consumed samples: 9364480 | consumed tokens: 19178455040 | elapsed time per iteration (s): 1.75 | learning rate: 1.528E-04 | global batch size: 512 | lm loss: 1.411124E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.097 | TFLOPs: 43.99 | 31: iteration 18300/ 52452 | consumed samples: 9369600 | consumed tokens: 19188940800 | elapsed time per iteration (s): 1.80 | learning rate: 1.528E-04 | global batch size: 512 | lm loss: 1.414983E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.154 | TFLOPs: 42.65 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18300 | lm loss value: 4.148310E+00 | lm loss PPL: 6.332687E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18300 | lm loss value: 4.104855E+00 | lm loss PPL: 6.063396E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18310/ 52452 | consumed samples: 9374720 | consumed tokens: 19199426560 | elapsed time per iteration (s): 9.21 | learning rate: 1.527E-04 | global batch size: 512 | lm loss: 1.421975E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.564 | TFLOPs: 8.34 | 31: iteration 18320/ 52452 | consumed samples: 9379840 | consumed tokens: 19209912320 | elapsed time per iteration (s): 1.84 | learning rate: 1.527E-04 | global batch size: 512 | lm loss: 1.428576E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.452 | TFLOPs: 41.79 | 31: iteration 18330/ 52452 | consumed samples: 9384960 | consumed tokens: 19220398080 | elapsed time per iteration (s): 1.73 | learning rate: 1.526E-04 | global batch size: 512 | lm loss: 1.419245E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.284 | TFLOPs: 44.47 | 31: iteration 18340/ 52452 | consumed samples: 9390080 | consumed tokens: 19230883840 | elapsed time per iteration (s): 1.76 | learning rate: 1.526E-04 | global batch size: 512 | lm loss: 1.411090E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.015 | TFLOPs: 43.68 | 31: iteration 18350/ 52452 | consumed samples: 9395200 | consumed tokens: 19241369600 | elapsed time per iteration (s): 1.84 | learning rate: 1.525E-04 | global batch size: 512 | lm loss: 1.424683E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.565 | TFLOPs: 41.66 | 31: iteration 18360/ 52452 | consumed samples: 9400320 | consumed tokens: 19251855360 | elapsed time per iteration (s): 1.73 | learning rate: 1.525E-04 | global batch size: 512 | lm loss: 1.396665E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.535 | TFLOPs: 44.36 | 31: iteration 18370/ 52452 | consumed samples: 9405440 | consumed tokens: 19262341120 | elapsed time per iteration (s): 1.80 | learning rate: 1.524E-04 | global batch size: 512 | lm loss: 1.422385E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.161 | TFLOPs: 42.65 | 31: iteration 18380/ 52452 | consumed samples: 9410560 | consumed tokens: 19272826880 | elapsed time per iteration (s): 1.84 | learning rate: 1.524E-04 | global batch size: 512 | lm loss: 1.421732E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.217 | TFLOPs: 41.76 | 31: iteration 18390/ 52452 | consumed samples: 9415680 | consumed tokens: 19283312640 | elapsed time per iteration (s): 1.74 | learning rate: 1.524E-04 | global batch size: 512 | lm loss: 1.406393E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.765 | TFLOPs: 44.09 | 31: iteration 18400/ 52452 | consumed samples: 9420800 | consumed tokens: 19293798400 | elapsed time per iteration (s): 1.75 | learning rate: 1.523E-04 | global batch size: 512 | lm loss: 1.428564E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.978 | TFLOPs: 43.97 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18400 | lm loss value: 4.015741E+00 | lm loss PPL: 5.546440E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18400 | lm loss value: 3.974258E+00 | lm loss PPL: 5.321060E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18410/ 52452 | consumed samples: 9425920 | consumed tokens: 19304284160 | elapsed time per iteration (s): 9.14 | learning rate: 1.523E-04 | global batch size: 512 | lm loss: 1.410723E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.031 | TFLOPs: 8.41 | 31: iteration 18420/ 52452 | consumed samples: 9431040 | consumed tokens: 19314769920 | elapsed time per iteration (s): 1.74 | learning rate: 1.522E-04 | global batch size: 512 | lm loss: 1.397855E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.737 | TFLOPs: 44.09 | 31: iteration 18430/ 52452 | consumed samples: 9436160 | consumed tokens: 19325255680 | elapsed time per iteration (s): 1.79 | learning rate: 1.522E-04 | global batch size: 512 | lm loss: 1.416327E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.386 | TFLOPs: 42.98 | 31: iteration 18440/ 52452 | consumed samples: 9441280 | consumed tokens: 19335741440 | elapsed time per iteration (s): 1.80 | learning rate: 1.521E-04 | global batch size: 512 | lm loss: 1.408623E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.003 | TFLOPs: 42.63 | 31: iteration 18450/ 52452 | consumed samples: 9446400 | consumed tokens: 19346227200 | elapsed time per iteration (s): 1.87 | learning rate: 1.521E-04 | global batch size: 512 | lm loss: 1.404788E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.265 | TFLOPs: 41.02 | 31: iteration 18460/ 52452 | consumed samples: 9451520 | consumed tokens: 19356712960 | elapsed time per iteration (s): 1.81 | learning rate: 1.520E-04 | global batch size: 512 | lm loss: 1.421810E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.016 | TFLOPs: 42.48 | 31: iteration 18470/ 52452 | consumed samples: 9456640 | consumed tokens: 19367198720 | elapsed time per iteration (s): 1.83 | learning rate: 1.520E-04 | global batch size: 512 | lm loss: 1.408513E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.275 | TFLOPs: 41.92 | 31: iteration 18480/ 52452 | consumed samples: 9461760 | consumed tokens: 19377684480 | elapsed time per iteration (s): 1.82 | learning rate: 1.519E-04 | global batch size: 512 | lm loss: 1.397425E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.313 | TFLOPs: 42.22 | 31: iteration 18490/ 52452 | consumed samples: 9466880 | consumed tokens: 19388170240 | elapsed time per iteration (s): 1.79 | learning rate: 1.519E-04 | global batch size: 512 | lm loss: 1.404787E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.578 | TFLOPs: 42.86 | 31: iteration 18500/ 52452 | consumed samples: 9472000 | consumed tokens: 19398656000 | elapsed time per iteration (s): 1.76 | learning rate: 1.518E-04 | global batch size: 512 | lm loss: 1.415469E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.454 | TFLOPs: 43.75 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18500 | lm loss value: 4.049451E+00 | lm loss PPL: 5.736597E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18500 | lm loss value: 4.002192E+00 | lm loss PPL: 5.471799E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18510/ 52452 | consumed samples: 9477120 | consumed tokens: 19409141760 | elapsed time per iteration (s): 9.19 | learning rate: 1.518E-04 | global batch size: 512 | lm loss: 1.416233E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.723 | TFLOPs: 8.36 | 31: iteration 18520/ 52452 | consumed samples: 9482240 | consumed tokens: 19419627520 | elapsed time per iteration (s): 1.90 | learning rate: 1.517E-04 | global batch size: 512 | lm loss: 1.406735E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.899 | TFLOPs: 40.36 | 31: iteration 18530/ 52452 | consumed samples: 9487360 | consumed tokens: 19430113280 | elapsed time per iteration (s): 1.80 | learning rate: 1.517E-04 | global batch size: 512 | lm loss: 1.416976E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.973 | TFLOPs: 42.77 | 31: iteration 18540/ 52452 | consumed samples: 9492480 | consumed tokens: 19440599040 | elapsed time per iteration (s): 1.87 | learning rate: 1.516E-04 | global batch size: 512 | lm loss: 1.417076E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.331 | TFLOPs: 41.18 | 31: iteration 18550/ 52452 | consumed samples: 9497600 | consumed tokens: 19451084800 | elapsed time per iteration (s): 1.78 | learning rate: 1.516E-04 | global batch size: 512 | lm loss: 1.416861E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.989 | TFLOPs: 43.08 | 31: iteration 18560/ 52452 | consumed samples: 9502720 | consumed tokens: 19461570560 | elapsed time per iteration (s): 1.74 | learning rate: 1.515E-04 | global batch size: 512 | lm loss: 1.407507E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.827 | TFLOPs: 44.10 | 31: iteration 18570/ 52452 | consumed samples: 9507840 | consumed tokens: 19472056320 | elapsed time per iteration (s): 1.76 | learning rate: 1.515E-04 | global batch size: 512 | lm loss: 1.399957E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.387 | TFLOPs: 43.74 | 31: iteration 18580/ 52452 | consumed samples: 9512960 | consumed tokens: 19482542080 | elapsed time per iteration (s): 1.81 | learning rate: 1.514E-04 | global batch size: 512 | lm loss: 1.398724E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.490 | TFLOPs: 42.40 | 31: iteration 18590/ 52452 | consumed samples: 9518080 | consumed tokens: 19493027840 | elapsed time per iteration (s): 1.78 | learning rate: 1.514E-04 | global batch size: 512 | lm loss: 1.414270E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.078 | TFLOPs: 43.24 | 31: iteration 18600/ 52452 | consumed samples: 9523200 | consumed tokens: 19503513600 | elapsed time per iteration (s): 1.86 | learning rate: 1.513E-04 | global batch size: 512 | lm loss: 1.413642E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.985 | TFLOPs: 41.27 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18600 | lm loss value: 4.150202E+00 | lm loss PPL: 6.344680E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18600 | lm loss value: 4.091503E+00 | lm loss PPL: 5.982973E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18610/ 52452 | consumed samples: 9528320 | consumed tokens: 19513999360 | elapsed time per iteration (s): 9.22 | learning rate: 1.513E-04 | global batch size: 512 | lm loss: 1.420868E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.503 | TFLOPs: 8.33 | 31: iteration 18620/ 52452 | consumed samples: 9533440 | consumed tokens: 19524485120 | elapsed time per iteration (s): 1.81 | learning rate: 1.512E-04 | global batch size: 512 | lm loss: 1.414700E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.933 | TFLOPs: 42.47 | 31: iteration 18630/ 52452 | consumed samples: 9538560 | consumed tokens: 19534970880 | elapsed time per iteration (s): 1.75 | learning rate: 1.512E-04 | global batch size: 512 | lm loss: 1.408962E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.818 | TFLOPs: 43.80 | 31: iteration 18640/ 52452 | consumed samples: 9543680 | consumed tokens: 19545456640 | elapsed time per iteration (s): 1.85 | learning rate: 1.511E-04 | global batch size: 512 | lm loss: 1.409759E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.289 | TFLOPs: 41.62 | 31: iteration 18650/ 52452 | consumed samples: 9548800 | consumed tokens: 19555942400 | elapsed time per iteration (s): 1.80 | learning rate: 1.511E-04 | global batch size: 512 | lm loss: 1.424703E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.798 | TFLOPs: 42.60 | 31: iteration 18660/ 52452 | consumed samples: 9553920 | consumed tokens: 19566428160 | elapsed time per iteration (s): 1.80 | learning rate: 1.510E-04 | global batch size: 512 | lm loss: 1.414053E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.048 | TFLOPs: 42.78 | 31: iteration 18670/ 52452 | consumed samples: 9559040 | consumed tokens: 19576913920 | elapsed time per iteration (s): 1.72 | learning rate: 1.510E-04 | global batch size: 512 | lm loss: 1.400588E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.006 | TFLOPs: 44.58 | 31: iteration 18680/ 52452 | consumed samples: 9564160 | consumed tokens: 19587399680 | elapsed time per iteration (s): 1.72 | learning rate: 1.510E-04 | global batch size: 512 | lm loss: 1.410826E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.303 | TFLOPs: 44.62 | 31: iteration 18690/ 52452 | consumed samples: 9569280 | consumed tokens: 19597885440 | elapsed time per iteration (s): 1.85 | learning rate: 1.509E-04 | global batch size: 512 | lm loss: 1.420122E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.808 | TFLOPs: 41.55 | 31: iteration 18700/ 52452 | consumed samples: 9574400 | consumed tokens: 19608371200 | elapsed time per iteration (s): 1.75 | learning rate: 1.509E-04 | global batch size: 512 | lm loss: 1.405972E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.828 | TFLOPs: 43.95 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18700 | lm loss value: 4.000746E+00 | lm loss PPL: 5.463891E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18700 | lm loss value: 3.982209E+00 | lm loss PPL: 5.363541E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18710/ 52452 | consumed samples: 9579520 | consumed tokens: 19618856960 | elapsed time per iteration (s): 9.20 | learning rate: 1.508E-04 | global batch size: 512 | lm loss: 1.403458E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.660 | TFLOPs: 8.35 | 31: iteration 18720/ 52452 | consumed samples: 9584640 | consumed tokens: 19629342720 | elapsed time per iteration (s): 1.76 | learning rate: 1.508E-04 | global batch size: 512 | lm loss: 1.411841E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.535 | TFLOPs: 43.76 | 31: iteration 18730/ 52452 | consumed samples: 9589760 | consumed tokens: 19639828480 | elapsed time per iteration (s): 1.78 | learning rate: 1.507E-04 | global batch size: 512 | lm loss: 1.401304E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.431 | TFLOPs: 43.29 | 31: iteration 18740/ 52452 | consumed samples: 9594880 | consumed tokens: 19650314240 | elapsed time per iteration (s): 1.72 | learning rate: 1.507E-04 | global batch size: 512 | lm loss: 1.405459E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.989 | TFLOPs: 44.58 | 31: iteration 18750/ 52452 | consumed samples: 9600000 | consumed tokens: 19660800000 | elapsed time per iteration (s): 1.75 | learning rate: 1.506E-04 | global batch size: 512 | lm loss: 1.403065E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.885 | TFLOPs: 43.96 | 31: iteration 18760/ 52452 | consumed samples: 9605120 | consumed tokens: 19671285760 | elapsed time per iteration (s): 1.79 | learning rate: 1.506E-04 | global batch size: 512 | lm loss: 1.403623E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.891 | TFLOPs: 42.91 | 31: iteration 18770/ 52452 | consumed samples: 9610240 | consumed tokens: 19681771520 | elapsed time per iteration (s): 1.79 | learning rate: 1.505E-04 | global batch size: 512 | lm loss: 1.407439E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.718 | TFLOPs: 42.88 | 31: iteration 18780/ 52452 | consumed samples: 9615360 | consumed tokens: 19692257280 | elapsed time per iteration (s): 1.78 | learning rate: 1.505E-04 | global batch size: 512 | lm loss: 1.410005E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.058 | TFLOPs: 43.24 | 31: iteration 18790/ 52452 | consumed samples: 9620480 | consumed tokens: 19702743040 | elapsed time per iteration (s): 1.74 | learning rate: 1.504E-04 | global batch size: 512 | lm loss: 1.403156E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.725 | TFLOPs: 44.24 | 31: iteration 18800/ 52452 | consumed samples: 9625600 | consumed tokens: 19713228800 | elapsed time per iteration (s): 1.76 | learning rate: 1.504E-04 | global batch size: 512 | lm loss: 1.409589E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.697 | TFLOPs: 43.63 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18800 | lm loss value: 4.313922E+00 | lm loss PPL: 7.473305E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18800 | lm loss value: 4.209241E+00 | lm loss PPL: 6.730546E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18810/ 52452 | consumed samples: 9630720 | consumed tokens: 19723714560 | elapsed time per iteration (s): 9.19 | learning rate: 1.503E-04 | global batch size: 512 | lm loss: 1.415028E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.739 | TFLOPs: 8.37 | 31: iteration 18820/ 52452 | consumed samples: 9635840 | consumed tokens: 19734200320 | elapsed time per iteration (s): 1.82 | learning rate: 1.503E-04 | global batch size: 512 | lm loss: 1.419963E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.014 | TFLOPs: 42.33 | 31: iteration 18830/ 52452 | consumed samples: 9640960 | consumed tokens: 19744686080 | elapsed time per iteration (s): 1.81 | learning rate: 1.502E-04 | global batch size: 512 | lm loss: 1.409405E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.760 | TFLOPs: 42.44 | 31: iteration 18840/ 52452 | consumed samples: 9646080 | consumed tokens: 19755171840 | elapsed time per iteration (s): 1.80 | learning rate: 1.502E-04 | global batch size: 512 | lm loss: 1.400327E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.382 | TFLOPs: 42.68 | 31: iteration 18850/ 52452 | consumed samples: 9651200 | consumed tokens: 19765657600 | elapsed time per iteration (s): 1.81 | learning rate: 1.501E-04 | global batch size: 512 | lm loss: 1.398017E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.355 | TFLOPs: 42.53 | 31: iteration 18860/ 52452 | consumed samples: 9656320 | consumed tokens: 19776143360 | elapsed time per iteration (s): 1.77 | learning rate: 1.501E-04 | global batch size: 512 | lm loss: 1.407055E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.782 | TFLOPs: 43.49 | 31: iteration 18870/ 52452 | consumed samples: 9661440 | consumed tokens: 19786629120 | elapsed time per iteration (s): 1.86 | learning rate: 1.500E-04 | global batch size: 512 | lm loss: 1.417142E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.886 | TFLOPs: 41.26 | 31: iteration 18880/ 52452 | consumed samples: 9666560 | consumed tokens: 19797114880 | elapsed time per iteration (s): 1.81 | learning rate: 1.500E-04 | global batch size: 512 | lm loss: 1.402876E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.895 | TFLOPs: 42.46 | 31: iteration 18890/ 52452 | consumed samples: 9671680 | consumed tokens: 19807600640 | elapsed time per iteration (s): 1.77 | learning rate: 1.499E-04 | global batch size: 512 | lm loss: 1.414821E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.977 | TFLOPs: 43.52 | 31: iteration 18900/ 52452 | consumed samples: 9676800 | consumed tokens: 19818086400 | elapsed time per iteration (s): 1.79 | learning rate: 1.499E-04 | global batch size: 512 | lm loss: 1.405888E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.162 | TFLOPs: 42.95 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 18900 | lm loss value: 4.150780E+00 | lm loss PPL: 6.348348E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 18900 | lm loss value: 4.075956E+00 | lm loss PPL: 5.890676E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 18910/ 52452 | consumed samples: 9681920 | consumed tokens: 19828572160 | elapsed time per iteration (s): 9.27 | learning rate: 1.498E-04 | global batch size: 512 | lm loss: 1.393220E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.254 | TFLOPs: 8.29 | 31: iteration 18920/ 52452 | consumed samples: 9687040 | consumed tokens: 19839057920 | elapsed time per iteration (s): 1.78 | learning rate: 1.498E-04 | global batch size: 512 | lm loss: 1.413810E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.954 | TFLOPs: 43.22 | 31: iteration 18930/ 52452 | consumed samples: 9692160 | consumed tokens: 19849543680 | elapsed time per iteration (s): 1.83 | learning rate: 1.497E-04 | global batch size: 512 | lm loss: 1.399255E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.854 | TFLOPs: 42.00 | 31: iteration 18940/ 52452 | consumed samples: 9697280 | consumed tokens: 19860029440 | elapsed time per iteration (s): 1.77 | learning rate: 1.497E-04 | global batch size: 512 | lm loss: 1.399604E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.554 | TFLOPs: 43.31 | 31: iteration 18950/ 52452 | consumed samples: 9702400 | consumed tokens: 19870515200 | elapsed time per iteration (s): 1.79 | learning rate: 1.496E-04 | global batch size: 512 | lm loss: 1.414278E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.799 | TFLOPs: 43.05 | 31: iteration 18960/ 52452 | consumed samples: 9707520 | consumed tokens: 19881000960 | elapsed time per iteration (s): 1.73 | learning rate: 1.496E-04 | global batch size: 512 | lm loss: 1.398516E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.949 | TFLOPs: 44.42 | 31: iteration 18970/ 52452 | consumed samples: 9712640 | consumed tokens: 19891486720 | elapsed time per iteration (s): 1.74 | learning rate: 1.495E-04 | global batch size: 512 | lm loss: 1.398617E+00 | grad norm: 0.188 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.752 | TFLOPs: 44.09 | 31: iteration 18980/ 52452 | consumed samples: 9717760 | consumed tokens: 19901972480 | elapsed time per iteration (s): 1.78 | learning rate: 1.495E-04 | global batch size: 512 | lm loss: 1.403434E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.038 | TFLOPs: 43.23 | 31: iteration 18990/ 52452 | consumed samples: 9722880 | consumed tokens: 19912458240 | elapsed time per iteration (s): 1.79 | learning rate: 1.494E-04 | global batch size: 512 | lm loss: 1.397700E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.794 | TFLOPs: 43.05 | 31: iteration 19000/ 52452 | consumed samples: 9728000 | consumed tokens: 19922944000 | elapsed time per iteration (s): 1.81 | learning rate: 1.494E-04 | global batch size: 512 | lm loss: 1.407915E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.987 | TFLOPs: 42.47 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19000 | lm loss value: 4.231021E+00 | lm loss PPL: 6.878742E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19000 | lm loss value: 4.142134E+00 | lm loss PPL: 6.293700E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19010/ 52452 | consumed samples: 9733120 | consumed tokens: 19933429760 | elapsed time per iteration (s): 9.10 | learning rate: 1.493E-04 | global batch size: 512 | lm loss: 1.409234E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.272 | TFLOPs: 8.45 | 31: iteration 19020/ 52452 | consumed samples: 9738240 | consumed tokens: 19943915520 | elapsed time per iteration (s): 1.79 | learning rate: 1.493E-04 | global batch size: 512 | lm loss: 1.415259E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.026 | TFLOPs: 42.93 | 31: iteration 19030/ 52452 | consumed samples: 9743360 | consumed tokens: 19954401280 | elapsed time per iteration (s): 1.77 | learning rate: 1.492E-04 | global batch size: 512 | lm loss: 1.410855E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.687 | TFLOPs: 43.48 | 31: iteration 19040/ 52452 | consumed samples: 9748480 | consumed tokens: 19964887040 | elapsed time per iteration (s): 1.75 | learning rate: 1.492E-04 | global batch size: 512 | lm loss: 1.391684E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.243 | TFLOPs: 43.86 | 31: iteration 19050/ 52452 | consumed samples: 9753600 | consumed tokens: 19975372800 | elapsed time per iteration (s): 1.86 | learning rate: 1.491E-04 | global batch size: 512 | lm loss: 1.408251E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.648 | TFLOPs: 41.37 | 31: iteration 19060/ 52452 | consumed samples: 9758720 | consumed tokens: 19985858560 | elapsed time per iteration (s): 1.78 | learning rate: 1.491E-04 | global batch size: 512 | lm loss: 1.409522E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.066 | TFLOPs: 43.09 | 31: iteration 19070/ 52452 | consumed samples: 9763840 | consumed tokens: 19996344320 | elapsed time per iteration (s): 1.78 | learning rate: 1.490E-04 | global batch size: 512 | lm loss: 1.398694E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.255 | TFLOPs: 43.27 | 31: iteration 19080/ 52452 | consumed samples: 9768960 | consumed tokens: 20006830080 | elapsed time per iteration (s): 1.76 | learning rate: 1.490E-04 | global batch size: 512 | lm loss: 1.404517E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.660 | TFLOPs: 43.63 | 31: iteration 19090/ 52452 | consumed samples: 9774080 | consumed tokens: 20017315840 | elapsed time per iteration (s): 1.83 | learning rate: 1.490E-04 | global batch size: 512 | lm loss: 1.405094E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.857 | TFLOPs: 42.00 | 31: iteration 19100/ 52452 | consumed samples: 9779200 | consumed tokens: 20027801600 | elapsed time per iteration (s): 1.76 | learning rate: 1.489E-04 | global batch size: 512 | lm loss: 1.413547E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.146 | TFLOPs: 43.70 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19100 | lm loss value: 4.204516E+00 | lm loss PPL: 6.698820E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19100 | lm loss value: 4.127206E+00 | lm loss PPL: 6.200443E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19110/ 52452 | consumed samples: 9784320 | consumed tokens: 20038287360 | elapsed time per iteration (s): 9.14 | learning rate: 1.489E-04 | global batch size: 512 | lm loss: 1.408219E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.012 | TFLOPs: 8.41 | 31: iteration 19120/ 52452 | consumed samples: 9789440 | consumed tokens: 20048773120 | elapsed time per iteration (s): 1.81 | learning rate: 1.488E-04 | global batch size: 512 | lm loss: 1.401171E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.323 | TFLOPs: 42.38 | 31: iteration 19130/ 52452 | consumed samples: 9794560 | consumed tokens: 20059258880 | elapsed time per iteration (s): 1.87 | learning rate: 1.488E-04 | global batch size: 512 | lm loss: 1.403177E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.345 | TFLOPs: 41.18 | 31: iteration 19140/ 52452 | consumed samples: 9799680 | consumed tokens: 20069744640 | elapsed time per iteration (s): 1.78 | learning rate: 1.487E-04 | global batch size: 512 | lm loss: 1.393768E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.028 | TFLOPs: 43.23 | 31: iteration 19150/ 52452 | consumed samples: 9804800 | consumed tokens: 20080230400 | elapsed time per iteration (s): 1.76 | learning rate: 1.487E-04 | global batch size: 512 | lm loss: 1.402714E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.585 | TFLOPs: 43.77 | 31: iteration 19160/ 52452 | consumed samples: 9809920 | consumed tokens: 20090716160 | elapsed time per iteration (s): 1.75 | learning rate: 1.486E-04 | global batch size: 512 | lm loss: 1.399759E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.993 | TFLOPs: 43.83 | 31: iteration 19170/ 52452 | consumed samples: 9815040 | consumed tokens: 20101201920 | elapsed time per iteration (s): 1.79 | learning rate: 1.486E-04 | global batch size: 512 | lm loss: 1.417396E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.324 | TFLOPs: 42.83 | 31: iteration 19180/ 52452 | consumed samples: 9820160 | consumed tokens: 20111687680 | elapsed time per iteration (s): 1.83 | learning rate: 1.485E-04 | global batch size: 512 | lm loss: 1.397487E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.201 | TFLOPs: 41.91 | 31: iteration 19190/ 52452 | consumed samples: 9825280 | consumed tokens: 20122173440 | elapsed time per iteration (s): 1.84 | learning rate: 1.485E-04 | global batch size: 512 | lm loss: 1.391309E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.410 | TFLOPs: 41.79 | 31: iteration 19200/ 52452 | consumed samples: 9830400 | consumed tokens: 20132659200 | elapsed time per iteration (s): 1.84 | learning rate: 1.484E-04 | global batch size: 512 | lm loss: 1.402959E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.029 | TFLOPs: 41.73 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19200 | lm loss value: 4.135253E+00 | lm loss PPL: 6.250543E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19200 | lm loss value: 4.081782E+00 | lm loss PPL: 5.925098E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19210/ 52452 | consumed samples: 9835520 | consumed tokens: 20143144960 | elapsed time per iteration (s): 9.17 | learning rate: 1.484E-04 | global batch size: 512 | lm loss: 1.405753E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.857 | TFLOPs: 8.38 | 31: iteration 19220/ 52452 | consumed samples: 9840640 | consumed tokens: 20153630720 | elapsed time per iteration (s): 1.94 | learning rate: 1.483E-04 | global batch size: 512 | lm loss: 1.398172E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 263.624 | TFLOPs: 39.57 | 31: iteration 19230/ 52452 | consumed samples: 9845760 | consumed tokens: 20164116480 | elapsed time per iteration (s): 1.76 | learning rate: 1.483E-04 | global batch size: 512 | lm loss: 1.396872E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.111 | TFLOPs: 43.54 | 31: iteration 19240/ 52452 | consumed samples: 9850880 | consumed tokens: 20174602240 | elapsed time per iteration (s): 1.80 | learning rate: 1.482E-04 | global batch size: 512 | lm loss: 1.400331E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.101 | TFLOPs: 42.79 | 31: iteration 19250/ 52452 | consumed samples: 9856000 | consumed tokens: 20185088000 | elapsed time per iteration (s): 1.76 | learning rate: 1.482E-04 | global batch size: 512 | lm loss: 1.408978E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.906 | TFLOPs: 43.66 | 31: iteration 19260/ 52452 | consumed samples: 9861120 | consumed tokens: 20195573760 | elapsed time per iteration (s): 1.77 | learning rate: 1.481E-04 | global batch size: 512 | lm loss: 1.394608E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.397 | TFLOPs: 43.44 | 31: iteration 19270/ 52452 | consumed samples: 9866240 | consumed tokens: 20206059520 | elapsed time per iteration (s): 1.79 | learning rate: 1.481E-04 | global batch size: 512 | lm loss: 1.397036E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.071 | TFLOPs: 42.94 | 31: iteration 19280/ 52452 | consumed samples: 9871360 | consumed tokens: 20216545280 | elapsed time per iteration (s): 1.81 | learning rate: 1.480E-04 | global batch size: 512 | lm loss: 1.394921E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.005 | TFLOPs: 42.48 | 31: iteration 19290/ 52452 | consumed samples: 9876480 | consumed tokens: 20227031040 | elapsed time per iteration (s): 1.75 | learning rate: 1.480E-04 | global batch size: 512 | lm loss: 1.404002E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.420 | TFLOPs: 43.89 | 31: iteration 19300/ 52452 | consumed samples: 9881600 | consumed tokens: 20237516800 | elapsed time per iteration (s): 1.75 | learning rate: 1.479E-04 | global batch size: 512 | lm loss: 1.399320E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.241 | TFLOPs: 43.86 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19300 | lm loss value: 4.079319E+00 | lm loss PPL: 5.910518E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19300 | lm loss value: 4.066116E+00 | lm loss PPL: 5.832996E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19310/ 52452 | consumed samples: 9886720 | consumed tokens: 20248002560 | elapsed time per iteration (s): 9.17 | learning rate: 1.479E-04 | global batch size: 512 | lm loss: 1.405456E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.821 | TFLOPs: 8.38 | 31: iteration 19320/ 52452 | consumed samples: 9891840 | consumed tokens: 20258488320 | elapsed time per iteration (s): 1.86 | learning rate: 1.478E-04 | global batch size: 512 | lm loss: 1.393436E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.104 | TFLOPs: 41.29 | 31: iteration 19330/ 52452 | consumed samples: 9896960 | consumed tokens: 20268974080 | elapsed time per iteration (s): 1.81 | learning rate: 1.478E-04 | global batch size: 512 | lm loss: 1.408372E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.550 | TFLOPs: 42.41 | 31: iteration 19340/ 52452 | consumed samples: 9902080 | consumed tokens: 20279459840 | elapsed time per iteration (s): 1.82 | learning rate: 1.477E-04 | global batch size: 512 | lm loss: 1.387544E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.721 | TFLOPs: 42.13 | 31: iteration 19350/ 52452 | consumed samples: 9907200 | consumed tokens: 20289945600 | elapsed time per iteration (s): 1.75 | learning rate: 1.477E-04 | global batch size: 512 | lm loss: 1.411142E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.840 | TFLOPs: 43.80 | 31: iteration 19360/ 52452 | consumed samples: 9912320 | consumed tokens: 20300431360 | elapsed time per iteration (s): 1.78 | learning rate: 1.476E-04 | global batch size: 512 | lm loss: 1.408193E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.438 | TFLOPs: 43.29 | 31: iteration 19370/ 52452 | consumed samples: 9917440 | consumed tokens: 20310917120 | elapsed time per iteration (s): 1.77 | learning rate: 1.476E-04 | global batch size: 512 | lm loss: 1.402031E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.685 | TFLOPs: 43.33 | 31: iteration 19380/ 52452 | consumed samples: 9922560 | consumed tokens: 20321402880 | elapsed time per iteration (s): 1.83 | learning rate: 1.475E-04 | global batch size: 512 | lm loss: 1.406408E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.366 | TFLOPs: 41.93 | 31: iteration 19390/ 52452 | consumed samples: 9927680 | consumed tokens: 20331888640 | elapsed time per iteration (s): 1.74 | learning rate: 1.475E-04 | global batch size: 512 | lm loss: 1.404751E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.703 | TFLOPs: 44.08 | 31: iteration 19400/ 52452 | consumed samples: 9932800 | consumed tokens: 20342374400 | elapsed time per iteration (s): 1.79 | learning rate: 1.474E-04 | global batch size: 512 | lm loss: 1.398831E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.755 | TFLOPs: 42.89 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19400 | lm loss value: 4.003684E+00 | lm loss PPL: 5.479964E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19400 | lm loss value: 3.989166E+00 | lm loss PPL: 5.400984E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19410/ 52452 | consumed samples: 9937920 | consumed tokens: 20352860160 | elapsed time per iteration (s): 9.11 | learning rate: 1.474E-04 | global batch size: 512 | lm loss: 1.390997E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.173 | TFLOPs: 8.43 | 31: iteration 19420/ 52452 | consumed samples: 9943040 | consumed tokens: 20363345920 | elapsed time per iteration (s): 1.75 | learning rate: 1.473E-04 | global batch size: 512 | lm loss: 1.398737E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.417 | TFLOPs: 43.89 | 31: iteration 19430/ 52452 | consumed samples: 9948160 | consumed tokens: 20373831680 | elapsed time per iteration (s): 1.74 | learning rate: 1.473E-04 | global batch size: 512 | lm loss: 1.391470E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.533 | TFLOPs: 44.06 | 31: iteration 19440/ 52452 | consumed samples: 9953280 | consumed tokens: 20384317440 | elapsed time per iteration (s): 1.75 | learning rate: 1.472E-04 | global batch size: 512 | lm loss: 1.404756E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.168 | TFLOPs: 43.85 | 31: iteration 19450/ 52452 | consumed samples: 9958400 | consumed tokens: 20394803200 | elapsed time per iteration (s): 1.82 | learning rate: 1.472E-04 | global batch size: 512 | lm loss: 1.397912E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.530 | TFLOPs: 42.26 | 31: iteration 19460/ 52452 | consumed samples: 9963520 | consumed tokens: 20405288960 | elapsed time per iteration (s): 1.75 | learning rate: 1.471E-04 | global batch size: 512 | lm loss: 1.390662E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.668 | TFLOPs: 43.93 | 31: iteration 19470/ 52452 | consumed samples: 9968640 | consumed tokens: 20415774720 | elapsed time per iteration (s): 1.75 | learning rate: 1.471E-04 | global batch size: 512 | lm loss: 1.387819E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.413 | TFLOPs: 43.89 | 31: iteration 19480/ 52452 | consumed samples: 9973760 | consumed tokens: 20426260480 | elapsed time per iteration (s): 1.82 | learning rate: 1.470E-04 | global batch size: 512 | lm loss: 1.394225E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.571 | TFLOPs: 42.26 | 31: iteration 19490/ 52452 | consumed samples: 9978880 | consumed tokens: 20436746240 | elapsed time per iteration (s): 1.79 | learning rate: 1.470E-04 | global batch size: 512 | lm loss: 1.407776E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.738 | TFLOPs: 42.89 | 31: iteration 19500/ 52452 | consumed samples: 9984000 | consumed tokens: 20447232000 | elapsed time per iteration (s): 1.78 | learning rate: 1.469E-04 | global batch size: 512 | lm loss: 1.394670E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.998 | TFLOPs: 43.23 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19500 | lm loss value: 4.087167E+00 | lm loss PPL: 5.957088E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19500 | lm loss value: 4.041183E+00 | lm loss PPL: 5.689363E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19510/ 52452 | consumed samples: 9989120 | consumed tokens: 20457717760 | elapsed time per iteration (s): 9.21 | learning rate: 1.469E-04 | global batch size: 512 | lm loss: 1.403921E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.577 | TFLOPs: 8.34 | 31: iteration 19520/ 52452 | consumed samples: 9994240 | consumed tokens: 20468203520 | elapsed time per iteration (s): 1.75 | learning rate: 1.468E-04 | global batch size: 512 | lm loss: 1.397186E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.759 | TFLOPs: 43.79 | 31: iteration 19530/ 52452 | consumed samples: 9999360 | consumed tokens: 20478689280 | elapsed time per iteration (s): 1.81 | learning rate: 1.468E-04 | global batch size: 512 | lm loss: 1.398728E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.395 | TFLOPs: 42.39 | 31: iteration 19540/ 52452 | consumed samples: 10004480 | consumed tokens: 20489175040 | elapsed time per iteration (s): 1.84 | learning rate: 1.467E-04 | global batch size: 512 | lm loss: 1.408545E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.139 | TFLOPs: 41.75 | 31: iteration 19550/ 52452 | consumed samples: 10009600 | consumed tokens: 20499660800 | elapsed time per iteration (s): 1.79 | learning rate: 1.467E-04 | global batch size: 512 | lm loss: 1.387568E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.663 | TFLOPs: 42.88 | 31: iteration 19560/ 52452 | consumed samples: 10014720 | consumed tokens: 20510146560 | elapsed time per iteration (s): 1.79 | learning rate: 1.466E-04 | global batch size: 512 | lm loss: 1.400314E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.561 | TFLOPs: 43.01 | 31: iteration 19570/ 52452 | consumed samples: 10019840 | consumed tokens: 20520632320 | elapsed time per iteration (s): 1.82 | learning rate: 1.466E-04 | global batch size: 512 | lm loss: 1.405464E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.922 | TFLOPs: 42.16 | 31: iteration 19580/ 52452 | consumed samples: 10024960 | consumed tokens: 20531118080 | elapsed time per iteration (s): 1.80 | learning rate: 1.465E-04 | global batch size: 512 | lm loss: 1.393844E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.450 | TFLOPs: 42.69 | 31: iteration 19590/ 52452 | consumed samples: 10030080 | consumed tokens: 20541603840 | elapsed time per iteration (s): 1.77 | learning rate: 1.465E-04 | global batch size: 512 | lm loss: 1.388936E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.700 | TFLOPs: 43.48 | 31: iteration 19600/ 52452 | consumed samples: 10035200 | consumed tokens: 20552089600 | elapsed time per iteration (s): 1.77 | learning rate: 1.464E-04 | global batch size: 512 | lm loss: 1.393599E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.069 | TFLOPs: 43.54 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19600 | lm loss value: 4.210323E+00 | lm loss PPL: 6.737829E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19600 | lm loss value: 4.118570E+00 | lm loss PPL: 6.147130E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19610/ 52452 | consumed samples: 10040320 | consumed tokens: 20562575360 | elapsed time per iteration (s): 9.25 | learning rate: 1.464E-04 | global batch size: 512 | lm loss: 1.407196E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.340 | TFLOPs: 8.31 | 31: iteration 19620/ 52452 | consumed samples: 10045440 | consumed tokens: 20573061120 | elapsed time per iteration (s): 1.74 | learning rate: 1.463E-04 | global batch size: 512 | lm loss: 1.388493E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.457 | TFLOPs: 44.20 | 31: iteration 19630/ 52452 | consumed samples: 10050560 | consumed tokens: 20583546880 | elapsed time per iteration (s): 1.81 | learning rate: 1.463E-04 | global batch size: 512 | lm loss: 1.396393E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.757 | TFLOPs: 42.44 | 31: iteration 19640/ 52452 | consumed samples: 10055680 | consumed tokens: 20594032640 | elapsed time per iteration (s): 1.79 | learning rate: 1.462E-04 | global batch size: 512 | lm loss: 1.393454E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.432 | TFLOPs: 42.99 | 31: iteration 19650/ 52452 | consumed samples: 10060800 | consumed tokens: 20604518400 | elapsed time per iteration (s): 1.74 | learning rate: 1.462E-04 | global batch size: 512 | lm loss: 1.407372E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.001 | TFLOPs: 44.13 | 31: iteration 19660/ 52452 | consumed samples: 10065920 | consumed tokens: 20615004160 | elapsed time per iteration (s): 1.78 | learning rate: 1.461E-04 | global batch size: 512 | lm loss: 1.402942E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.487 | TFLOPs: 43.15 | 31: iteration 19670/ 52452 | consumed samples: 10071040 | consumed tokens: 20625489920 | elapsed time per iteration (s): 1.79 | learning rate: 1.461E-04 | global batch size: 512 | lm loss: 1.401680E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.506 | TFLOPs: 42.85 | 31: iteration 19680/ 52452 | consumed samples: 10076160 | consumed tokens: 20635975680 | elapsed time per iteration (s): 1.81 | learning rate: 1.460E-04 | global batch size: 512 | lm loss: 1.402817E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.297 | TFLOPs: 42.52 | 31: iteration 19690/ 52452 | consumed samples: 10081280 | consumed tokens: 20646461440 | elapsed time per iteration (s): 1.79 | learning rate: 1.460E-04 | global batch size: 512 | lm loss: 1.401990E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.536 | TFLOPs: 42.86 | 31: iteration 19700/ 52452 | consumed samples: 10086400 | consumed tokens: 20656947200 | elapsed time per iteration (s): 1.81 | learning rate: 1.459E-04 | global batch size: 512 | lm loss: 1.403618E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.525 | TFLOPs: 42.56 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19700 | lm loss value: 4.091269E+00 | lm loss PPL: 5.981575E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19700 | lm loss value: 4.042895E+00 | lm loss PPL: 5.699111E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19710/ 52452 | consumed samples: 10091520 | consumed tokens: 20667432960 | elapsed time per iteration (s): 9.15 | learning rate: 1.459E-04 | global batch size: 512 | lm loss: 1.384518E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.983 | TFLOPs: 8.40 | 31: iteration 19720/ 52452 | consumed samples: 10096640 | consumed tokens: 20677918720 | elapsed time per iteration (s): 1.82 | learning rate: 1.458E-04 | global batch size: 512 | lm loss: 1.387039E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.152 | TFLOPs: 42.20 | 31: iteration 19730/ 52452 | consumed samples: 10101760 | consumed tokens: 20688404480 | elapsed time per iteration (s): 1.83 | learning rate: 1.458E-04 | global batch size: 512 | lm loss: 1.390072E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.688 | TFLOPs: 41.98 | 31: iteration 19740/ 52452 | consumed samples: 10106880 | consumed tokens: 20698890240 | elapsed time per iteration (s): 1.74 | learning rate: 1.457E-04 | global batch size: 512 | lm loss: 1.385749E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.975 | TFLOPs: 44.27 | 31: iteration 19750/ 52452 | consumed samples: 10112000 | consumed tokens: 20709376000 | elapsed time per iteration (s): 1.87 | learning rate: 1.457E-04 | global batch size: 512 | lm loss: 1.407262E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.854 | TFLOPs: 41.10 | 31: iteration 19760/ 52452 | consumed samples: 10117120 | consumed tokens: 20719861760 | elapsed time per iteration (s): 1.89 | learning rate: 1.456E-04 | global batch size: 512 | lm loss: 1.399064E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.342 | TFLOPs: 40.58 | 31: iteration 19770/ 52452 | consumed samples: 10122240 | consumed tokens: 20730347520 | elapsed time per iteration (s): 1.84 | learning rate: 1.456E-04 | global batch size: 512 | lm loss: 1.386925E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.332 | TFLOPs: 41.78 | 31: iteration 19780/ 52452 | consumed samples: 10127360 | consumed tokens: 20740833280 | elapsed time per iteration (s): 1.77 | learning rate: 1.455E-04 | global batch size: 512 | lm loss: 1.401532E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.931 | TFLOPs: 43.52 | 31: iteration 19790/ 52452 | consumed samples: 10132480 | consumed tokens: 20751319040 | elapsed time per iteration (s): 1.81 | learning rate: 1.455E-04 | global batch size: 512 | lm loss: 1.401493E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.195 | TFLOPs: 42.51 | 31: iteration 19800/ 52452 | consumed samples: 10137600 | consumed tokens: 20761804800 | elapsed time per iteration (s): 1.78 | learning rate: 1.454E-04 | global batch size: 512 | lm loss: 1.387740E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.064 | TFLOPs: 43.24 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19800 | lm loss value: 4.214413E+00 | lm loss PPL: 6.765442E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19800 | lm loss value: 4.131591E+00 | lm loss PPL: 6.227692E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19810/ 52452 | consumed samples: 10142720 | consumed tokens: 20772290560 | elapsed time per iteration (s): 9.14 | learning rate: 1.454E-04 | global batch size: 512 | lm loss: 1.391393E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.005 | TFLOPs: 8.41 | 31: iteration 19820/ 52452 | consumed samples: 10147840 | consumed tokens: 20782776320 | elapsed time per iteration (s): 1.86 | learning rate: 1.453E-04 | global batch size: 512 | lm loss: 1.403650E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.988 | TFLOPs: 41.27 | 31: iteration 19830/ 52452 | consumed samples: 10152960 | consumed tokens: 20793262080 | elapsed time per iteration (s): 1.82 | learning rate: 1.453E-04 | global batch size: 512 | lm loss: 1.393131E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.843 | TFLOPs: 42.15 | 31: iteration 19840/ 52452 | consumed samples: 10158080 | consumed tokens: 20803747840 | elapsed time per iteration (s): 1.80 | learning rate: 1.452E-04 | global batch size: 512 | lm loss: 1.384704E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.921 | TFLOPs: 42.61 | 31: iteration 19850/ 52452 | consumed samples: 10163200 | consumed tokens: 20814233600 | elapsed time per iteration (s): 1.82 | learning rate: 1.452E-04 | global batch size: 512 | lm loss: 1.396256E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.655 | TFLOPs: 42.27 | 31: iteration 19860/ 52452 | consumed samples: 10168320 | consumed tokens: 20824719360 | elapsed time per iteration (s): 1.81 | learning rate: 1.451E-04 | global batch size: 512 | lm loss: 1.406127E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.524 | TFLOPs: 42.56 | 31: iteration 19870/ 52452 | consumed samples: 10173440 | consumed tokens: 20835205120 | elapsed time per iteration (s): 1.76 | learning rate: 1.451E-04 | global batch size: 512 | lm loss: 1.386056E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.135 | TFLOPs: 43.70 | 31: iteration 19880/ 52452 | consumed samples: 10178560 | consumed tokens: 20845690880 | elapsed time per iteration (s): 1.92 | learning rate: 1.450E-04 | global batch size: 512 | lm loss: 1.385019E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.249 | TFLOPs: 40.11 | 31: iteration 19890/ 52452 | consumed samples: 10183680 | consumed tokens: 20856176640 | elapsed time per iteration (s): 1.80 | learning rate: 1.450E-04 | global batch size: 512 | lm loss: 1.402772E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.396 | TFLOPs: 42.69 | 31: iteration 19900/ 52452 | consumed samples: 10188800 | consumed tokens: 20866662400 | elapsed time per iteration (s): 1.78 | learning rate: 1.449E-04 | global batch size: 512 | lm loss: 1.395400E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.321 | TFLOPs: 43.13 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 19900 | lm loss value: 4.226639E+00 | lm loss PPL: 6.848668E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 19900 | lm loss value: 4.132376E+00 | lm loss PPL: 6.232585E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 19910/ 52452 | consumed samples: 10193920 | consumed tokens: 20877148160 | elapsed time per iteration (s): 9.19 | learning rate: 1.449E-04 | global batch size: 512 | lm loss: 1.402729E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.740 | TFLOPs: 8.37 | 31: iteration 19920/ 52452 | consumed samples: 10199040 | consumed tokens: 20887633920 | elapsed time per iteration (s): 1.83 | learning rate: 1.448E-04 | global batch size: 512 | lm loss: 1.404543E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.501 | TFLOPs: 42.10 | 31: iteration 19930/ 52452 | consumed samples: 10204160 | consumed tokens: 20898119680 | elapsed time per iteration (s): 1.78 | learning rate: 1.448E-04 | global batch size: 512 | lm loss: 1.399297E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.376 | TFLOPs: 43.13 | 31: iteration 19940/ 52452 | consumed samples: 10209280 | consumed tokens: 20908605440 | elapsed time per iteration (s): 1.85 | learning rate: 1.447E-04 | global batch size: 512 | lm loss: 1.381390E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.104 | TFLOPs: 41.59 | 31: iteration 19950/ 52452 | consumed samples: 10214400 | consumed tokens: 20919091200 | elapsed time per iteration (s): 1.82 | learning rate: 1.447E-04 | global batch size: 512 | lm loss: 1.380389E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.813 | TFLOPs: 42.15 | 31: iteration 19960/ 52452 | consumed samples: 10219520 | consumed tokens: 20929576960 | elapsed time per iteration (s): 1.83 | learning rate: 1.446E-04 | global batch size: 512 | lm loss: 1.396169E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.306 | TFLOPs: 41.92 | 31: iteration 19970/ 52452 | consumed samples: 10224640 | consumed tokens: 20940062720 | elapsed time per iteration (s): 1.82 | learning rate: 1.446E-04 | global batch size: 512 | lm loss: 1.396078E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.833 | TFLOPs: 42.15 | 31: iteration 19980/ 52452 | consumed samples: 10229760 | consumed tokens: 20950548480 | elapsed time per iteration (s): 1.84 | learning rate: 1.445E-04 | global batch size: 512 | lm loss: 1.404868E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.923 | TFLOPs: 41.71 | 31: iteration 19990/ 52452 | consumed samples: 10234880 | consumed tokens: 20961034240 | elapsed time per iteration (s): 1.77 | learning rate: 1.445E-04 | global batch size: 512 | lm loss: 1.388044E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.037 | TFLOPs: 43.53 | 0: [2022-12-25 23:05:42,256] [INFO] [logging.py:68:log_dist] [Rank 0] step=20000, skipped=0, lr=[0.00014442822239555864, 0.00014442822239555864, 0.00014442822239555864], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 20000/ 52452 | consumed samples: 10240000 | consumed tokens: 20971520000 | elapsed time per iteration (s): 1.76 | learning rate: 1.444E-04 | global batch size: 512 | lm loss: 1.406541E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.195 | TFLOPs: 43.56 | 0: steps: 20000 loss: 1.3819 iter time (s): 1.792 samples/sec: 285.637 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20000 | lm loss value: 4.109913E+00 | lm loss PPL: 6.094144E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20000 | lm loss value: 4.053877E+00 | lm loss PPL: 5.762044E+01 | 31: ------------------------------------------------------------------------------------------------------ 0: saving checkpoint at iteration 20000 to checkpoints_2b855brealtasky 0: [2022-12-25 23:06:56,066] [INFO] [logging.py:68:log_dist] [Rank 0] [Torch] Checkpoint global_step20000 is begin to save! 0: [2022-12-25 23:06:56,076] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_01-model_00-model_states.pt... 0: [2022-12-25 23:06:56,438] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_01-model_00-model_states.pt. 0: [2022-12-25 23:06:56,438] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_03-model_00-model_states.pt... 0: [2022-12-25 23:06:56,619] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_03-model_00-model_states.pt. 0: [2022-12-25 23:06:56,619] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_04-model_00-model_states.pt... 0: [2022-12-25 23:06:56,802] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_04-model_00-model_states.pt. 0: [2022-12-25 23:06:56,803] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_05-model_00-model_states.pt... 0: [2022-12-25 23:06:56,976] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_05-model_00-model_states.pt. 0: [2022-12-25 23:06:56,976] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_06-model_00-model_states.pt... 0: [2022-12-25 23:06:57,150] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_06-model_00-model_states.pt. 0: [2022-12-25 23:06:57,150] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_07-model_00-model_states.pt... 0: [2022-12-25 23:06:57,330] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_07-model_00-model_states.pt. 0: [2022-12-25 23:06:57,330] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_08-model_00-model_states.pt... 0: [2022-12-25 23:06:57,505] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_08-model_00-model_states.pt. 0: [2022-12-25 23:06:57,505] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_09-model_00-model_states.pt... 0: [2022-12-25 23:06:57,678] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_09-model_00-model_states.pt. 0: [2022-12-25 23:06:57,678] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_10-model_00-model_states.pt... 0: [2022-12-25 23:06:57,850] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_10-model_00-model_states.pt. 0: [2022-12-25 23:06:57,850] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_11-model_00-model_states.pt... 0: [2022-12-25 23:06:58,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_11-model_00-model_states.pt. 0: [2022-12-25 23:06:58,021] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_12-model_00-model_states.pt... 0: [2022-12-25 23:06:58,187] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_12-model_00-model_states.pt. 0: [2022-12-25 23:06:58,188] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_13-model_00-model_states.pt... 0: [2022-12-25 23:06:58,353] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_13-model_00-model_states.pt. 0: [2022-12-25 23:06:58,353] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_14-model_00-model_states.pt... 0: [2022-12-25 23:06:58,521] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_14-model_00-model_states.pt. 0: [2022-12-25 23:06:58,521] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_15-model_00-model_states.pt... 0: [2022-12-25 23:06:58,688] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_15-model_00-model_states.pt. 0: [2022-12-25 23:06:58,689] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_16-model_00-model_states.pt... 0: [2022-12-25 23:06:58,857] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_16-model_00-model_states.pt. 0: [2022-12-25 23:06:58,858] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_17-model_00-model_states.pt... 0: [2022-12-25 23:06:59,028] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_17-model_00-model_states.pt. 0: [2022-12-25 23:06:59,028] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_18-model_00-model_states.pt... 0: [2022-12-25 23:06:59,192] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_18-model_00-model_states.pt. 0: [2022-12-25 23:06:59,192] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_19-model_00-model_states.pt... 0: [2022-12-25 23:06:59,364] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_19-model_00-model_states.pt. 0: [2022-12-25 23:06:59,364] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_20-model_00-model_states.pt... 0: [2022-12-25 23:06:59,529] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_20-model_00-model_states.pt. 0: [2022-12-25 23:06:59,529] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_21-model_00-model_states.pt... 0: [2022-12-25 23:06:59,694] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_21-model_00-model_states.pt. 0: [2022-12-25 23:06:59,695] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_22-model_00-model_states.pt... 0: [2022-12-25 23:06:59,861] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_22-model_00-model_states.pt. 0: [2022-12-25 23:06:59,862] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_23-model_00-model_states.pt... 0: [2022-12-25 23:07:00,029] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_23-model_00-model_states.pt. 0: [2022-12-25 23:07:00,030] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_24-model_00-model_states.pt... 0: [2022-12-25 23:07:00,201] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_24-model_00-model_states.pt. 0: [2022-12-25 23:07:00,201] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_25-model_00-model_states.pt... 0: [2022-12-25 23:07:00,364] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_25-model_00-model_states.pt. 0: [2022-12-25 23:07:00,365] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_26-model_00-model_states.pt... 0: [2022-12-25 23:07:00,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_26-model_00-model_states.pt. 0: [2022-12-25 23:07:00,532] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_27-model_00-model_states.pt... 0: [2022-12-25 23:07:00,706] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_27-model_00-model_states.pt. 0: [2022-12-25 23:07:00,706] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_28-model_00-model_states.pt... 0: [2022-12-25 23:07:00,870] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_28-model_00-model_states.pt. 0: [2022-12-25 23:07:00,870] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_29-model_00-model_states.pt... 0: [2022-12-25 23:07:01,037] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_29-model_00-model_states.pt. 0: [2022-12-25 23:07:01,038] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_30-model_00-model_states.pt... 0: [2022-12-25 23:07:01,208] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_30-model_00-model_states.pt. 0: [2022-12-25 23:07:01,209] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_31-model_00-model_states.pt... 0: [2022-12-25 23:07:01,374] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_31-model_00-model_states.pt. 0: [2022-12-25 23:07:01,374] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_32-model_00-model_states.pt... 0: [2022-12-25 23:07:01,543] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_32-model_00-model_states.pt. 0: [2022-12-25 23:07:01,544] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_33-model_00-model_states.pt... 0: [2022-12-25 23:07:01,707] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_33-model_00-model_states.pt. 0: [2022-12-25 23:07:01,707] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_34-model_00-model_states.pt... 0: [2022-12-25 23:07:01,881] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_34-model_00-model_states.pt. 0: [2022-12-25 23:07:01,882] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_35-model_00-model_states.pt... 0: [2022-12-25 23:07:02,048] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_35-model_00-model_states.pt. 0: [2022-12-25 23:07:02,048] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_36-model_00-model_states.pt... 0: [2022-12-25 23:07:02,214] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_36-model_00-model_states.pt. 0: [2022-12-25 23:07:02,214] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/layer_38-model_00-model_states.pt... 0: [2022-12-25 23:07:02,217] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/layer_38-model_00-model_states.pt. 0: [2022-12-25 23:07:02,219] [INFO] [logging.py:68:log_dist] [Rank 0] Saving model checkpoint: checkpoints_2b855brealtasky/global_step20000/mp_rank_00_model_states.pt 0: [2022-12-25 23:07:02,219] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/mp_rank_00_model_states.pt... 0: [2022-12-25 23:07:02,224] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/mp_rank_00_model_states.pt. 0: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... 0: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... 0: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... 12: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt... 8: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt... 8: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt... 24: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt... 24: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt... 6: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt... 6: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt... 0: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... 16: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt... 16: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt... 16: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt... 16: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt... 16: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt... 14: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt... 12: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt... 12: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt... 12: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt... 12: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt... 3: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt... 3: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt... 3: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt... 3: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt... 3: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt... 18: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt... 8: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt... 8: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt... 4: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt... 4: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt... 4: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt... 4: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt... 4: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt... 7: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt... 7: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt... 7: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt... 7: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt... 9: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt... 9: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt... 13: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt... 24: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt... 24: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt... 17: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt... 17: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt... 17: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt... 17: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt... 17: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt... 28: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt... 28: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt... 28: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt... 21: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt... 21: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt... 21: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt... 30: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt... 30: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt... 26: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt... 26: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt... 26: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt... 26: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt... 26: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt... 19: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt... 19: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt... 19: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt... 25: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt... 25: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt... 25: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt... 25: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt... 31: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt... 31: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt... 31: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt... 31: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt... 29: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt... 29: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt... 29: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt... 29: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt... 29: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt... 11: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt... 11: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt... 11: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt... 11: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt... 11: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt... 10: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt... 10: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt... 10: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt... 10: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt... 10: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt... 23: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt... 23: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt... 23: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt... 23: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt... 23: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt... 20: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt... 20: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt... 20: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt... 20: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt... 20: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt... 2: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt... 2: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt... 27: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt... 27: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt... 27: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt... 1: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... 6: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt... 6: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt... 6: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt... 15: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt... 15: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt... 15: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt... 15: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt... 0: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... 0: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... 16: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt... 16: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt... 14: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt... 14: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt... 12: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt... 3: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt... 3: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt... 18: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt... 8: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt... 8: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt... 8: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt... 5: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt... 5: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt... 5: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt... 5: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt... 5: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt... 4: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt... 4: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt... 4: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt... 7: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt... 7: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt... 7: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt... 9: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt... 9: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt... 9: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt... 9: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt... 13: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt... 13: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt... 13: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt... 13: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt... 24: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt... 24: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt... 24: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt... 24: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt... 17: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt... 17: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt... 28: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt... 28: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt... 28: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt... 28: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt... 28: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt... 21: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt... 21: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt... 21: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt... 21: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt... 30: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt... 30: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt... 30: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt... 26: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt... 19: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt... 19: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt... 25: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt... 25: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt... 31: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt... 31: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt... 31: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt... 29: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt... 29: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt... 29: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt... 11: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt... 10: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt... 23: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt... 20: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt... 2: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt... 2: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt... 2: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt... 2: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt... 2: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt... 27: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt... 27: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt... 27: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt... 22: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt... 22: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt... 22: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt... 22: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt... 22: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt... 1: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... 1: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... 1: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... 1: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... 6: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt... 6: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt... 6: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt... 15: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt... 15: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt... 0: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... 16: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt... 14: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt... 14: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt... 12: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt... 3: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt... 18: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt... 8: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt... 5: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt... 7: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt... 9: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt... 13: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt... 13: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt... 17: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt... 21: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt... 30: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt... 26: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt... 26: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt... 19: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt... 25: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt... 31: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt... 11: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt... 11: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt... 10: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt... 10: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt... 23: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt... 23: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt... 20: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt... 20: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt... 2: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt... 27: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt... 27: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt... 22: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt... 1: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... 1: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... 1: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... 15: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt... 0: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... 14: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt... 14: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt... 12: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt... 18: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt... 5: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt... 5: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt... 9: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt... 13: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt... 30: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt... 19: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt... 25: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt... 22: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt... 22: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt... 15: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt... 14: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt... 18: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt... 19: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt... 18: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt... 18: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt... 18: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt... 30: [2022-12-25 23:07:02,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt... 0: [2022-12-25 23:07:02,454] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. 0: [2022-12-25 23:07:02,454] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt 0: [2022-12-25 23:07:02,454] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 0: [2022-12-25 23:07:02,464] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. 0: [2022-12-25 23:07:02,464] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. 0: [2022-12-25 23:07:02,464] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt 0: [2022-12-25 23:07:02,464] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt 0: [2022-12-25 23:07:02,464] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 0: [2022-12-25 23:07:02,464] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 0: [2022-12-25 23:07:02,466] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. 0: [2022-12-25 23:07:02,466] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. 0: [2022-12-25 23:07:02,466] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt 0: [2022-12-25 23:07:02,466] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt 0: [2022-12-25 23:07:02,466] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 0: [2022-12-25 23:07:02,466] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 17: [2022-12-25 23:07:02,466] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt. 17: [2022-12-25 23:07:02,466] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt 17: [2022-12-25 23:07:02,467] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 17: [2022-12-25 23:07:02,467] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt. 17: [2022-12-25 23:07:02,467] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt. 17: [2022-12-25 23:07:02,467] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt 17: [2022-12-25 23:07:02,467] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt 17: [2022-12-25 23:07:02,467] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 17: [2022-12-25 23:07:02,467] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 17: [2022-12-25 23:07:02,471] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt. 17: [2022-12-25 23:07:02,471] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt 17: [2022-12-25 23:07:02,471] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 19: [2022-12-25 23:07:02,481] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt. 19: [2022-12-25 23:07:02,481] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt 19: [2022-12-25 23:07:02,481] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 19: [2022-12-25 23:07:02,486] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt. 19: [2022-12-25 23:07:02,487] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt 19: [2022-12-25 23:07:02,487] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 19: [2022-12-25 23:07:02,495] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt. 19: [2022-12-25 23:07:02,495] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt 19: [2022-12-25 23:07:02,495] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 19: [2022-12-25 23:07:02,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt. 19: [2022-12-25 23:07:02,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt 19: [2022-12-25 23:07:02,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 19: [2022-12-25 23:07:02,497] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt. 19: [2022-12-25 23:07:02,497] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt 19: [2022-12-25 23:07:02,497] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt. 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt. 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt. 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt. 16: [2022-12-25 23:07:02,498] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt. 16: [2022-12-25 23:07:02,498] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt 16: [2022-12-25 23:07:02,498] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 16: [2022-12-25 23:07:02,498] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt 16: [2022-12-25 23:07:02,498] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 16: [2022-12-25 23:07:02,498] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 0: [2022-12-25 23:07:02,499] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. 0: [2022-12-25 23:07:02,499] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. 0: [2022-12-25 23:07:02,499] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt 0: [2022-12-25 23:07:02,499] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 0: [2022-12-25 23:07:02,499] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. 0: [2022-12-25 23:07:02,499] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt 0: [2022-12-25 23:07:02,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 25: [2022-12-25 23:07:02,500] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt. 25: [2022-12-25 23:07:02,500] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt. 25: [2022-12-25 23:07:02,500] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt 25: [2022-12-25 23:07:02,500] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt 25: [2022-12-25 23:07:02,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 25: [2022-12-25 23:07:02,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt. 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt. 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt. 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt. 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt. 22: [2022-12-25 23:07:02,503] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt 22: [2022-12-25 23:07:02,503] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt 22: [2022-12-25 23:07:02,503] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 22: [2022-12-25 23:07:02,503] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt 22: [2022-12-25 23:07:02,503] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 22: [2022-12-25 23:07:02,503] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt. 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt. 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt. 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt. 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt. 12: [2022-12-25 23:07:02,509] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt 12: [2022-12-25 23:07:02,509] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt 12: [2022-12-25 23:07:02,509] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt 12: [2022-12-25 23:07:02,509] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt 12: [2022-12-25 23:07:02,509] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 12: [2022-12-25 23:07:02,509] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 25: [2022-12-25 23:07:02,512] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt. 25: [2022-12-25 23:07:02,512] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt 25: [2022-12-25 23:07:02,512] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 25: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt. 25: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt. 25: [2022-12-25 23:07:02,513] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt 25: [2022-12-25 23:07:02,513] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt 25: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 25: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 17: [2022-12-25 23:07:02,516] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt. 17: [2022-12-25 23:07:02,517] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt 17: [2022-12-25 23:07:02,517] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 19: [2022-12-25 23:07:02,520] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt. 19: [2022-12-25 23:07:02,520] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt 19: [2022-12-25 23:07:02,520] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt. 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt. 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt. 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt. 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt. 11: [2022-12-25 23:07:02,539] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt. 11: [2022-12-25 23:07:02,539] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt. 11: [2022-12-25 23:07:02,539] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt. 11: [2022-12-25 23:07:02,539] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt 11: [2022-12-25 23:07:02,539] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt 11: [2022-12-25 23:07:02,539] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt 11: [2022-12-25 23:07:02,539] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 11: [2022-12-25 23:07:02,539] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 11: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 11: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt. 11: [2022-12-25 23:07:02,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt 11: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt. 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt. 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt. 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt. 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt. 18: [2022-12-25 23:07:02,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt 18: [2022-12-25 23:07:02,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt 18: [2022-12-25 23:07:02,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt 18: [2022-12-25 23:07:02,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt 18: [2022-12-25 23:07:02,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 18: [2022-12-25 23:07:02,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt. 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt. 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt. 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt. 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt. 28: [2022-12-25 23:07:02,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt 28: [2022-12-25 23:07:02,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt 28: [2022-12-25 23:07:02,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 28: [2022-12-25 23:07:02,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt 28: [2022-12-25 23:07:02,544] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 28: [2022-12-25 23:07:02,544] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 15: [2022-12-25 23:07:02,545] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt. 15: [2022-12-25 23:07:02,545] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt. 15: [2022-12-25 23:07:02,545] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt. 15: [2022-12-25 23:07:02,545] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt. 15: [2022-12-25 23:07:02,545] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt. 15: [2022-12-25 23:07:02,545] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt 15: [2022-12-25 23:07:02,545] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt 15: [2022-12-25 23:07:02,545] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt 15: [2022-12-25 23:07:02,545] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt 15: [2022-12-25 23:07:02,545] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt 15: [2022-12-25 23:07:02,545] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 15: [2022-12-25 23:07:02,545] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 15: [2022-12-25 23:07:02,545] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 15: [2022-12-25 23:07:02,546] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 15: [2022-12-25 23:07:02,546] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 27: [2022-12-25 23:07:02,547] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt. 27: [2022-12-25 23:07:02,547] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt. 27: [2022-12-25 23:07:02,547] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt. 27: [2022-12-25 23:07:02,547] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt 27: [2022-12-25 23:07:02,547] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt 27: [2022-12-25 23:07:02,547] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt 27: [2022-12-25 23:07:02,547] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 27: [2022-12-25 23:07:02,547] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 27: [2022-12-25 23:07:02,547] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 27: [2022-12-25 23:07:02,548] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt. 27: [2022-12-25 23:07:02,548] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt 27: [2022-12-25 23:07:02,548] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 27: [2022-12-25 23:07:02,548] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt. 27: [2022-12-25 23:07:02,549] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt 27: [2022-12-25 23:07:02,549] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt. 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt. 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt. 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt. 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt. 24: [2022-12-25 23:07:02,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt 24: [2022-12-25 23:07:02,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt 24: [2022-12-25 23:07:02,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt 24: [2022-12-25 23:07:02,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt 24: [2022-12-25 23:07:02,550] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 24: [2022-12-25 23:07:02,550] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt. 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt. 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt. 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt. 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt. 20: [2022-12-25 23:07:02,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt 20: [2022-12-25 23:07:02,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt 20: [2022-12-25 23:07:02,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt 20: [2022-12-25 23:07:02,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt 20: [2022-12-25 23:07:02,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 20: [2022-12-25 23:07:02,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 10: [2022-12-25 23:07:02,554] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt. 10: [2022-12-25 23:07:02,554] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt 10: [2022-12-25 23:07:02,554] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 10: [2022-12-25 23:07:02,554] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt. 10: [2022-12-25 23:07:02,554] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt. 10: [2022-12-25 23:07:02,554] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt 10: [2022-12-25 23:07:02,554] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt 10: [2022-12-25 23:07:02,554] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 10: [2022-12-25 23:07:02,554] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 12: [2022-12-25 23:07:02,555] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt. 12: [2022-12-25 23:07:02,555] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt 12: [2022-12-25 23:07:02,555] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 10: [2022-12-25 23:07:02,565] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt. 10: [2022-12-25 23:07:02,565] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt 17: [2022-12-25 23:07:02,565] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt. 17: [2022-12-25 23:07:02,566] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt 10: [2022-12-25 23:07:02,566] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 17: [2022-12-25 23:07:02,566] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 10: [2022-12-25 23:07:02,571] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt. 10: [2022-12-25 23:07:02,572] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt 10: [2022-12-25 23:07:02,572] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt. 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt. 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt. 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt. 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt. 21: [2022-12-25 23:07:02,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt 21: [2022-12-25 23:07:02,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 21: [2022-12-25 23:07:02,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt 21: [2022-12-25 23:07:02,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt 21: [2022-12-25 23:07:02,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 21: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 9: [2022-12-25 23:07:02,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt. 9: [2022-12-25 23:07:02,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt. 9: [2022-12-25 23:07:02,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt. 9: [2022-12-25 23:07:02,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt. 9: [2022-12-25 23:07:02,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt. 9: [2022-12-25 23:07:02,577] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt 9: [2022-12-25 23:07:02,577] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt 9: [2022-12-25 23:07:02,577] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt 9: [2022-12-25 23:07:02,577] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt 9: [2022-12-25 23:07:02,577] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt 9: [2022-12-25 23:07:02,577] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 9: [2022-12-25 23:07:02,577] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 9: [2022-12-25 23:07:02,577] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 9: [2022-12-25 23:07:02,577] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 9: [2022-12-25 23:07:02,577] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt. 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt. 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt. 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt. 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt. 13: [2022-12-25 23:07:02,578] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt 13: [2022-12-25 23:07:02,578] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt 13: [2022-12-25 23:07:02,578] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 13: [2022-12-25 23:07:02,578] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt 13: [2022-12-25 23:07:02,578] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 13: [2022-12-25 23:07:02,578] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt. 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt. 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt. 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt. 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt. 31: [2022-12-25 23:07:02,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt 31: [2022-12-25 23:07:02,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt 31: [2022-12-25 23:07:02,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt 31: [2022-12-25 23:07:02,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt 31: [2022-12-25 23:07:02,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 31: [2022-12-25 23:07:02,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 30: [2022-12-25 23:07:02,513] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt 30: [2022-12-25 23:07:02,513] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt 30: [2022-12-25 23:07:02,513] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt 30: [2022-12-25 23:07:02,513] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt 30: [2022-12-25 23:07:02,513] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt. 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt. 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt. 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt. 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt. 5: [2022-12-25 23:07:02,523] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt. 5: [2022-12-25 23:07:02,523] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt. 5: [2022-12-25 23:07:02,523] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt. 5: [2022-12-25 23:07:02,523] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt. 5: [2022-12-25 23:07:02,523] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt. 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 26: [2022-12-25 23:07:02,523] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt. 26: [2022-12-25 23:07:02,523] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt. 29: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt. 29: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt. 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 30: [2022-12-25 23:07:02,513] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 5: [2022-12-25 23:07:02,524] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt 5: [2022-12-25 23:07:02,524] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt 5: [2022-12-25 23:07:02,524] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt 5: [2022-12-25 23:07:02,524] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt 5: [2022-12-25 23:07:02,524] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt 26: [2022-12-25 23:07:02,523] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt 26: [2022-12-25 23:07:02,523] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt 26: [2022-12-25 23:07:02,523] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 3: [2022-12-25 23:07:02,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 5: [2022-12-25 23:07:02,524] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 5: [2022-12-25 23:07:02,524] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 5: [2022-12-25 23:07:02,524] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 5: [2022-12-25 23:07:02,524] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 26: [2022-12-25 23:07:02,523] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 29: [2022-12-25 23:07:02,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 3: [2022-12-25 23:07:02,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 5: [2022-12-25 23:07:02,524] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 26: [2022-12-25 23:07:02,527] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt. 29: [2022-12-25 23:07:02,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt 29: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 26: [2022-12-25 23:07:02,527] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt 29: [2022-12-25 23:07:02,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 26: [2022-12-25 23:07:02,527] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 29: [2022-12-25 23:07:02,582] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt. 26: [2022-12-25 23:07:02,527] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt. 29: [2022-12-25 23:07:02,582] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt. 26: [2022-12-25 23:07:02,528] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt 29: [2022-12-25 23:07:02,583] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt 26: [2022-12-25 23:07:02,528] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 29: [2022-12-25 23:07:02,583] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt. 26: [2022-12-25 23:07:02,528] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt. 29: [2022-12-25 23:07:02,583] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt 26: [2022-12-25 23:07:02,528] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt 29: [2022-12-25 23:07:02,583] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 26: [2022-12-25 23:07:02,528] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 29: [2022-12-25 23:07:02,583] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 29: [2022-12-25 23:07:02,583] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt 26: [2022-12-25 23:07:02,553] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt. 29: [2022-12-25 23:07:02,583] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 26: [2022-12-25 23:07:02,553] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt 26: [2022-12-25 23:07:02,553] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt. 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt. 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt. 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt. 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt. 14: [2022-12-25 23:07:02,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt 14: [2022-12-25 23:07:02,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt 14: [2022-12-25 23:07:02,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt 14: [2022-12-25 23:07:02,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt 14: [2022-12-25 23:07:02,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 14: [2022-12-25 23:07:02,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt. 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt. 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt. 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt. 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt. 7: [2022-12-25 23:07:02,589] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt 7: [2022-12-25 23:07:02,589] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt 7: [2022-12-25 23:07:02,589] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt 7: [2022-12-25 23:07:02,589] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt 7: [2022-12-25 23:07:02,589] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 7: [2022-12-25 23:07:02,589] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 25: [2022-12-25 23:07:02,595] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt. 25: [2022-12-25 23:07:02,595] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt 25: [2022-12-25 23:07:02,595] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 11: [2022-12-25 23:07:02,595] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt. 11: [2022-12-25 23:07:02,595] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt 11: [2022-12-25 23:07:02,595] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt. 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt. 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt. 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt. 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt. 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt. 4: [2022-12-25 23:07:02,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt 4: [2022-12-25 23:07:02,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt 4: [2022-12-25 23:07:02,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt 4: [2022-12-25 23:07:02,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt 4: [2022-12-25 23:07:02,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt 4: [2022-12-25 23:07:02,620] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 4: [2022-12-25 23:07:02,620] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt. 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt. 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt. 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt. 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt. 6: [2022-12-25 23:07:02,622] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt 6: [2022-12-25 23:07:02,622] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt 6: [2022-12-25 23:07:02,622] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt. 6: [2022-12-25 23:07:02,622] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 6: [2022-12-25 23:07:02,622] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 6: [2022-12-25 23:07:02,622] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 6: [2022-12-25 23:07:02,622] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 10: [2022-12-25 23:07:02,623] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt. 10: [2022-12-25 23:07:02,623] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt 10: [2022-12-25 23:07:02,623] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt. 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt. 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt. 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt. 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt. 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt. 23: [2022-12-25 23:07:02,632] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt 23: [2022-12-25 23:07:02,632] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt 23: [2022-12-25 23:07:02,632] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt 23: [2022-12-25 23:07:02,632] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt 23: [2022-12-25 23:07:02,632] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt 23: [2022-12-25 23:07:02,632] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 23: [2022-12-25 23:07:02,632] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 17: [2022-12-25 23:07:02,636] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt. 17: [2022-12-25 23:07:02,636] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt 17: [2022-12-25 23:07:02,636] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 17: [2022-12-25 23:07:02,637] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt. 17: [2022-12-25 23:07:02,637] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt 17: [2022-12-25 23:07:02,637] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 15: [2022-12-25 23:07:02,642] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt. 15: [2022-12-25 23:07:02,642] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt 15: [2022-12-25 23:07:02,642] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 24: [2022-12-25 23:07:02,647] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt. 24: [2022-12-25 23:07:02,647] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt 24: [2022-12-25 23:07:02,647] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 20: [2022-12-25 23:07:02,660] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt. 20: [2022-12-25 23:07:02,660] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt 20: [2022-12-25 23:07:02,660] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 0: [2022-12-25 23:07:02,667] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt 0: [2022-12-25 23:07:02,667] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 27: [2022-12-25 23:07:02,671] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt. 27: [2022-12-25 23:07:02,671] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt 27: [2022-12-25 23:07:02,671] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 5: [2022-12-25 23:07:02,672] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt. 5: [2022-12-25 23:07:02,672] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt 5: [2022-12-25 23:07:02,672] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 23: [2022-12-25 23:07:02,680] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt. 23: [2022-12-25 23:07:02,680] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt 23: [2022-12-25 23:07:02,680] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. 22: [2022-12-25 23:07:02,698] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt. 22: [2022-12-25 23:07:02,698] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt 22: [2022-12-25 23:07:02,698] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 30: [2022-12-25 23:07:02,700] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt. 30: [2022-12-25 23:07:02,700] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt 30: [2022-12-25 23:07:02,700] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 8: [2022-12-25 23:07:02,702] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt. 8: [2022-12-25 23:07:02,702] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt. 8: [2022-12-25 23:07:02,702] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt. 8: [2022-12-25 23:07:02,702] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt. 8: [2022-12-25 23:07:02,702] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt. 16: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt. 16: [2022-12-25 23:07:02,703] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt 16: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 8: [2022-12-25 23:07:02,702] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt. 8: [2022-12-25 23:07:02,702] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt. 8: [2022-12-25 23:07:02,702] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt. 8: [2022-12-25 23:07:02,702] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt 8: [2022-12-25 23:07:02,702] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt 8: [2022-12-25 23:07:02,702] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt 8: [2022-12-25 23:07:02,702] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt 8: [2022-12-25 23:07:02,702] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt 8: [2022-12-25 23:07:02,702] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt 8: [2022-12-25 23:07:02,702] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt 8: [2022-12-25 23:07:02,702] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt 8: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 8: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 8: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 8: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 8: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 8: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 8: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 8: [2022-12-25 23:07:02,703] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. 1: [2022-12-25 23:07:02,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt 1: [2022-12-25 23:07:02,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt 1: [2022-12-25 23:07:02,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt 1: [2022-12-25 23:07:02,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt 1: [2022-12-25 23:07:02,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt 1: [2022-12-25 23:07:02,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 1: [2022-12-25 23:07:02,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt 1: [2022-12-25 23:07:02,697] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 1: [2022-12-25 23:07:02,697] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 9: [2022-12-25 23:07:02,712] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt. 9: [2022-12-25 23:07:02,712] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt 9: [2022-12-25 23:07:02,712] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 29: [2022-12-25 23:07:02,713] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt. 29: [2022-12-25 23:07:02,713] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt 29: [2022-12-25 23:07:02,714] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 19: [2022-12-25 23:07:02,720] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt. 19: [2022-12-25 23:07:02,720] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt 19: [2022-12-25 23:07:02,720] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 11: [2022-12-25 23:07:02,722] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt. 11: [2022-12-25 23:07:02,723] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt 11: [2022-12-25 23:07:02,723] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt. 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt. 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt. 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt. 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt. 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt. 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt. 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt. 2: [2022-12-25 23:07:02,726] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt 2: [2022-12-25 23:07:02,726] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt 2: [2022-12-25 23:07:02,726] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt 2: [2022-12-25 23:07:02,726] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt 2: [2022-12-25 23:07:02,726] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt 2: [2022-12-25 23:07:02,726] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 2: [2022-12-25 23:07:02,726] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt 2: [2022-12-25 23:07:02,726] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 2: [2022-12-25 23:07:02,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 3: [2022-12-25 23:07:02,735] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt. 3: [2022-12-25 23:07:02,735] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,735] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 28: [2022-12-25 23:07:02,736] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt. 28: [2022-12-25 23:07:02,736] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt 28: [2022-12-25 23:07:02,736] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 25: [2022-12-25 23:07:02,738] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt. 25: [2022-12-25 23:07:02,739] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt 25: [2022-12-25 23:07:02,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 6: [2022-12-25 23:07:02,739] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt. 6: [2022-12-25 23:07:02,739] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt 6: [2022-12-25 23:07:02,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 18: [2022-12-25 23:07:02,743] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt. 31: [2022-12-25 23:07:02,743] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt. 31: [2022-12-25 23:07:02,743] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt 18: [2022-12-25 23:07:02,743] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt 31: [2022-12-25 23:07:02,743] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 18: [2022-12-25 23:07:02,743] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 21: [2022-12-25 23:07:02,743] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt. 21: [2022-12-25 23:07:02,743] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt 21: [2022-12-25 23:07:02,743] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 7: [2022-12-25 23:07:02,744] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt. 7: [2022-12-25 23:07:02,744] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt 7: [2022-12-25 23:07:02,744] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 13: [2022-12-25 23:07:02,745] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt. 13: [2022-12-25 23:07:02,745] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt 13: [2022-12-25 23:07:02,745] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 12: [2022-12-25 23:07:02,751] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt. 12: [2022-12-25 23:07:02,751] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt 12: [2022-12-25 23:07:02,751] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 4: [2022-12-25 23:07:02,752] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt. 4: [2022-12-25 23:07:02,752] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt 4: [2022-12-25 23:07:02,752] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 24: [2022-12-25 23:07:02,753] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt. 24: [2022-12-25 23:07:02,753] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt 24: [2022-12-25 23:07:02,753] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 14: [2022-12-25 23:07:02,755] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt. 14: [2022-12-25 23:07:02,755] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt 14: [2022-12-25 23:07:02,755] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 20: [2022-12-25 23:07:02,755] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt. 20: [2022-12-25 23:07:02,756] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt 20: [2022-12-25 23:07:02,756] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 15: [2022-12-25 23:07:02,757] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt. 15: [2022-12-25 23:07:02,757] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt 15: [2022-12-25 23:07:02,757] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 26: [2022-12-25 23:07:02,758] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt. 26: [2022-12-25 23:07:02,759] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt 26: [2022-12-25 23:07:02,759] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 23: [2022-12-25 23:07:02,761] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt. 23: [2022-12-25 23:07:02,761] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt 23: [2022-12-25 23:07:02,761] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 16: [2022-12-25 23:07:02,767] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt. 16: [2022-12-25 23:07:02,768] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt 16: [2022-12-25 23:07:02,768] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 27: [2022-12-25 23:07:02,771] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt. 27: [2022-12-25 23:07:02,771] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt 27: [2022-12-25 23:07:02,771] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 5: [2022-12-25 23:07:02,773] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt. 5: [2022-12-25 23:07:02,773] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt 5: [2022-12-25 23:07:02,773] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 10: [2022-12-25 23:07:02,776] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt. 10: [2022-12-25 23:07:02,776] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt 10: [2022-12-25 23:07:02,776] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 9: [2022-12-25 23:07:02,779] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt. 9: [2022-12-25 23:07:02,779] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt 9: [2022-12-25 23:07:02,779] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 30: [2022-12-25 23:07:02,787] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt. 30: [2022-12-25 23:07:02,787] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt 30: [2022-12-25 23:07:02,787] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 4: [2022-12-25 23:07:02,787] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt. 4: [2022-12-25 23:07:02,787] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt 15: [2022-12-25 23:07:02,787] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt. 4: [2022-12-25 23:07:02,787] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 15: [2022-12-25 23:07:02,788] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt 20: [2022-12-25 23:07:02,786] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt. 15: [2022-12-25 23:07:02,788] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 20: [2022-12-25 23:07:02,787] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt 20: [2022-12-25 23:07:02,787] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 24: [2022-12-25 23:07:02,788] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt. 24: [2022-12-25 23:07:02,788] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt 24: [2022-12-25 23:07:02,788] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 26: [2022-12-25 23:07:02,791] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt. 26: [2022-12-25 23:07:02,791] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt 26: [2022-12-25 23:07:02,791] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 10: [2022-12-25 23:07:02,791] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt. 10: [2022-12-25 23:07:02,791] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt 10: [2022-12-25 23:07:02,791] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 29: [2022-12-25 23:07:02,791] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt. 29: [2022-12-25 23:07:02,791] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt 29: [2022-12-25 23:07:02,791] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt. 29: [2022-12-25 23:07:02,791] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 29: [2022-12-25 23:07:02,791] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt 29: [2022-12-25 23:07:02,792] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 16: [2022-12-25 23:07:02,793] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt. 16: [2022-12-25 23:07:02,793] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt 16: [2022-12-25 23:07:02,793] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 30: [2022-12-25 23:07:02,793] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt. 30: [2022-12-25 23:07:02,793] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt 30: [2022-12-25 23:07:02,793] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 21: [2022-12-25 23:07:02,794] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt. 21: [2022-12-25 23:07:02,794] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt 21: [2022-12-25 23:07:02,794] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 6: [2022-12-25 23:07:02,795] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt. 6: [2022-12-25 23:07:02,795] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt 6: [2022-12-25 23:07:02,795] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 12: [2022-12-25 23:07:02,795] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt. 12: [2022-12-25 23:07:02,796] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt 12: [2022-12-25 23:07:02,796] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 7: [2022-12-25 23:07:02,796] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt. 7: [2022-12-25 23:07:02,796] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt 7: [2022-12-25 23:07:02,796] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 22: [2022-12-25 23:07:02,796] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt. 22: [2022-12-25 23:07:02,797] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt 22: [2022-12-25 23:07:02,797] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 31: [2022-12-25 23:07:02,797] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt. 31: [2022-12-25 23:07:02,797] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt 31: [2022-12-25 23:07:02,797] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 11: [2022-12-25 23:07:02,797] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt. 11: [2022-12-25 23:07:02,798] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt 11: [2022-12-25 23:07:02,798] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 25: [2022-12-25 23:07:02,800] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt. 25: [2022-12-25 23:07:02,800] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt 25: [2022-12-25 23:07:02,800] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 28: [2022-12-25 23:07:02,800] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt. 28: [2022-12-25 23:07:02,800] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt. 28: [2022-12-25 23:07:02,800] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt 28: [2022-12-25 23:07:02,800] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt 28: [2022-12-25 23:07:02,800] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 28: [2022-12-25 23:07:02,800] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 19: [2022-12-25 23:07:02,802] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt. 19: [2022-12-25 23:07:02,803] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt 19: [2022-12-25 23:07:02,803] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 3: [2022-12-25 23:07:02,803] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt. 3: [2022-12-25 23:07:02,803] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,803] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 22: [2022-12-25 23:07:02,803] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt. 22: [2022-12-25 23:07:02,804] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt 22: [2022-12-25 23:07:02,804] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 27: [2022-12-25 23:07:02,804] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt. 27: [2022-12-25 23:07:02,804] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt 27: [2022-12-25 23:07:02,804] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 9: [2022-12-25 23:07:02,804] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt. 9: [2022-12-25 23:07:02,805] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt 9: [2022-12-25 23:07:02,805] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 3: [2022-12-25 23:07:02,805] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt. 3: [2022-12-25 23:07:02,805] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt 3: [2022-12-25 23:07:02,805] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 5: [2022-12-25 23:07:02,805] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt. 5: [2022-12-25 23:07:02,805] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt 5: [2022-12-25 23:07:02,805] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 31: [2022-12-25 23:07:02,806] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt. 31: [2022-12-25 23:07:02,806] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt 31: [2022-12-25 23:07:02,806] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 11: [2022-12-25 23:07:02,806] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt. 11: [2022-12-25 23:07:02,806] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt 11: [2022-12-25 23:07:02,806] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 7: [2022-12-25 23:07:02,807] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt. 13: [2022-12-25 23:07:02,807] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt. 7: [2022-12-25 23:07:02,807] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt 13: [2022-12-25 23:07:02,807] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt. 7: [2022-12-25 23:07:02,807] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 13: [2022-12-25 23:07:02,807] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt 13: [2022-12-25 23:07:02,807] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt 13: [2022-12-25 23:07:02,807] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 13: [2022-12-25 23:07:02,807] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 21: [2022-12-25 23:07:02,808] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt. 21: [2022-12-25 23:07:02,808] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt 21: [2022-12-25 23:07:02,808] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 18: [2022-12-25 23:07:02,809] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt. 18: [2022-12-25 23:07:02,810] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt 18: [2022-12-25 23:07:02,810] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 14: [2022-12-25 23:07:02,813] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt. 14: [2022-12-25 23:07:02,813] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt 14: [2022-12-25 23:07:02,813] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 18: [2022-12-25 23:07:02,815] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt. 18: [2022-12-25 23:07:02,815] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt 18: [2022-12-25 23:07:02,815] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 14: [2022-12-25 23:07:02,819] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt. 14: [2022-12-25 23:07:02,819] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step20000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt 14: [2022-12-25 23:07:02,819] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step20000 is ready now! 0: successfully saved checkpoint at iteration 20000 to checkpoints_2b855brealtasky 31: time (ms) | save-checkpoint: 6851.84 31: iteration 20010/ 52452 | consumed samples: 10245120 | consumed tokens: 20982005760 | elapsed time per iteration (s): 9.85 | learning rate: 1.444E-04 | global batch size: 512 | lm loss: 1.381345E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 51.954 | TFLOPs: 7.80 | 31: iteration 20020/ 52452 | consumed samples: 10250240 | consumed tokens: 20992491520 | elapsed time per iteration (s): 1.76 | learning rate: 1.443E-04 | global batch size: 512 | lm loss: 1.387321E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.905 | TFLOPs: 43.66 | 31: iteration 20030/ 52452 | consumed samples: 10255360 | consumed tokens: 21002977280 | elapsed time per iteration (s): 1.77 | learning rate: 1.443E-04 | global batch size: 512 | lm loss: 1.396236E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.311 | TFLOPs: 43.42 | 31: iteration 20040/ 52452 | consumed samples: 10260480 | consumed tokens: 21013463040 | elapsed time per iteration (s): 1.80 | learning rate: 1.442E-04 | global batch size: 512 | lm loss: 1.387490E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.086 | TFLOPs: 42.79 | 31: iteration 20050/ 52452 | consumed samples: 10265600 | consumed tokens: 21023948800 | elapsed time per iteration (s): 1.83 | learning rate: 1.442E-04 | global batch size: 512 | lm loss: 1.396171E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.297 | TFLOPs: 41.92 | 31: iteration 20060/ 52452 | consumed samples: 10270720 | consumed tokens: 21034434560 | elapsed time per iteration (s): 1.85 | learning rate: 1.441E-04 | global batch size: 512 | lm loss: 1.376953E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.424 | TFLOPs: 41.49 | 31: iteration 20070/ 52452 | consumed samples: 10275840 | consumed tokens: 21044920320 | elapsed time per iteration (s): 1.79 | learning rate: 1.441E-04 | global batch size: 512 | lm loss: 1.401761E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.459 | TFLOPs: 42.85 | 31: iteration 20080/ 52452 | consumed samples: 10280960 | consumed tokens: 21055406080 | elapsed time per iteration (s): 1.81 | learning rate: 1.440E-04 | global batch size: 512 | lm loss: 1.385703E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.505 | TFLOPs: 42.55 | 31: iteration 20090/ 52452 | consumed samples: 10286080 | consumed tokens: 21065891840 | elapsed time per iteration (s): 1.77 | learning rate: 1.440E-04 | global batch size: 512 | lm loss: 1.392374E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.034 | TFLOPs: 43.38 | 31: iteration 20100/ 52452 | consumed samples: 10291200 | consumed tokens: 21076377600 | elapsed time per iteration (s): 1.83 | learning rate: 1.439E-04 | global batch size: 512 | lm loss: 1.379725E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.006 | TFLOPs: 42.03 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20100 | lm loss value: 4.106543E+00 | lm loss PPL: 6.073639E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20100 | lm loss value: 4.055366E+00 | lm loss PPL: 5.770625E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 20110/ 52452 | consumed samples: 10296320 | consumed tokens: 21086863360 | elapsed time per iteration (s): 9.19 | learning rate: 1.439E-04 | global batch size: 512 | lm loss: 1.389979E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.688 | TFLOPs: 8.36 | 31: iteration 20120/ 52452 | consumed samples: 10301440 | consumed tokens: 21097349120 | elapsed time per iteration (s): 1.79 | learning rate: 1.438E-04 | global batch size: 512 | lm loss: 1.385445E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.721 | TFLOPs: 42.89 | 31: iteration 20130/ 52452 | consumed samples: 10306560 | consumed tokens: 21107834880 | elapsed time per iteration (s): 1.79 | learning rate: 1.438E-04 | global batch size: 512 | lm loss: 1.384429E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.577 | TFLOPs: 43.01 | 31: iteration 20140/ 52452 | consumed samples: 10311680 | consumed tokens: 21118320640 | elapsed time per iteration (s): 1.76 | learning rate: 1.437E-04 | global batch size: 512 | lm loss: 1.388336E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.818 | TFLOPs: 43.65 | 31: iteration 20150/ 52452 | consumed samples: 10316800 | consumed tokens: 21128806400 | elapsed time per iteration (s): 1.77 | learning rate: 1.437E-04 | global batch size: 512 | lm loss: 1.379496E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.861 | TFLOPs: 43.51 | 31: iteration 20160/ 52452 | consumed samples: 10321920 | consumed tokens: 21139292160 | elapsed time per iteration (s): 1.77 | learning rate: 1.436E-04 | global batch size: 512 | lm loss: 1.402033E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.515 | TFLOPs: 43.45 | 31: iteration 20170/ 52452 | consumed samples: 10327040 | consumed tokens: 21149777920 | elapsed time per iteration (s): 1.81 | learning rate: 1.436E-04 | global batch size: 512 | lm loss: 1.383338E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.005 | TFLOPs: 42.48 | 31: iteration 20180/ 52452 | consumed samples: 10332160 | consumed tokens: 21160263680 | elapsed time per iteration (s): 1.77 | learning rate: 1.435E-04 | global batch size: 512 | lm loss: 1.380350E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.182 | TFLOPs: 43.40 | 31: iteration 20190/ 52452 | consumed samples: 10337280 | consumed tokens: 21170749440 | elapsed time per iteration (s): 1.82 | learning rate: 1.435E-04 | global batch size: 512 | lm loss: 1.394584E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.233 | TFLOPs: 42.21 | 31: iteration 20200/ 52452 | consumed samples: 10342400 | consumed tokens: 21181235200 | elapsed time per iteration (s): 1.87 | learning rate: 1.434E-04 | global batch size: 512 | lm loss: 1.398624E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.307 | TFLOPs: 41.17 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20200 | lm loss value: 4.181778E+00 | lm loss PPL: 6.548217E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20200 | lm loss value: 4.113899E+00 | lm loss PPL: 6.118480E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 20210/ 52452 | consumed samples: 10347520 | consumed tokens: 21191720960 | elapsed time per iteration (s): 9.32 | learning rate: 1.434E-04 | global batch size: 512 | lm loss: 1.389759E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.917 | TFLOPs: 8.24 | 31: iteration 20220/ 52452 | consumed samples: 10352640 | consumed tokens: 21202206720 | elapsed time per iteration (s): 1.74 | learning rate: 1.433E-04 | global batch size: 512 | lm loss: 1.392177E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.356 | TFLOPs: 44.18 | 31: iteration 20230/ 52452 | consumed samples: 10357760 | consumed tokens: 21212692480 | elapsed time per iteration (s): 1.79 | learning rate: 1.433E-04 | global batch size: 512 | lm loss: 1.397864E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.237 | TFLOPs: 42.96 | 31: iteration 20240/ 52452 | consumed samples: 10362880 | consumed tokens: 21223178240 | elapsed time per iteration (s): 1.73 | learning rate: 1.432E-04 | global batch size: 512 | lm loss: 1.389327E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.284 | TFLOPs: 44.47 | 31: iteration 20250/ 52452 | consumed samples: 10368000 | consumed tokens: 21233664000 | elapsed time per iteration (s): 1.79 | learning rate: 1.432E-04 | global batch size: 512 | lm loss: 1.380482E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.416 | TFLOPs: 42.84 | 31: iteration 20260/ 52452 | consumed samples: 10373120 | consumed tokens: 21244149760 | elapsed time per iteration (s): 1.84 | learning rate: 1.431E-04 | global batch size: 512 | lm loss: 1.379063E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.392 | TFLOPs: 41.79 | 31: iteration 20270/ 52452 | consumed samples: 10378240 | consumed tokens: 21254635520 | elapsed time per iteration (s): 1.76 | learning rate: 1.431E-04 | global batch size: 512 | lm loss: 1.383571E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.603 | TFLOPs: 43.62 | 31: iteration 20280/ 52452 | consumed samples: 10383360 | consumed tokens: 21265121280 | elapsed time per iteration (s): 1.75 | learning rate: 1.430E-04 | global batch size: 512 | lm loss: 1.383534E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.612 | TFLOPs: 43.92 | 31: iteration 20290/ 52452 | consumed samples: 10388480 | consumed tokens: 21275607040 | elapsed time per iteration (s): 1.75 | learning rate: 1.430E-04 | global batch size: 512 | lm loss: 1.390881E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.444 | TFLOPs: 43.89 | 31: iteration 20300/ 52452 | consumed samples: 10393600 | consumed tokens: 21286092800 | elapsed time per iteration (s): 1.76 | learning rate: 1.429E-04 | global batch size: 512 | lm loss: 1.379895E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.549 | TFLOPs: 43.76 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20300 | lm loss value: 4.140336E+00 | lm loss PPL: 6.282393E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20300 | lm loss value: 4.084621E+00 | lm loss PPL: 5.941944E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 20310/ 52452 | consumed samples: 10398720 | consumed tokens: 21296578560 | elapsed time per iteration (s): 9.13 | learning rate: 1.429E-04 | global batch size: 512 | lm loss: 1.391478E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.049 | TFLOPs: 8.41 | 31: iteration 20320/ 52452 | consumed samples: 10403840 | consumed tokens: 21307064320 | elapsed time per iteration (s): 1.74 | learning rate: 1.428E-04 | global batch size: 512 | lm loss: 1.391612E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.948 | TFLOPs: 44.12 | 31: iteration 20330/ 52452 | consumed samples: 10408960 | consumed tokens: 21317550080 | elapsed time per iteration (s): 1.78 | learning rate: 1.428E-04 | global batch size: 512 | lm loss: 1.403983E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.598 | TFLOPs: 43.17 | 31: iteration 20340/ 52452 | consumed samples: 10414080 | consumed tokens: 21328035840 | elapsed time per iteration (s): 1.80 | learning rate: 1.427E-04 | global batch size: 512 | lm loss: 1.393406E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.064 | TFLOPs: 42.64 | 31: iteration 20350/ 52452 | consumed samples: 10419200 | consumed tokens: 21338521600 | elapsed time per iteration (s): 1.77 | learning rate: 1.427E-04 | global batch size: 512 | lm loss: 1.391639E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.363 | TFLOPs: 43.43 | 31: iteration 20360/ 52452 | consumed samples: 10424320 | consumed tokens: 21349007360 | elapsed time per iteration (s): 1.83 | learning rate: 1.426E-04 | global batch size: 512 | lm loss: 1.385619E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.405 | TFLOPs: 42.09 | 31: iteration 20370/ 52452 | consumed samples: 10429440 | consumed tokens: 21359493120 | elapsed time per iteration (s): 1.79 | learning rate: 1.426E-04 | global batch size: 512 | lm loss: 1.385680E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.394 | TFLOPs: 42.99 | 31: iteration 20380/ 52452 | consumed samples: 10434560 | consumed tokens: 21369978880 | elapsed time per iteration (s): 1.87 | learning rate: 1.425E-04 | global batch size: 512 | lm loss: 1.388738E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.926 | TFLOPs: 41.11 | 31: iteration 20390/ 52452 | consumed samples: 10439680 | consumed tokens: 21380464640 | elapsed time per iteration (s): 1.81 | learning rate: 1.425E-04 | global batch size: 512 | lm loss: 1.385628E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.963 | TFLOPs: 42.47 | 31: iteration 20400/ 52452 | consumed samples: 10444800 | consumed tokens: 21390950400 | elapsed time per iteration (s): 1.76 | learning rate: 1.424E-04 | global batch size: 512 | lm loss: 1.389950E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.185 | TFLOPs: 43.71 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20400 | lm loss value: 4.102358E+00 | lm loss PPL: 6.048273E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20400 | lm loss value: 4.051383E+00 | lm loss PPL: 5.747692E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 20410/ 52452 | consumed samples: 10449920 | consumed tokens: 21401436160 | elapsed time per iteration (s): 9.12 | learning rate: 1.424E-04 | global batch size: 512 | lm loss: 1.389152E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.127 | TFLOPs: 8.42 | 31: iteration 20420/ 52452 | consumed samples: 10455040 | consumed tokens: 21411921920 | elapsed time per iteration (s): 1.80 | learning rate: 1.423E-04 | global batch size: 512 | lm loss: 1.381927E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.860 | TFLOPs: 42.61 | 31: iteration 20430/ 52452 | consumed samples: 10460160 | consumed tokens: 21422407680 | elapsed time per iteration (s): 1.85 | learning rate: 1.423E-04 | global batch size: 512 | lm loss: 1.389288E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.466 | TFLOPs: 41.50 | 31: iteration 20440/ 52452 | consumed samples: 10465280 | consumed tokens: 21432893440 | elapsed time per iteration (s): 2.21 | learning rate: 1.422E-04 | global batch size: 512 | lm loss: 1.374922E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 231.881 | TFLOPs: 34.80 | 31: iteration 20450/ 52452 | consumed samples: 10470400 | consumed tokens: 21443379200 | elapsed time per iteration (s): 1.81 | learning rate: 1.422E-04 | global batch size: 512 | lm loss: 1.370615E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.033 | TFLOPs: 42.48 | 31: iteration 20460/ 52452 | consumed samples: 10475520 | consumed tokens: 21453864960 | elapsed time per iteration (s): 1.79 | learning rate: 1.421E-04 | global batch size: 512 | lm loss: 1.388130E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.727 | TFLOPs: 42.89 | 31: iteration 20470/ 52452 | consumed samples: 10480640 | consumed tokens: 21464350720 | elapsed time per iteration (s): 1.77 | learning rate: 1.421E-04 | global batch size: 512 | lm loss: 1.397560E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.758 | TFLOPs: 43.34 | 31: iteration 20480/ 52452 | consumed samples: 10485760 | consumed tokens: 21474836480 | elapsed time per iteration (s): 1.77 | learning rate: 1.420E-04 | global batch size: 512 | lm loss: 1.373910E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.861 | TFLOPs: 43.36 | 31: iteration 20490/ 52452 | consumed samples: 10490880 | consumed tokens: 21485322240 | elapsed time per iteration (s): 1.79 | learning rate: 1.419E-04 | global batch size: 512 | lm loss: 1.392713E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.533 | TFLOPs: 43.01 | 31: iteration 20500/ 52452 | consumed samples: 10496000 | consumed tokens: 21495808000 | elapsed time per iteration (s): 1.75 | learning rate: 1.419E-04 | global batch size: 512 | lm loss: 1.398291E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.314 | TFLOPs: 43.87 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20500 | lm loss value: 4.307053E+00 | lm loss PPL: 7.422141E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20500 | lm loss value: 4.192196E+00 | lm loss PPL: 6.616796E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 20510/ 52452 | consumed samples: 10501120 | consumed tokens: 21506293760 | elapsed time per iteration (s): 9.21 | learning rate: 1.418E-04 | global batch size: 512 | lm loss: 1.383223E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.607 | TFLOPs: 8.35 | 31: iteration 20520/ 52452 | consumed samples: 10506240 | consumed tokens: 21516779520 | elapsed time per iteration (s): 1.85 | learning rate: 1.418E-04 | global batch size: 512 | lm loss: 1.381123E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.913 | TFLOPs: 41.56 | 31: iteration 20530/ 52452 | consumed samples: 10511360 | consumed tokens: 21527265280 | elapsed time per iteration (s): 1.85 | learning rate: 1.417E-04 | global batch size: 512 | lm loss: 1.378539E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.205 | TFLOPs: 41.61 | 31: iteration 20540/ 52452 | consumed samples: 10516480 | consumed tokens: 21537751040 | elapsed time per iteration (s): 1.75 | learning rate: 1.417E-04 | global batch size: 512 | lm loss: 1.386541E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.868 | TFLOPs: 43.96 | 31: iteration 20550/ 52452 | consumed samples: 10521600 | consumed tokens: 21548236800 | elapsed time per iteration (s): 1.76 | learning rate: 1.416E-04 | global batch size: 512 | lm loss: 1.401015E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.828 | TFLOPs: 43.65 | 31: iteration 20560/ 52452 | consumed samples: 10526720 | consumed tokens: 21558722560 | elapsed time per iteration (s): 1.74 | learning rate: 1.416E-04 | global batch size: 512 | lm loss: 1.387187E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.867 | TFLOPs: 44.26 | 31: iteration 20570/ 52452 | consumed samples: 10531840 | consumed tokens: 21569208320 | elapsed time per iteration (s): 1.78 | learning rate: 1.415E-04 | global batch size: 512 | lm loss: 1.376322E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.672 | TFLOPs: 43.18 | 31: iteration 20580/ 52452 | consumed samples: 10536960 | consumed tokens: 21579694080 | elapsed time per iteration (s): 1.77 | learning rate: 1.415E-04 | global batch size: 512 | lm loss: 1.380000E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.493 | TFLOPs: 43.45 | 31: iteration 20590/ 52452 | consumed samples: 10542080 | consumed tokens: 21590179840 | elapsed time per iteration (s): 1.74 | learning rate: 1.414E-04 | global batch size: 512 | lm loss: 1.395466E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.838 | TFLOPs: 44.10 | 31: iteration 20600/ 52452 | consumed samples: 10547200 | consumed tokens: 21600665600 | elapsed time per iteration (s): 1.80 | learning rate: 1.414E-04 | global batch size: 512 | lm loss: 1.381630E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.470 | TFLOPs: 42.70 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20600 | lm loss value: 4.175066E+00 | lm loss PPL: 6.504410E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20600 | lm loss value: 4.149166E+00 | lm loss PPL: 6.338109E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 20610/ 52452 | consumed samples: 10552320 | consumed tokens: 21611151360 | elapsed time per iteration (s): 9.13 | learning rate: 1.413E-04 | global batch size: 512 | lm loss: 1.383997E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.049 | TFLOPs: 8.41 | 31: iteration 20620/ 52452 | consumed samples: 10557440 | consumed tokens: 21621637120 | elapsed time per iteration (s): 1.73 | learning rate: 1.413E-04 | global batch size: 512 | lm loss: 1.385634E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.955 | TFLOPs: 44.42 | 31: iteration 20630/ 52452 | consumed samples: 10562560 | consumed tokens: 21632122880 | elapsed time per iteration (s): 1.79 | learning rate: 1.412E-04 | global batch size: 512 | lm loss: 1.386545E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.809 | TFLOPs: 43.05 | 31: iteration 20640/ 52452 | consumed samples: 10567680 | consumed tokens: 21642608640 | elapsed time per iteration (s): 1.75 | learning rate: 1.412E-04 | global batch size: 512 | lm loss: 1.366857E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.984 | TFLOPs: 43.98 | 31: iteration 20650/ 52452 | consumed samples: 10572800 | consumed tokens: 21653094400 | elapsed time per iteration (s): 1.81 | learning rate: 1.411E-04 | global batch size: 512 | lm loss: 1.388597E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.599 | TFLOPs: 42.57 | 31: iteration 20660/ 52452 | consumed samples: 10577920 | consumed tokens: 21663580160 | elapsed time per iteration (s): 1.79 | learning rate: 1.411E-04 | global batch size: 512 | lm loss: 1.380006E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.958 | TFLOPs: 42.92 | 31: iteration 20670/ 52452 | consumed samples: 10583040 | consumed tokens: 21674065920 | elapsed time per iteration (s): 1.76 | learning rate: 1.410E-04 | global batch size: 512 | lm loss: 1.374175E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.170 | TFLOPs: 43.70 | 31: iteration 20680/ 52452 | consumed samples: 10588160 | consumed tokens: 21684551680 | elapsed time per iteration (s): 1.77 | learning rate: 1.410E-04 | global batch size: 512 | lm loss: 1.374315E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.653 | TFLOPs: 43.48 | 31: iteration 20690/ 52452 | consumed samples: 10593280 | consumed tokens: 21695037440 | elapsed time per iteration (s): 1.81 | learning rate: 1.409E-04 | global batch size: 512 | lm loss: 1.382548E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.042 | TFLOPs: 42.48 | 31: iteration 20700/ 52452 | consumed samples: 10598400 | consumed tokens: 21705523200 | elapsed time per iteration (s): 1.83 | learning rate: 1.409E-04 | global batch size: 512 | lm loss: 1.376528E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.614 | TFLOPs: 41.97 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20700 | lm loss value: 4.220490E+00 | lm loss PPL: 6.806683E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20700 | lm loss value: 4.135365E+00 | lm loss PPL: 6.251243E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 20710/ 52452 | consumed samples: 10603520 | consumed tokens: 21716008960 | elapsed time per iteration (s): 9.17 | learning rate: 1.408E-04 | global batch size: 512 | lm loss: 1.380984E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.809 | TFLOPs: 8.38 | 31: iteration 20720/ 52452 | consumed samples: 10608640 | consumed tokens: 21726494720 | elapsed time per iteration (s): 1.86 | learning rate: 1.408E-04 | global batch size: 512 | lm loss: 1.384716E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.747 | TFLOPs: 41.39 | 31: iteration 20730/ 52452 | consumed samples: 10613760 | consumed tokens: 21736980480 | elapsed time per iteration (s): 1.78 | learning rate: 1.407E-04 | global batch size: 512 | lm loss: 1.375327E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.116 | TFLOPs: 43.09 | 31: iteration 20740/ 52452 | consumed samples: 10618880 | consumed tokens: 21747466240 | elapsed time per iteration (s): 1.77 | learning rate: 1.407E-04 | global batch size: 512 | lm loss: 1.377598E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.298 | TFLOPs: 43.42 | 31: iteration 20750/ 52452 | consumed samples: 10624000 | consumed tokens: 21757952000 | elapsed time per iteration (s): 1.81 | learning rate: 1.406E-04 | global batch size: 512 | lm loss: 1.379648E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.637 | TFLOPs: 42.42 | 31: iteration 20760/ 52452 | consumed samples: 10629120 | consumed tokens: 21768437760 | elapsed time per iteration (s): 1.81 | learning rate: 1.406E-04 | global batch size: 512 | lm loss: 1.389811E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.485 | TFLOPs: 42.40 | 31: iteration 20770/ 52452 | consumed samples: 10634240 | consumed tokens: 21778923520 | elapsed time per iteration (s): 1.75 | learning rate: 1.405E-04 | global batch size: 512 | lm loss: 1.365255E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.999 | TFLOPs: 43.98 | 31: iteration 20780/ 52452 | consumed samples: 10639360 | consumed tokens: 21789409280 | elapsed time per iteration (s): 1.77 | learning rate: 1.405E-04 | global batch size: 512 | lm loss: 1.374988E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.928 | TFLOPs: 43.37 | 31: iteration 20790/ 52452 | consumed samples: 10644480 | consumed tokens: 21799895040 | elapsed time per iteration (s): 1.75 | learning rate: 1.404E-04 | global batch size: 512 | lm loss: 1.387372E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.953 | TFLOPs: 43.82 | 31: iteration 20800/ 52452 | consumed samples: 10649600 | consumed tokens: 21810380800 | elapsed time per iteration (s): 1.90 | learning rate: 1.404E-04 | global batch size: 512 | lm loss: 1.383843E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.542 | TFLOPs: 40.46 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20800 | lm loss value: 4.130514E+00 | lm loss PPL: 6.220987E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20800 | lm loss value: 4.077050E+00 | lm loss PPL: 5.897126E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 20810/ 52452 | consumed samples: 10654720 | consumed tokens: 21820866560 | elapsed time per iteration (s): 9.19 | learning rate: 1.403E-04 | global batch size: 512 | lm loss: 1.371469E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.689 | TFLOPs: 8.36 | 31: iteration 20820/ 52452 | consumed samples: 10659840 | consumed tokens: 21831352320 | elapsed time per iteration (s): 1.79 | learning rate: 1.403E-04 | global batch size: 512 | lm loss: 1.391440E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.209 | TFLOPs: 42.96 | 31: iteration 20830/ 52452 | consumed samples: 10664960 | consumed tokens: 21841838080 | elapsed time per iteration (s): 1.77 | learning rate: 1.402E-04 | global batch size: 512 | lm loss: 1.376332E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.480 | TFLOPs: 43.30 | 31: iteration 20840/ 52452 | consumed samples: 10670080 | consumed tokens: 21852323840 | elapsed time per iteration (s): 1.75 | learning rate: 1.402E-04 | global batch size: 512 | lm loss: 1.386368E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.334 | TFLOPs: 43.88 | 31: iteration 20850/ 52452 | consumed samples: 10675200 | consumed tokens: 21862809600 | elapsed time per iteration (s): 1.76 | learning rate: 1.401E-04 | global batch size: 512 | lm loss: 1.382475E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.571 | TFLOPs: 43.61 | 31: iteration 20860/ 52452 | consumed samples: 10680320 | consumed tokens: 21873295360 | elapsed time per iteration (s): 1.75 | learning rate: 1.401E-04 | global batch size: 512 | lm loss: 1.379266E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.886 | TFLOPs: 43.96 | 31: iteration 20870/ 52452 | consumed samples: 10685440 | consumed tokens: 21883781120 | elapsed time per iteration (s): 1.77 | learning rate: 1.400E-04 | global batch size: 512 | lm loss: 1.371612E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.663 | TFLOPs: 43.48 | 31: iteration 20880/ 52452 | consumed samples: 10690560 | consumed tokens: 21894266880 | elapsed time per iteration (s): 1.80 | learning rate: 1.400E-04 | global batch size: 512 | lm loss: 1.385420E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.979 | TFLOPs: 42.77 | 31: iteration 20890/ 52452 | consumed samples: 10695680 | consumed tokens: 21904752640 | elapsed time per iteration (s): 1.83 | learning rate: 1.399E-04 | global batch size: 512 | lm loss: 1.370934E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.661 | TFLOPs: 41.98 | 31: iteration 20900/ 52452 | consumed samples: 10700800 | consumed tokens: 21915238400 | elapsed time per iteration (s): 1.80 | learning rate: 1.399E-04 | global batch size: 512 | lm loss: 1.368748E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.704 | TFLOPs: 42.58 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 20900 | lm loss value: 4.131628E+00 | lm loss PPL: 6.227923E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 20900 | lm loss value: 4.077166E+00 | lm loss PPL: 5.897807E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 20910/ 52452 | consumed samples: 10705920 | consumed tokens: 21925724160 | elapsed time per iteration (s): 9.18 | learning rate: 1.398E-04 | global batch size: 512 | lm loss: 1.377211E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.747 | TFLOPs: 8.37 | 31: iteration 20920/ 52452 | consumed samples: 10711040 | consumed tokens: 21936209920 | elapsed time per iteration (s): 1.76 | learning rate: 1.397E-04 | global batch size: 512 | lm loss: 1.375055E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.359 | TFLOPs: 43.73 | 31: iteration 20930/ 52452 | consumed samples: 10716160 | consumed tokens: 21946695680 | elapsed time per iteration (s): 1.77 | learning rate: 1.397E-04 | global batch size: 512 | lm loss: 1.379120E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.000 | TFLOPs: 43.53 | 31: iteration 20940/ 52452 | consumed samples: 10721280 | consumed tokens: 21957181440 | elapsed time per iteration (s): 1.78 | learning rate: 1.396E-04 | global batch size: 512 | lm loss: 1.381881E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.045 | TFLOPs: 43.08 | 31: iteration 20950/ 52452 | consumed samples: 10726400 | consumed tokens: 21967667200 | elapsed time per iteration (s): 1.76 | learning rate: 1.396E-04 | global batch size: 512 | lm loss: 1.373563E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.501 | TFLOPs: 43.75 | 31: iteration 20960/ 52452 | consumed samples: 10731520 | consumed tokens: 21978152960 | elapsed time per iteration (s): 1.76 | learning rate: 1.395E-04 | global batch size: 512 | lm loss: 1.370097E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.267 | TFLOPs: 43.57 | 31: iteration 20970/ 52452 | consumed samples: 10736640 | consumed tokens: 21988638720 | elapsed time per iteration (s): 1.81 | learning rate: 1.395E-04 | global batch size: 512 | lm loss: 1.389187E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.687 | TFLOPs: 42.43 | 31: iteration 20980/ 52452 | consumed samples: 10741760 | consumed tokens: 21999124480 | elapsed time per iteration (s): 1.74 | learning rate: 1.394E-04 | global batch size: 512 | lm loss: 1.372721E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.486 | TFLOPs: 44.20 | 31: iteration 20990/ 52452 | consumed samples: 10746880 | consumed tokens: 22009610240 | elapsed time per iteration (s): 1.77 | learning rate: 1.394E-04 | global batch size: 512 | lm loss: 1.373383E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.360 | TFLOPs: 43.43 | 31: iteration 21000/ 52452 | consumed samples: 10752000 | consumed tokens: 22020096000 | elapsed time per iteration (s): 1.80 | learning rate: 1.393E-04 | global batch size: 512 | lm loss: 1.382040E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.973 | TFLOPs: 42.77 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21000 | lm loss value: 4.191329E+00 | lm loss PPL: 6.611059E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21000 | lm loss value: 4.127972E+00 | lm loss PPL: 6.205193E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21010/ 52452 | consumed samples: 10757120 | consumed tokens: 22030581760 | elapsed time per iteration (s): 9.21 | learning rate: 1.393E-04 | global batch size: 512 | lm loss: 1.380599E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.605 | TFLOPs: 8.35 | 31: iteration 21020/ 52452 | consumed samples: 10762240 | consumed tokens: 22041067520 | elapsed time per iteration (s): 1.76 | learning rate: 1.392E-04 | global batch size: 512 | lm loss: 1.371335E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.031 | TFLOPs: 43.68 | 31: iteration 21030/ 52452 | consumed samples: 10767360 | consumed tokens: 22051553280 | elapsed time per iteration (s): 1.77 | learning rate: 1.392E-04 | global batch size: 512 | lm loss: 1.381798E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.698 | TFLOPs: 43.33 | 31: iteration 21040/ 52452 | consumed samples: 10772480 | consumed tokens: 22062039040 | elapsed time per iteration (s): 1.79 | learning rate: 1.391E-04 | global batch size: 512 | lm loss: 1.375860E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.270 | TFLOPs: 42.82 | 31: iteration 21050/ 52452 | consumed samples: 10777600 | consumed tokens: 22072524800 | elapsed time per iteration (s): 1.84 | learning rate: 1.391E-04 | global batch size: 512 | lm loss: 1.371230E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.171 | TFLOPs: 41.75 | 31: iteration 21060/ 52452 | consumed samples: 10782720 | consumed tokens: 22083010560 | elapsed time per iteration (s): 1.74 | learning rate: 1.390E-04 | global batch size: 512 | lm loss: 1.375990E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.107 | TFLOPs: 44.14 | 31: iteration 21070/ 52452 | consumed samples: 10787840 | consumed tokens: 22093496320 | elapsed time per iteration (s): 1.82 | learning rate: 1.390E-04 | global batch size: 512 | lm loss: 1.375534E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.788 | TFLOPs: 42.29 | 31: iteration 21080/ 52452 | consumed samples: 10792960 | consumed tokens: 22103982080 | elapsed time per iteration (s): 1.77 | learning rate: 1.389E-04 | global batch size: 512 | lm loss: 1.387354E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.237 | TFLOPs: 43.41 | 31: iteration 21090/ 52452 | consumed samples: 10798080 | consumed tokens: 22114467840 | elapsed time per iteration (s): 1.77 | learning rate: 1.389E-04 | global batch size: 512 | lm loss: 1.375673E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.860 | TFLOPs: 43.51 | 31: iteration 21100/ 52452 | consumed samples: 10803200 | consumed tokens: 22124953600 | elapsed time per iteration (s): 1.79 | learning rate: 1.388E-04 | global batch size: 512 | lm loss: 1.380154E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.600 | TFLOPs: 42.87 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21100 | lm loss value: 4.160568E+00 | lm loss PPL: 6.410794E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21100 | lm loss value: 4.101198E+00 | lm loss PPL: 6.041263E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21110/ 52452 | consumed samples: 10808320 | consumed tokens: 22135439360 | elapsed time per iteration (s): 9.16 | learning rate: 1.388E-04 | global batch size: 512 | lm loss: 1.385309E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.909 | TFLOPs: 8.39 | 31: iteration 21120/ 52452 | consumed samples: 10813440 | consumed tokens: 22145925120 | elapsed time per iteration (s): 1.78 | learning rate: 1.387E-04 | global batch size: 512 | lm loss: 1.370569E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.333 | TFLOPs: 43.13 | 31: iteration 21130/ 52452 | consumed samples: 10818560 | consumed tokens: 22156410880 | elapsed time per iteration (s): 1.82 | learning rate: 1.387E-04 | global batch size: 512 | lm loss: 1.372673E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.918 | TFLOPs: 42.16 | 31: iteration 21140/ 52452 | consumed samples: 10823680 | consumed tokens: 22166896640 | elapsed time per iteration (s): 1.76 | learning rate: 1.386E-04 | global batch size: 512 | lm loss: 1.372850E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.117 | TFLOPs: 43.54 | 31: iteration 21150/ 52452 | consumed samples: 10828800 | consumed tokens: 22177382400 | elapsed time per iteration (s): 1.79 | learning rate: 1.386E-04 | global batch size: 512 | lm loss: 1.377305E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.121 | TFLOPs: 42.95 | 31: iteration 21160/ 52452 | consumed samples: 10833920 | consumed tokens: 22187868160 | elapsed time per iteration (s): 1.82 | learning rate: 1.385E-04 | global batch size: 512 | lm loss: 1.384426E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.431 | TFLOPs: 42.24 | 31: iteration 21170/ 52452 | consumed samples: 10839040 | consumed tokens: 22198353920 | elapsed time per iteration (s): 1.83 | learning rate: 1.385E-04 | global batch size: 512 | lm loss: 1.379044E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.718 | TFLOPs: 41.98 | 31: iteration 21180/ 52452 | consumed samples: 10844160 | consumed tokens: 22208839680 | elapsed time per iteration (s): 1.76 | learning rate: 1.384E-04 | global batch size: 512 | lm loss: 1.389796E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.232 | TFLOPs: 43.71 | 31: iteration 21190/ 52452 | consumed samples: 10849280 | consumed tokens: 22219325440 | elapsed time per iteration (s): 1.76 | learning rate: 1.384E-04 | global batch size: 512 | lm loss: 1.372733E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.364 | TFLOPs: 43.73 | 31: iteration 21200/ 52452 | consumed samples: 10854400 | consumed tokens: 22229811200 | elapsed time per iteration (s): 1.80 | learning rate: 1.383E-04 | global batch size: 512 | lm loss: 1.366312E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.716 | TFLOPs: 42.73 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21200 | lm loss value: 4.204573E+00 | lm loss PPL: 6.699200E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21200 | lm loss value: 4.128729E+00 | lm loss PPL: 6.209894E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21210/ 52452 | consumed samples: 10859520 | consumed tokens: 22240296960 | elapsed time per iteration (s): 9.21 | learning rate: 1.383E-04 | global batch size: 512 | lm loss: 1.368539E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.582 | TFLOPs: 8.34 | 31: iteration 21220/ 52452 | consumed samples: 10864640 | consumed tokens: 22250782720 | elapsed time per iteration (s): 1.81 | learning rate: 1.382E-04 | global batch size: 512 | lm loss: 1.383556E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.473 | TFLOPs: 42.40 | 31: iteration 21230/ 52452 | consumed samples: 10869760 | consumed tokens: 22261268480 | elapsed time per iteration (s): 1.77 | learning rate: 1.382E-04 | global batch size: 512 | lm loss: 1.370709E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.875 | TFLOPs: 43.36 | 31: iteration 21240/ 52452 | consumed samples: 10874880 | consumed tokens: 22271754240 | elapsed time per iteration (s): 1.76 | learning rate: 1.381E-04 | global batch size: 512 | lm loss: 1.368154E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.623 | TFLOPs: 43.77 | 31: iteration 21250/ 52452 | consumed samples: 10880000 | consumed tokens: 22282240000 | elapsed time per iteration (s): 1.76 | learning rate: 1.380E-04 | global batch size: 512 | lm loss: 1.380512E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.610 | TFLOPs: 43.77 | 31: iteration 21260/ 52452 | consumed samples: 10885120 | consumed tokens: 22292725760 | elapsed time per iteration (s): 1.75 | learning rate: 1.380E-04 | global batch size: 512 | lm loss: 1.370684E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.345 | TFLOPs: 43.88 | 31: iteration 21270/ 52452 | consumed samples: 10890240 | consumed tokens: 22303211520 | elapsed time per iteration (s): 1.80 | learning rate: 1.379E-04 | global batch size: 512 | lm loss: 1.376814E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.183 | TFLOPs: 42.80 | 31: iteration 21280/ 52452 | consumed samples: 10895360 | consumed tokens: 22313697280 | elapsed time per iteration (s): 1.79 | learning rate: 1.379E-04 | global batch size: 512 | lm loss: 1.370910E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.364 | TFLOPs: 42.98 | 31: iteration 21290/ 52452 | consumed samples: 10900480 | consumed tokens: 22324183040 | elapsed time per iteration (s): 1.76 | learning rate: 1.378E-04 | global batch size: 512 | lm loss: 1.376227E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.428 | TFLOPs: 43.74 | 31: iteration 21300/ 52452 | consumed samples: 10905600 | consumed tokens: 22334668800 | elapsed time per iteration (s): 1.82 | learning rate: 1.378E-04 | global batch size: 512 | lm loss: 1.368257E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.707 | TFLOPs: 42.13 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21300 | lm loss value: 4.049829E+00 | lm loss PPL: 5.738762E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21300 | lm loss value: 4.009635E+00 | lm loss PPL: 5.512674E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21310/ 52452 | consumed samples: 10910720 | consumed tokens: 22345154560 | elapsed time per iteration (s): 9.20 | learning rate: 1.377E-04 | global batch size: 512 | lm loss: 1.374357E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.626 | TFLOPs: 8.35 | 31: iteration 21320/ 52452 | consumed samples: 10915840 | consumed tokens: 22355640320 | elapsed time per iteration (s): 1.82 | learning rate: 1.377E-04 | global batch size: 512 | lm loss: 1.383456E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.620 | TFLOPs: 42.27 | 31: iteration 21330/ 52452 | consumed samples: 10920960 | consumed tokens: 22366126080 | elapsed time per iteration (s): 1.75 | learning rate: 1.376E-04 | global batch size: 512 | lm loss: 1.362254E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.463 | TFLOPs: 43.90 | 31: iteration 21340/ 52452 | consumed samples: 10926080 | consumed tokens: 22376611840 | elapsed time per iteration (s): 1.78 | learning rate: 1.376E-04 | global batch size: 512 | lm loss: 1.383622E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.593 | TFLOPs: 43.17 | 31: iteration 21350/ 52452 | consumed samples: 10931200 | consumed tokens: 22387097600 | elapsed time per iteration (s): 1.75 | learning rate: 1.375E-04 | global batch size: 512 | lm loss: 1.379433E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.929 | TFLOPs: 43.82 | 31: iteration 21360/ 52452 | consumed samples: 10936320 | consumed tokens: 22397583360 | elapsed time per iteration (s): 1.79 | learning rate: 1.375E-04 | global batch size: 512 | lm loss: 1.380178E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.305 | TFLOPs: 42.82 | 31: iteration 21370/ 52452 | consumed samples: 10941440 | consumed tokens: 22408069120 | elapsed time per iteration (s): 1.81 | learning rate: 1.374E-04 | global batch size: 512 | lm loss: 1.378721E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.482 | TFLOPs: 42.55 | 31: iteration 21380/ 52452 | consumed samples: 10946560 | consumed tokens: 22418554880 | elapsed time per iteration (s): 1.76 | learning rate: 1.374E-04 | global batch size: 512 | lm loss: 1.367354E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.378 | TFLOPs: 43.58 | 31: iteration 21390/ 52452 | consumed samples: 10951680 | consumed tokens: 22429040640 | elapsed time per iteration (s): 1.78 | learning rate: 1.373E-04 | global batch size: 512 | lm loss: 1.383265E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.136 | TFLOPs: 43.10 | 31: iteration 21400/ 52452 | consumed samples: 10956800 | consumed tokens: 22439526400 | elapsed time per iteration (s): 1.80 | learning rate: 1.373E-04 | global batch size: 512 | lm loss: 1.373514E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.214 | TFLOPs: 42.66 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21400 | lm loss value: 4.258382E+00 | lm loss PPL: 7.069553E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21400 | lm loss value: 4.162382E+00 | lm loss PPL: 6.422430E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21410/ 52452 | consumed samples: 10961920 | consumed tokens: 22450012160 | elapsed time per iteration (s): 9.24 | learning rate: 1.372E-04 | global batch size: 512 | lm loss: 1.369144E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.386 | TFLOPs: 8.31 | 31: iteration 21420/ 52452 | consumed samples: 10967040 | consumed tokens: 22460497920 | elapsed time per iteration (s): 1.77 | learning rate: 1.372E-04 | global batch size: 512 | lm loss: 1.380750E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.413 | TFLOPs: 43.44 | 31: iteration 21430/ 52452 | consumed samples: 10972160 | consumed tokens: 22470983680 | elapsed time per iteration (s): 1.80 | learning rate: 1.371E-04 | global batch size: 512 | lm loss: 1.369175E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.220 | TFLOPs: 42.81 | 31: iteration 21440/ 52452 | consumed samples: 10977280 | consumed tokens: 22481469440 | elapsed time per iteration (s): 1.80 | learning rate: 1.371E-04 | global batch size: 512 | lm loss: 1.376365E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.385 | TFLOPs: 42.68 | 31: iteration 21450/ 52452 | consumed samples: 10982400 | consumed tokens: 22491955200 | elapsed time per iteration (s): 1.74 | learning rate: 1.370E-04 | global batch size: 512 | lm loss: 1.384705E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.044 | TFLOPs: 44.13 | 31: iteration 21460/ 52452 | consumed samples: 10987520 | consumed tokens: 22502440960 | elapsed time per iteration (s): 1.77 | learning rate: 1.370E-04 | global batch size: 512 | lm loss: 1.359670E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.570 | TFLOPs: 43.31 | 31: iteration 21470/ 52452 | consumed samples: 10992640 | consumed tokens: 22512926720 | elapsed time per iteration (s): 1.74 | learning rate: 1.369E-04 | global batch size: 512 | lm loss: 1.365836E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.621 | TFLOPs: 44.07 | 31: iteration 21480/ 52452 | consumed samples: 10997760 | consumed tokens: 22523412480 | elapsed time per iteration (s): 1.79 | learning rate: 1.369E-04 | global batch size: 512 | lm loss: 1.376418E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.916 | TFLOPs: 42.91 | 31: iteration 21490/ 52452 | consumed samples: 11002880 | consumed tokens: 22533898240 | elapsed time per iteration (s): 1.83 | learning rate: 1.368E-04 | global batch size: 512 | lm loss: 1.381587E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.065 | TFLOPs: 42.04 | 31: iteration 21500/ 52452 | consumed samples: 11008000 | consumed tokens: 22544384000 | elapsed time per iteration (s): 1.82 | learning rate: 1.368E-04 | global batch size: 512 | lm loss: 1.351439E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.124 | TFLOPs: 42.20 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21500 | lm loss value: 4.119561E+00 | lm loss PPL: 6.153224E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21500 | lm loss value: 4.058949E+00 | lm loss PPL: 5.791344E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21510/ 52452 | consumed samples: 11013120 | consumed tokens: 22554869760 | elapsed time per iteration (s): 9.18 | learning rate: 1.367E-04 | global batch size: 512 | lm loss: 1.361783E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.751 | TFLOPs: 8.37 | 31: iteration 21520/ 52452 | consumed samples: 11018240 | consumed tokens: 22565355520 | elapsed time per iteration (s): 1.75 | learning rate: 1.366E-04 | global batch size: 512 | lm loss: 1.379628E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.181 | TFLOPs: 43.85 | 31: iteration 21530/ 52452 | consumed samples: 11023360 | consumed tokens: 22575841280 | elapsed time per iteration (s): 1.78 | learning rate: 1.366E-04 | global batch size: 512 | lm loss: 1.381925E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.409 | TFLOPs: 43.29 | 31: iteration 21540/ 52452 | consumed samples: 11028480 | consumed tokens: 22586327040 | elapsed time per iteration (s): 1.74 | learning rate: 1.365E-04 | global batch size: 512 | lm loss: 1.362457E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.943 | TFLOPs: 44.12 | 31: iteration 21550/ 52452 | consumed samples: 11033600 | consumed tokens: 22596812800 | elapsed time per iteration (s): 1.81 | learning rate: 1.365E-04 | global batch size: 512 | lm loss: 1.384467E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.883 | TFLOPs: 42.46 | 31: iteration 21560/ 52452 | consumed samples: 11038720 | consumed tokens: 22607298560 | elapsed time per iteration (s): 1.74 | learning rate: 1.364E-04 | global batch size: 512 | lm loss: 1.369822E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.456 | TFLOPs: 44.05 | 31: iteration 21570/ 52452 | consumed samples: 11043840 | consumed tokens: 22617784320 | elapsed time per iteration (s): 1.77 | learning rate: 1.364E-04 | global batch size: 512 | lm loss: 1.381904E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.997 | TFLOPs: 43.38 | 31: iteration 21580/ 52452 | consumed samples: 11048960 | consumed tokens: 22628270080 | elapsed time per iteration (s): 1.77 | learning rate: 1.363E-04 | global batch size: 512 | lm loss: 1.374626E+00 | grad norm: 0.147 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.540 | TFLOPs: 43.31 | 31: iteration 21590/ 52452 | consumed samples: 11054080 | consumed tokens: 22638755840 | elapsed time per iteration (s): 1.98 | learning rate: 1.363E-04 | global batch size: 512 | lm loss: 1.362221E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 259.142 | TFLOPs: 38.90 | 31: iteration 21600/ 52452 | consumed samples: 11059200 | consumed tokens: 22649241600 | elapsed time per iteration (s): 1.83 | learning rate: 1.362E-04 | global batch size: 512 | lm loss: 1.376125E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.313 | TFLOPs: 42.07 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21600 | lm loss value: 4.107525E+00 | lm loss PPL: 6.079608E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21600 | lm loss value: 4.046285E+00 | lm loss PPL: 5.718463E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21610/ 52452 | consumed samples: 11064320 | consumed tokens: 22659727360 | elapsed time per iteration (s): 9.19 | learning rate: 1.362E-04 | global batch size: 512 | lm loss: 1.364196E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.703 | TFLOPs: 8.36 | 31: iteration 21620/ 52452 | consumed samples: 11069440 | consumed tokens: 22670213120 | elapsed time per iteration (s): 1.75 | learning rate: 1.361E-04 | global batch size: 512 | lm loss: 1.365841E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.320 | TFLOPs: 43.88 | 31: iteration 21630/ 52452 | consumed samples: 11074560 | consumed tokens: 22680698880 | elapsed time per iteration (s): 1.76 | learning rate: 1.361E-04 | global batch size: 512 | lm loss: 1.377660E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.139 | TFLOPs: 43.55 | 31: iteration 21640/ 52452 | consumed samples: 11079680 | consumed tokens: 22691184640 | elapsed time per iteration (s): 1.85 | learning rate: 1.360E-04 | global batch size: 512 | lm loss: 1.371377E+00 | grad norm: 0.147 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.239 | TFLOPs: 41.61 | 31: iteration 21650/ 52452 | consumed samples: 11084800 | consumed tokens: 22701670400 | elapsed time per iteration (s): 1.82 | learning rate: 1.360E-04 | global batch size: 512 | lm loss: 1.375731E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.634 | TFLOPs: 42.27 | 31: iteration 21660/ 52452 | consumed samples: 11089920 | consumed tokens: 22712156160 | elapsed time per iteration (s): 1.73 | learning rate: 1.359E-04 | global batch size: 512 | lm loss: 1.382007E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.351 | TFLOPs: 44.48 | 31: iteration 21670/ 52452 | consumed samples: 11095040 | consumed tokens: 22722641920 | elapsed time per iteration (s): 1.78 | learning rate: 1.359E-04 | global batch size: 512 | lm loss: 1.377876E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.051 | TFLOPs: 43.08 | 31: iteration 21680/ 52452 | consumed samples: 11100160 | consumed tokens: 22733127680 | elapsed time per iteration (s): 1.93 | learning rate: 1.358E-04 | global batch size: 512 | lm loss: 1.383882E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.228 | TFLOPs: 39.81 | 31: iteration 21690/ 52452 | consumed samples: 11105280 | consumed tokens: 22743613440 | elapsed time per iteration (s): 1.75 | learning rate: 1.358E-04 | global batch size: 512 | lm loss: 1.362276E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.606 | TFLOPs: 43.92 | 31: iteration 21700/ 52452 | consumed samples: 11110400 | consumed tokens: 22754099200 | elapsed time per iteration (s): 1.75 | learning rate: 1.357E-04 | global batch size: 512 | lm loss: 1.367659E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.790 | TFLOPs: 43.95 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21700 | lm loss value: 4.113703E+00 | lm loss PPL: 6.117281E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21700 | lm loss value: 4.055323E+00 | lm loss PPL: 5.770381E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21710/ 52452 | consumed samples: 11115520 | consumed tokens: 22764584960 | elapsed time per iteration (s): 9.11 | learning rate: 1.357E-04 | global batch size: 512 | lm loss: 1.360445E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.174 | TFLOPs: 8.43 | 31: iteration 21720/ 52452 | consumed samples: 11120640 | consumed tokens: 22775070720 | elapsed time per iteration (s): 1.77 | learning rate: 1.356E-04 | global batch size: 512 | lm loss: 1.382607E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.994 | TFLOPs: 43.53 | 31: iteration 21730/ 52452 | consumed samples: 11125760 | consumed tokens: 22785556480 | elapsed time per iteration (s): 1.77 | learning rate: 1.356E-04 | global batch size: 512 | lm loss: 1.370256E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.944 | TFLOPs: 43.52 | 31: iteration 21740/ 52452 | consumed samples: 11130880 | consumed tokens: 22796042240 | elapsed time per iteration (s): 1.77 | learning rate: 1.355E-04 | global batch size: 512 | lm loss: 1.372579E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.806 | TFLOPs: 43.50 | 31: iteration 21750/ 52452 | consumed samples: 11136000 | consumed tokens: 22806528000 | elapsed time per iteration (s): 1.76 | learning rate: 1.354E-04 | global batch size: 512 | lm loss: 1.364440E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.470 | TFLOPs: 43.75 | 31: iteration 21760/ 52452 | consumed samples: 11141120 | consumed tokens: 22817013760 | elapsed time per iteration (s): 1.78 | learning rate: 1.354E-04 | global batch size: 512 | lm loss: 1.371845E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.109 | TFLOPs: 43.24 | 31: iteration 21770/ 52452 | consumed samples: 11146240 | consumed tokens: 22827499520 | elapsed time per iteration (s): 1.81 | learning rate: 1.353E-04 | global batch size: 512 | lm loss: 1.371392E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.579 | TFLOPs: 42.56 | 31: iteration 21780/ 52452 | consumed samples: 11151360 | consumed tokens: 22837985280 | elapsed time per iteration (s): 1.76 | learning rate: 1.353E-04 | global batch size: 512 | lm loss: 1.373419E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.046 | TFLOPs: 43.68 | 31: iteration 21790/ 52452 | consumed samples: 11156480 | consumed tokens: 22848471040 | elapsed time per iteration (s): 1.75 | learning rate: 1.352E-04 | global batch size: 512 | lm loss: 1.369021E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.869 | TFLOPs: 43.81 | 31: iteration 21800/ 52452 | consumed samples: 11161600 | consumed tokens: 22858956800 | elapsed time per iteration (s): 1.79 | learning rate: 1.352E-04 | global batch size: 512 | lm loss: 1.373208E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.072 | TFLOPs: 42.94 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21800 | lm loss value: 4.131706E+00 | lm loss PPL: 6.228410E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21800 | lm loss value: 4.077406E+00 | lm loss PPL: 5.899227E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21810/ 52452 | consumed samples: 11166720 | consumed tokens: 22869442560 | elapsed time per iteration (s): 9.18 | learning rate: 1.351E-04 | global batch size: 512 | lm loss: 1.365551E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.789 | TFLOPs: 8.37 | 31: iteration 21820/ 52452 | consumed samples: 11171840 | consumed tokens: 22879928320 | elapsed time per iteration (s): 1.77 | learning rate: 1.351E-04 | global batch size: 512 | lm loss: 1.366821E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.124 | TFLOPs: 43.40 | 31: iteration 21830/ 52452 | consumed samples: 11176960 | consumed tokens: 22890414080 | elapsed time per iteration (s): 1.77 | learning rate: 1.350E-04 | global batch size: 512 | lm loss: 1.365635E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.975 | TFLOPs: 43.52 | 31: iteration 21840/ 52452 | consumed samples: 11182080 | consumed tokens: 22900899840 | elapsed time per iteration (s): 1.76 | learning rate: 1.350E-04 | global batch size: 512 | lm loss: 1.376744E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.359 | TFLOPs: 43.73 | 31: iteration 21850/ 52452 | consumed samples: 11187200 | consumed tokens: 22911385600 | elapsed time per iteration (s): 1.89 | learning rate: 1.349E-04 | global batch size: 512 | lm loss: 1.364105E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.847 | TFLOPs: 40.65 | 31: iteration 21860/ 52452 | consumed samples: 11192320 | consumed tokens: 22921871360 | elapsed time per iteration (s): 1.93 | learning rate: 1.349E-04 | global batch size: 512 | lm loss: 1.360581E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.335 | TFLOPs: 39.83 | 31: iteration 21870/ 52452 | consumed samples: 11197440 | consumed tokens: 22932357120 | elapsed time per iteration (s): 1.91 | learning rate: 1.348E-04 | global batch size: 512 | lm loss: 1.375147E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.403 | TFLOPs: 40.14 | 31: iteration 21880/ 52452 | consumed samples: 11202560 | consumed tokens: 22942842880 | elapsed time per iteration (s): 1.81 | learning rate: 1.348E-04 | global batch size: 512 | lm loss: 1.358274E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.559 | TFLOPs: 42.41 | 31: iteration 21890/ 52452 | consumed samples: 11207680 | consumed tokens: 22953328640 | elapsed time per iteration (s): 1.88 | learning rate: 1.347E-04 | global batch size: 512 | lm loss: 1.373771E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.066 | TFLOPs: 40.84 | 31: iteration 21900/ 52452 | consumed samples: 11212800 | consumed tokens: 22963814400 | elapsed time per iteration (s): 1.80 | learning rate: 1.347E-04 | global batch size: 512 | lm loss: 1.365601E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.497 | TFLOPs: 42.70 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 21900 | lm loss value: 4.195755E+00 | lm loss PPL: 6.640382E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 21900 | lm loss value: 4.125864E+00 | lm loss PPL: 6.192126E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 21910/ 52452 | consumed samples: 11217920 | consumed tokens: 22974300160 | elapsed time per iteration (s): 9.19 | learning rate: 1.346E-04 | global batch size: 512 | lm loss: 1.356774E+00 | grad norm: 0.144 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.699 | TFLOPs: 8.36 | 31: iteration 21920/ 52452 | consumed samples: 11223040 | consumed tokens: 22984785920 | elapsed time per iteration (s): 1.74 | learning rate: 1.346E-04 | global batch size: 512 | lm loss: 1.364904E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.676 | TFLOPs: 44.08 | 31: iteration 21930/ 52452 | consumed samples: 11228160 | consumed tokens: 22995271680 | elapsed time per iteration (s): 1.77 | learning rate: 1.345E-04 | global batch size: 512 | lm loss: 1.373322E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.653 | TFLOPs: 43.33 | 31: iteration 21940/ 52452 | consumed samples: 11233280 | consumed tokens: 23005757440 | elapsed time per iteration (s): 1.78 | learning rate: 1.345E-04 | global batch size: 512 | lm loss: 1.368656E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.966 | TFLOPs: 43.22 | 31: iteration 21950/ 52452 | consumed samples: 11238400 | consumed tokens: 23016243200 | elapsed time per iteration (s): 1.73 | learning rate: 1.344E-04 | global batch size: 512 | lm loss: 1.367115E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.298 | TFLOPs: 44.32 | 31: iteration 21960/ 52452 | consumed samples: 11243520 | consumed tokens: 23026728960 | elapsed time per iteration (s): 1.84 | learning rate: 1.343E-04 | global batch size: 512 | lm loss: 1.358135E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.548 | TFLOPs: 41.66 | 31: iteration 21970/ 52452 | consumed samples: 11248640 | consumed tokens: 23037214720 | elapsed time per iteration (s): 1.83 | learning rate: 1.343E-04 | global batch size: 512 | lm loss: 1.364218E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.383 | TFLOPs: 41.93 | 31: iteration 21980/ 52452 | consumed samples: 11253760 | consumed tokens: 23047700480 | elapsed time per iteration (s): 1.79 | learning rate: 1.342E-04 | global batch size: 512 | lm loss: 1.366363E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.947 | TFLOPs: 42.92 | 31: iteration 21990/ 52452 | consumed samples: 11258880 | consumed tokens: 23058186240 | elapsed time per iteration (s): 1.86 | learning rate: 1.342E-04 | global batch size: 512 | lm loss: 1.370682E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.788 | TFLOPs: 41.39 | 0: [2022-12-26 00:30:09,986] [INFO] [logging.py:68:log_dist] [Rank 0] step=22000, skipped=0, lr=[0.000134139402222714, 0.000134139402222714, 0.000134139402222714], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 22000/ 52452 | consumed samples: 11264000 | consumed tokens: 23068672000 | elapsed time per iteration (s): 1.77 | learning rate: 1.341E-04 | global batch size: 512 | lm loss: 1.368864E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.755 | TFLOPs: 43.49 | 0: steps: 22000 loss: 1.4145 iter time (s): 1.791 samples/sec: 285.797 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22000 | lm loss value: 3.991511E+00 | lm loss PPL: 5.413662E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22000 | lm loss value: 3.996226E+00 | lm loss PPL: 5.439249E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22010/ 52452 | consumed samples: 11269120 | consumed tokens: 23079157760 | elapsed time per iteration (s): 9.18 | learning rate: 1.341E-04 | global batch size: 512 | lm loss: 1.372619E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.795 | TFLOPs: 8.37 | 31: iteration 22020/ 52452 | consumed samples: 11274240 | consumed tokens: 23089643520 | elapsed time per iteration (s): 1.79 | learning rate: 1.340E-04 | global batch size: 512 | lm loss: 1.374691E+00 | grad norm: 0.185 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.341 | TFLOPs: 42.98 | 31: iteration 22030/ 52452 | consumed samples: 11279360 | consumed tokens: 23100129280 | elapsed time per iteration (s): 1.77 | learning rate: 1.340E-04 | global batch size: 512 | lm loss: 1.369069E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.645 | TFLOPs: 43.47 | 31: iteration 22040/ 52452 | consumed samples: 11284480 | consumed tokens: 23110615040 | elapsed time per iteration (s): 1.81 | learning rate: 1.339E-04 | global batch size: 512 | lm loss: 1.372301E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.516 | TFLOPs: 42.55 | 31: iteration 22050/ 52452 | consumed samples: 11289600 | consumed tokens: 23121100800 | elapsed time per iteration (s): 1.77 | learning rate: 1.339E-04 | global batch size: 512 | lm loss: 1.372152E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.705 | TFLOPs: 43.33 | 31: iteration 22060/ 52452 | consumed samples: 11294720 | consumed tokens: 23131586560 | elapsed time per iteration (s): 1.81 | learning rate: 1.338E-04 | global batch size: 512 | lm loss: 1.382392E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.156 | TFLOPs: 42.50 | 31: iteration 22070/ 52452 | consumed samples: 11299840 | consumed tokens: 23142072320 | elapsed time per iteration (s): 1.75 | learning rate: 1.338E-04 | global batch size: 512 | lm loss: 1.370983E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.779 | TFLOPs: 43.94 | 31: iteration 22080/ 52452 | consumed samples: 11304960 | consumed tokens: 23152558080 | elapsed time per iteration (s): 1.79 | learning rate: 1.337E-04 | global batch size: 512 | lm loss: 1.365106E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.755 | TFLOPs: 43.04 | 31: iteration 22090/ 52452 | consumed samples: 11310080 | consumed tokens: 23163043840 | elapsed time per iteration (s): 1.77 | learning rate: 1.337E-04 | global batch size: 512 | lm loss: 1.376562E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.481 | TFLOPs: 43.30 | 31: iteration 22100/ 52452 | consumed samples: 11315200 | consumed tokens: 23173529600 | elapsed time per iteration (s): 1.82 | learning rate: 1.336E-04 | global batch size: 512 | lm loss: 1.357660E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.727 | TFLOPs: 42.14 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22100 | lm loss value: 4.160878E+00 | lm loss PPL: 6.412778E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22100 | lm loss value: 4.098159E+00 | lm loss PPL: 6.022929E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22110/ 52452 | consumed samples: 11320320 | consumed tokens: 23184015360 | elapsed time per iteration (s): 9.19 | learning rate: 1.336E-04 | global batch size: 512 | lm loss: 1.360320E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.728 | TFLOPs: 8.36 | 31: iteration 22120/ 52452 | consumed samples: 11325440 | consumed tokens: 23194501120 | elapsed time per iteration (s): 1.74 | learning rate: 1.335E-04 | global batch size: 512 | lm loss: 1.374711E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.010 | TFLOPs: 44.13 | 31: iteration 22130/ 52452 | consumed samples: 11330560 | consumed tokens: 23204986880 | elapsed time per iteration (s): 1.78 | learning rate: 1.335E-04 | global batch size: 512 | lm loss: 1.360711E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.233 | TFLOPs: 43.26 | 31: iteration 22140/ 52452 | consumed samples: 11335680 | consumed tokens: 23215472640 | elapsed time per iteration (s): 1.79 | learning rate: 1.334E-04 | global batch size: 512 | lm loss: 1.363561E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.606 | TFLOPs: 43.02 | 31: iteration 22150/ 52452 | consumed samples: 11340800 | consumed tokens: 23225958400 | elapsed time per iteration (s): 1.80 | learning rate: 1.334E-04 | global batch size: 512 | lm loss: 1.367787E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.064 | TFLOPs: 42.79 | 31: iteration 22160/ 52452 | consumed samples: 11345920 | consumed tokens: 23236444160 | elapsed time per iteration (s): 1.85 | learning rate: 1.333E-04 | global batch size: 512 | lm loss: 1.365219E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.142 | TFLOPs: 41.60 | 31: iteration 22170/ 52452 | consumed samples: 11351040 | consumed tokens: 23246929920 | elapsed time per iteration (s): 1.78 | learning rate: 1.332E-04 | global batch size: 512 | lm loss: 1.357431E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.131 | TFLOPs: 43.25 | 31: iteration 22180/ 52452 | consumed samples: 11356160 | consumed tokens: 23257415680 | elapsed time per iteration (s): 1.77 | learning rate: 1.332E-04 | global batch size: 512 | lm loss: 1.361853E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.186 | TFLOPs: 43.41 | 31: iteration 22190/ 52452 | consumed samples: 11361280 | consumed tokens: 23267901440 | elapsed time per iteration (s): 1.75 | learning rate: 1.331E-04 | global batch size: 512 | lm loss: 1.356965E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.835 | TFLOPs: 43.95 | 31: iteration 22200/ 52452 | consumed samples: 11366400 | consumed tokens: 23278387200 | elapsed time per iteration (s): 1.82 | learning rate: 1.331E-04 | global batch size: 512 | lm loss: 1.357225E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.393 | TFLOPs: 42.24 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22200 | lm loss value: 4.063395E+00 | lm loss PPL: 5.817144E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22200 | lm loss value: 3.997973E+00 | lm loss PPL: 5.448759E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22210/ 52452 | consumed samples: 11371520 | consumed tokens: 23288872960 | elapsed time per iteration (s): 9.18 | learning rate: 1.330E-04 | global batch size: 512 | lm loss: 1.362963E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.778 | TFLOPs: 8.37 | 31: iteration 22220/ 52452 | consumed samples: 11376640 | consumed tokens: 23299358720 | elapsed time per iteration (s): 1.77 | learning rate: 1.330E-04 | global batch size: 512 | lm loss: 1.355866E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.720 | TFLOPs: 43.49 | 31: iteration 22230/ 52452 | consumed samples: 11381760 | consumed tokens: 23309844480 | elapsed time per iteration (s): 1.75 | learning rate: 1.329E-04 | global batch size: 512 | lm loss: 1.369120E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.069 | TFLOPs: 43.99 | 31: iteration 22240/ 52452 | consumed samples: 11386880 | consumed tokens: 23320330240 | elapsed time per iteration (s): 1.76 | learning rate: 1.329E-04 | global batch size: 512 | lm loss: 1.350794E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.205 | TFLOPs: 43.71 | 31: iteration 22250/ 52452 | consumed samples: 11392000 | consumed tokens: 23330816000 | elapsed time per iteration (s): 1.74 | learning rate: 1.328E-04 | global batch size: 512 | lm loss: 1.369883E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.557 | TFLOPs: 44.21 | 31: iteration 22260/ 52452 | consumed samples: 11397120 | consumed tokens: 23341301760 | elapsed time per iteration (s): 1.73 | learning rate: 1.328E-04 | global batch size: 512 | lm loss: 1.355481E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.181 | TFLOPs: 44.31 | 31: iteration 22270/ 52452 | consumed samples: 11402240 | consumed tokens: 23351787520 | elapsed time per iteration (s): 1.79 | learning rate: 1.327E-04 | global batch size: 512 | lm loss: 1.357449E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.782 | TFLOPs: 43.04 | 31: iteration 22280/ 52452 | consumed samples: 11407360 | consumed tokens: 23362273280 | elapsed time per iteration (s): 1.74 | learning rate: 1.327E-04 | global batch size: 512 | lm loss: 1.369806E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.314 | TFLOPs: 44.17 | 31: iteration 22290/ 52452 | consumed samples: 11412480 | consumed tokens: 23372759040 | elapsed time per iteration (s): 1.74 | learning rate: 1.326E-04 | global batch size: 512 | lm loss: 1.367868E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.015 | TFLOPs: 44.28 | 31: iteration 22300/ 52452 | consumed samples: 11417600 | consumed tokens: 23383244800 | elapsed time per iteration (s): 1.76 | learning rate: 1.326E-04 | global batch size: 512 | lm loss: 1.359280E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.746 | TFLOPs: 43.64 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22300 | lm loss value: 4.065682E+00 | lm loss PPL: 5.830468E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22300 | lm loss value: 4.017487E+00 | lm loss PPL: 5.556128E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22310/ 52452 | consumed samples: 11422720 | consumed tokens: 23393730560 | elapsed time per iteration (s): 9.19 | learning rate: 1.325E-04 | global batch size: 512 | lm loss: 1.358998E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.689 | TFLOPs: 8.36 | 31: iteration 22320/ 52452 | consumed samples: 11427840 | consumed tokens: 23404216320 | elapsed time per iteration (s): 1.73 | learning rate: 1.325E-04 | global batch size: 512 | lm loss: 1.369337E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.923 | TFLOPs: 44.42 | 31: iteration 22330/ 52452 | consumed samples: 11432960 | consumed tokens: 23414702080 | elapsed time per iteration (s): 1.74 | learning rate: 1.324E-04 | global batch size: 512 | lm loss: 1.357300E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.480 | TFLOPs: 44.05 | 31: iteration 22340/ 52452 | consumed samples: 11438080 | consumed tokens: 23425187840 | elapsed time per iteration (s): 1.76 | learning rate: 1.324E-04 | global batch size: 512 | lm loss: 1.360577E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.112 | TFLOPs: 43.54 | 31: iteration 22350/ 52452 | consumed samples: 11443200 | consumed tokens: 23435673600 | elapsed time per iteration (s): 1.74 | learning rate: 1.323E-04 | global batch size: 512 | lm loss: 1.367213E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.878 | TFLOPs: 44.26 | 31: iteration 22360/ 52452 | consumed samples: 11448320 | consumed tokens: 23446159360 | elapsed time per iteration (s): 1.74 | learning rate: 1.322E-04 | global batch size: 512 | lm loss: 1.359207E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.800 | TFLOPs: 44.25 | 31: iteration 22370/ 52452 | consumed samples: 11453440 | consumed tokens: 23456645120 | elapsed time per iteration (s): 1.82 | learning rate: 1.322E-04 | global batch size: 512 | lm loss: 1.362238E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.983 | TFLOPs: 42.17 | 31: iteration 22380/ 52452 | consumed samples: 11458560 | consumed tokens: 23467130880 | elapsed time per iteration (s): 1.74 | learning rate: 1.321E-04 | global batch size: 512 | lm loss: 1.363268E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.088 | TFLOPs: 44.14 | 31: iteration 22390/ 52452 | consumed samples: 11463680 | consumed tokens: 23477616640 | elapsed time per iteration (s): 1.84 | learning rate: 1.321E-04 | global batch size: 512 | lm loss: 1.353632E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.960 | TFLOPs: 41.87 | 31: iteration 22400/ 52452 | consumed samples: 11468800 | consumed tokens: 23488102400 | elapsed time per iteration (s): 1.79 | learning rate: 1.320E-04 | global batch size: 512 | lm loss: 1.351816E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.527 | TFLOPs: 43.01 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22400 | lm loss value: 4.081028E+00 | lm loss PPL: 5.920633E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22400 | lm loss value: 4.038769E+00 | lm loss PPL: 5.675645E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22410/ 52452 | consumed samples: 11473920 | consumed tokens: 23498588160 | elapsed time per iteration (s): 9.09 | learning rate: 1.320E-04 | global batch size: 512 | lm loss: 1.355033E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.324 | TFLOPs: 8.45 | 31: iteration 22420/ 52452 | consumed samples: 11479040 | consumed tokens: 23509073920 | elapsed time per iteration (s): 1.74 | learning rate: 1.319E-04 | global batch size: 512 | lm loss: 1.367088E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.532 | TFLOPs: 44.06 | 31: iteration 22430/ 52452 | consumed samples: 11484160 | consumed tokens: 23519559680 | elapsed time per iteration (s): 1.76 | learning rate: 1.319E-04 | global batch size: 512 | lm loss: 1.351698E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.143 | TFLOPs: 43.55 | 31: iteration 22440/ 52452 | consumed samples: 11489280 | consumed tokens: 23530045440 | elapsed time per iteration (s): 1.75 | learning rate: 1.318E-04 | global batch size: 512 | lm loss: 1.362182E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.832 | TFLOPs: 43.95 | 31: iteration 22450/ 52452 | consumed samples: 11494400 | consumed tokens: 23540531200 | elapsed time per iteration (s): 1.74 | learning rate: 1.318E-04 | global batch size: 512 | lm loss: 1.365277E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.341 | TFLOPs: 44.18 | 31: iteration 22460/ 52452 | consumed samples: 11499520 | consumed tokens: 23551016960 | elapsed time per iteration (s): 1.75 | learning rate: 1.317E-04 | global batch size: 512 | lm loss: 1.357048E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.725 | TFLOPs: 43.94 | 31: iteration 22470/ 52452 | consumed samples: 11504640 | consumed tokens: 23561502720 | elapsed time per iteration (s): 1.73 | learning rate: 1.317E-04 | global batch size: 512 | lm loss: 1.362061E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.259 | TFLOPs: 44.47 | 31: iteration 22480/ 52452 | consumed samples: 11509760 | consumed tokens: 23571988480 | elapsed time per iteration (s): 1.75 | learning rate: 1.316E-04 | global batch size: 512 | lm loss: 1.362801E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.669 | TFLOPs: 43.93 | 31: iteration 22490/ 52452 | consumed samples: 11514880 | consumed tokens: 23582474240 | elapsed time per iteration (s): 1.73 | learning rate: 1.316E-04 | global batch size: 512 | lm loss: 1.356950E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.295 | TFLOPs: 44.32 | 31: iteration 22500/ 52452 | consumed samples: 11520000 | consumed tokens: 23592960000 | elapsed time per iteration (s): 1.76 | learning rate: 1.315E-04 | global batch size: 512 | lm loss: 1.360413E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.438 | TFLOPs: 43.74 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22500 | lm loss value: 4.102280E+00 | lm loss PPL: 6.047800E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22500 | lm loss value: 4.030669E+00 | lm loss PPL: 5.629855E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22510/ 52452 | consumed samples: 11525120 | consumed tokens: 23603445760 | elapsed time per iteration (s): 9.14 | learning rate: 1.315E-04 | global batch size: 512 | lm loss: 1.361297E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.022 | TFLOPs: 8.41 | 31: iteration 22520/ 52452 | consumed samples: 11530240 | consumed tokens: 23613931520 | elapsed time per iteration (s): 1.73 | learning rate: 1.314E-04 | global batch size: 512 | lm loss: 1.361633E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.283 | TFLOPs: 44.47 | 31: iteration 22530/ 52452 | consumed samples: 11535360 | consumed tokens: 23624417280 | elapsed time per iteration (s): 1.73 | learning rate: 1.313E-04 | global batch size: 512 | lm loss: 1.356236E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.778 | TFLOPs: 44.39 | 31: iteration 22540/ 52452 | consumed samples: 11540480 | consumed tokens: 23634903040 | elapsed time per iteration (s): 1.77 | learning rate: 1.313E-04 | global batch size: 512 | lm loss: 1.354839E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.027 | TFLOPs: 43.53 | 31: iteration 22550/ 52452 | consumed samples: 11545600 | consumed tokens: 23645388800 | elapsed time per iteration (s): 1.84 | learning rate: 1.312E-04 | global batch size: 512 | lm loss: 1.364898E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.482 | TFLOPs: 41.80 | 31: iteration 22560/ 52452 | consumed samples: 11550720 | consumed tokens: 23655874560 | elapsed time per iteration (s): 1.74 | learning rate: 1.312E-04 | global batch size: 512 | lm loss: 1.370059E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.443 | TFLOPs: 44.04 | 31: iteration 22570/ 52452 | consumed samples: 11555840 | consumed tokens: 23666360320 | elapsed time per iteration (s): 1.78 | learning rate: 1.311E-04 | global batch size: 512 | lm loss: 1.360501E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.126 | TFLOPs: 43.25 | 31: iteration 22580/ 52452 | consumed samples: 11560960 | consumed tokens: 23676846080 | elapsed time per iteration (s): 1.75 | learning rate: 1.311E-04 | global batch size: 512 | lm loss: 1.352624E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.199 | TFLOPs: 44.01 | 31: iteration 22590/ 52452 | consumed samples: 11566080 | consumed tokens: 23687331840 | elapsed time per iteration (s): 1.75 | learning rate: 1.310E-04 | global batch size: 512 | lm loss: 1.359160E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.747 | TFLOPs: 43.79 | 31: iteration 22600/ 52452 | consumed samples: 11571200 | consumed tokens: 23697817600 | elapsed time per iteration (s): 1.75 | learning rate: 1.310E-04 | global batch size: 512 | lm loss: 1.363016E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.820 | TFLOPs: 43.80 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22600 | lm loss value: 4.029831E+00 | lm loss PPL: 5.625143E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22600 | lm loss value: 3.995814E+00 | lm loss PPL: 5.437010E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22610/ 52452 | consumed samples: 11576320 | consumed tokens: 23708303360 | elapsed time per iteration (s): 9.12 | learning rate: 1.309E-04 | global batch size: 512 | lm loss: 1.366263E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.119 | TFLOPs: 8.42 | 31: iteration 22620/ 52452 | consumed samples: 11581440 | consumed tokens: 23718789120 | elapsed time per iteration (s): 1.81 | learning rate: 1.309E-04 | global batch size: 512 | lm loss: 1.366934E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.507 | TFLOPs: 42.55 | 31: iteration 22630/ 52452 | consumed samples: 11586560 | consumed tokens: 23729274880 | elapsed time per iteration (s): 1.74 | learning rate: 1.308E-04 | global batch size: 512 | lm loss: 1.368281E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.603 | TFLOPs: 44.22 | 31: iteration 22640/ 52452 | consumed samples: 11591680 | consumed tokens: 23739760640 | elapsed time per iteration (s): 1.78 | learning rate: 1.308E-04 | global batch size: 512 | lm loss: 1.368069E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.295 | TFLOPs: 43.12 | 31: iteration 22650/ 52452 | consumed samples: 11596800 | consumed tokens: 23750246400 | elapsed time per iteration (s): 1.77 | learning rate: 1.307E-04 | global batch size: 512 | lm loss: 1.363793E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.239 | TFLOPs: 43.41 | 31: iteration 22660/ 52452 | consumed samples: 11601920 | consumed tokens: 23760732160 | elapsed time per iteration (s): 1.75 | learning rate: 1.307E-04 | global batch size: 512 | lm loss: 1.367486E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.973 | TFLOPs: 43.82 | 31: iteration 22670/ 52452 | consumed samples: 11607040 | consumed tokens: 23771217920 | elapsed time per iteration (s): 1.80 | learning rate: 1.306E-04 | global batch size: 512 | lm loss: 1.360766E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.159 | TFLOPs: 42.65 | 31: iteration 22680/ 52452 | consumed samples: 11612160 | consumed tokens: 23781703680 | elapsed time per iteration (s): 1.79 | learning rate: 1.306E-04 | global batch size: 512 | lm loss: 1.355085E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.348 | TFLOPs: 42.83 | 31: iteration 22690/ 52452 | consumed samples: 11617280 | consumed tokens: 23792189440 | elapsed time per iteration (s): 1.79 | learning rate: 1.305E-04 | global batch size: 512 | lm loss: 1.350631E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.627 | TFLOPs: 42.87 | 31: iteration 22700/ 52452 | consumed samples: 11622400 | consumed tokens: 23802675200 | elapsed time per iteration (s): 1.83 | learning rate: 1.304E-04 | global batch size: 512 | lm loss: 1.365031E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.542 | TFLOPs: 42.11 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22700 | lm loss value: 3.934854E+00 | lm loss PPL: 5.115467E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22700 | lm loss value: 3.933434E+00 | lm loss PPL: 5.108207E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22710/ 52452 | consumed samples: 11627520 | consumed tokens: 23813160960 | elapsed time per iteration (s): 9.14 | learning rate: 1.304E-04 | global batch size: 512 | lm loss: 1.361016E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.027 | TFLOPs: 8.41 | 31: iteration 22720/ 52452 | consumed samples: 11632640 | consumed tokens: 23823646720 | elapsed time per iteration (s): 1.84 | learning rate: 1.303E-04 | global batch size: 512 | lm loss: 1.346298E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.919 | TFLOPs: 41.86 | 31: iteration 22730/ 52452 | consumed samples: 11637760 | consumed tokens: 23834132480 | elapsed time per iteration (s): 1.81 | learning rate: 1.303E-04 | global batch size: 512 | lm loss: 1.360955E+00 | grad norm: 0.147 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.944 | TFLOPs: 42.47 | 31: iteration 22740/ 52452 | consumed samples: 11642880 | consumed tokens: 23844618240 | elapsed time per iteration (s): 1.73 | learning rate: 1.302E-04 | global batch size: 512 | lm loss: 1.351662E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.293 | TFLOPs: 44.47 | 31: iteration 22750/ 52452 | consumed samples: 11648000 | consumed tokens: 23855104000 | elapsed time per iteration (s): 1.73 | learning rate: 1.302E-04 | global batch size: 512 | lm loss: 1.359896E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.727 | TFLOPs: 44.39 | 31: iteration 22760/ 52452 | consumed samples: 11653120 | consumed tokens: 23865589760 | elapsed time per iteration (s): 1.76 | learning rate: 1.301E-04 | global batch size: 512 | lm loss: 1.352347E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.558 | TFLOPs: 43.76 | 31: iteration 22770/ 52452 | consumed samples: 11658240 | consumed tokens: 23876075520 | elapsed time per iteration (s): 1.75 | learning rate: 1.301E-04 | global batch size: 512 | lm loss: 1.343518E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.491 | TFLOPs: 43.90 | 31: iteration 22780/ 52452 | consumed samples: 11663360 | consumed tokens: 23886561280 | elapsed time per iteration (s): 1.75 | learning rate: 1.300E-04 | global batch size: 512 | lm loss: 1.362264E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.737 | TFLOPs: 43.94 | 31: iteration 22790/ 52452 | consumed samples: 11668480 | consumed tokens: 23897047040 | elapsed time per iteration (s): 1.74 | learning rate: 1.300E-04 | global batch size: 512 | lm loss: 1.357150E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.546 | TFLOPs: 44.21 | 31: iteration 22800/ 52452 | consumed samples: 11673600 | consumed tokens: 23907532800 | elapsed time per iteration (s): 1.74 | learning rate: 1.299E-04 | global batch size: 512 | lm loss: 1.359019E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.912 | TFLOPs: 44.26 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22800 | lm loss value: 4.037504E+00 | lm loss PPL: 5.668467E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22800 | lm loss value: 3.986511E+00 | lm loss PPL: 5.386665E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22810/ 52452 | consumed samples: 11678720 | consumed tokens: 23918018560 | elapsed time per iteration (s): 9.20 | learning rate: 1.299E-04 | global batch size: 512 | lm loss: 1.347554E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.651 | TFLOPs: 8.35 | 31: iteration 22820/ 52452 | consumed samples: 11683840 | consumed tokens: 23928504320 | elapsed time per iteration (s): 1.78 | learning rate: 1.298E-04 | global batch size: 512 | lm loss: 1.359148E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.901 | TFLOPs: 43.06 | 31: iteration 22830/ 52452 | consumed samples: 11688960 | consumed tokens: 23938990080 | elapsed time per iteration (s): 1.73 | learning rate: 1.298E-04 | global batch size: 512 | lm loss: 1.342425E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.262 | TFLOPs: 44.47 | 31: iteration 22840/ 52452 | consumed samples: 11694080 | consumed tokens: 23949475840 | elapsed time per iteration (s): 1.82 | learning rate: 1.297E-04 | global batch size: 512 | lm loss: 1.356385E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.084 | TFLOPs: 42.19 | 31: iteration 22850/ 52452 | consumed samples: 11699200 | consumed tokens: 23959961600 | elapsed time per iteration (s): 1.75 | learning rate: 1.297E-04 | global batch size: 512 | lm loss: 1.375220E+00 | grad norm: 0.147 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.043 | TFLOPs: 43.83 | 31: iteration 22860/ 52452 | consumed samples: 11704320 | consumed tokens: 23970447360 | elapsed time per iteration (s): 1.72 | learning rate: 1.296E-04 | global batch size: 512 | lm loss: 1.358509E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.263 | TFLOPs: 44.62 | 31: iteration 22870/ 52452 | consumed samples: 11709440 | consumed tokens: 23980933120 | elapsed time per iteration (s): 1.73 | learning rate: 1.295E-04 | global batch size: 512 | lm loss: 1.363373E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.766 | TFLOPs: 44.39 | 31: iteration 22880/ 52452 | consumed samples: 11714560 | consumed tokens: 23991418880 | elapsed time per iteration (s): 1.76 | learning rate: 1.295E-04 | global batch size: 512 | lm loss: 1.349449E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.358 | TFLOPs: 43.58 | 31: iteration 22890/ 52452 | consumed samples: 11719680 | consumed tokens: 24001904640 | elapsed time per iteration (s): 1.83 | learning rate: 1.294E-04 | global batch size: 512 | lm loss: 1.343270E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.302 | TFLOPs: 42.07 | 31: iteration 22900/ 52452 | consumed samples: 11724800 | consumed tokens: 24012390400 | elapsed time per iteration (s): 1.80 | learning rate: 1.294E-04 | global batch size: 512 | lm loss: 1.367412E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.021 | TFLOPs: 42.63 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 22900 | lm loss value: 4.078603E+00 | lm loss PPL: 5.906289E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 22900 | lm loss value: 4.019114E+00 | lm loss PPL: 5.565180E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 22910/ 52452 | consumed samples: 11729920 | consumed tokens: 24022876160 | elapsed time per iteration (s): 9.18 | learning rate: 1.293E-04 | global batch size: 512 | lm loss: 1.348751E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.774 | TFLOPs: 8.37 | 31: iteration 22920/ 52452 | consumed samples: 11735040 | consumed tokens: 24033361920 | elapsed time per iteration (s): 1.76 | learning rate: 1.293E-04 | global batch size: 512 | lm loss: 1.355796E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.623 | TFLOPs: 43.77 | 31: iteration 22930/ 52452 | consumed samples: 11740160 | consumed tokens: 24043847680 | elapsed time per iteration (s): 1.79 | learning rate: 1.292E-04 | global batch size: 512 | lm loss: 1.378856E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.713 | TFLOPs: 42.88 | 31: iteration 22940/ 52452 | consumed samples: 11745280 | consumed tokens: 24054333440 | elapsed time per iteration (s): 1.78 | learning rate: 1.292E-04 | global batch size: 512 | lm loss: 1.361077E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.402 | TFLOPs: 43.29 | 31: iteration 22950/ 52452 | consumed samples: 11750400 | consumed tokens: 24064819200 | elapsed time per iteration (s): 1.79 | learning rate: 1.291E-04 | global batch size: 512 | lm loss: 1.369214E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.821 | TFLOPs: 42.90 | 31: iteration 22960/ 52452 | consumed samples: 11755520 | consumed tokens: 24075304960 | elapsed time per iteration (s): 1.79 | learning rate: 1.291E-04 | global batch size: 512 | lm loss: 1.351896E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.537 | TFLOPs: 42.86 | 31: iteration 22970/ 52452 | consumed samples: 11760640 | consumed tokens: 24085790720 | elapsed time per iteration (s): 1.76 | learning rate: 1.290E-04 | global batch size: 512 | lm loss: 1.349735E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.713 | TFLOPs: 43.63 | 31: iteration 22980/ 52452 | consumed samples: 11765760 | consumed tokens: 24096276480 | elapsed time per iteration (s): 1.77 | learning rate: 1.290E-04 | global batch size: 512 | lm loss: 1.340825E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.436 | TFLOPs: 43.44 | 31: iteration 22990/ 52452 | consumed samples: 11770880 | consumed tokens: 24106762240 | elapsed time per iteration (s): 1.73 | learning rate: 1.289E-04 | global batch size: 512 | lm loss: 1.352134E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.211 | TFLOPs: 44.31 | 31: iteration 23000/ 52452 | consumed samples: 11776000 | consumed tokens: 24117248000 | elapsed time per iteration (s): 1.75 | learning rate: 1.289E-04 | global batch size: 512 | lm loss: 1.349107E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.571 | TFLOPs: 43.91 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23000 | lm loss value: 4.095190E+00 | lm loss PPL: 6.005072E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23000 | lm loss value: 4.040283E+00 | lm loss PPL: 5.684244E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23010/ 52452 | consumed samples: 11781120 | consumed tokens: 24127733760 | elapsed time per iteration (s): 9.18 | learning rate: 1.288E-04 | global batch size: 512 | lm loss: 1.367841E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.763 | TFLOPs: 8.37 | 31: iteration 23020/ 52452 | consumed samples: 11786240 | consumed tokens: 24138219520 | elapsed time per iteration (s): 1.82 | learning rate: 1.287E-04 | global batch size: 512 | lm loss: 1.364420E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.320 | TFLOPs: 42.22 | 31: iteration 23030/ 52452 | consumed samples: 11791360 | consumed tokens: 24148705280 | elapsed time per iteration (s): 1.77 | learning rate: 1.287E-04 | global batch size: 512 | lm loss: 1.368155E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.638 | TFLOPs: 43.47 | 31: iteration 23040/ 52452 | consumed samples: 11796480 | consumed tokens: 24159191040 | elapsed time per iteration (s): 1.77 | learning rate: 1.286E-04 | global batch size: 512 | lm loss: 1.350071E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.794 | TFLOPs: 43.35 | 31: iteration 23050/ 52452 | consumed samples: 11801600 | consumed tokens: 24169676800 | elapsed time per iteration (s): 1.78 | learning rate: 1.286E-04 | global batch size: 512 | lm loss: 1.348615E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.759 | TFLOPs: 43.19 | 31: iteration 23060/ 52452 | consumed samples: 11806720 | consumed tokens: 24180162560 | elapsed time per iteration (s): 1.82 | learning rate: 1.285E-04 | global batch size: 512 | lm loss: 1.357580E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.461 | TFLOPs: 42.25 | 31: iteration 23070/ 52452 | consumed samples: 11811840 | consumed tokens: 24190648320 | elapsed time per iteration (s): 1.82 | learning rate: 1.285E-04 | global batch size: 512 | lm loss: 1.354970E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.797 | TFLOPs: 42.30 | 31: iteration 23080/ 52452 | consumed samples: 11816960 | consumed tokens: 24201134080 | elapsed time per iteration (s): 1.80 | learning rate: 1.284E-04 | global batch size: 512 | lm loss: 1.360317E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.975 | TFLOPs: 42.77 | 31: iteration 23090/ 52452 | consumed samples: 11822080 | consumed tokens: 24211619840 | elapsed time per iteration (s): 1.85 | learning rate: 1.284E-04 | global batch size: 512 | lm loss: 1.344763E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.442 | TFLOPs: 41.64 | 31: iteration 23100/ 52452 | consumed samples: 11827200 | consumed tokens: 24222105600 | elapsed time per iteration (s): 1.83 | learning rate: 1.283E-04 | global batch size: 512 | lm loss: 1.352637E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.938 | TFLOPs: 42.02 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23100 | lm loss value: 3.957144E+00 | lm loss PPL: 5.230775E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23100 | lm loss value: 3.939501E+00 | lm loss PPL: 5.139293E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23110/ 52452 | consumed samples: 11832320 | consumed tokens: 24232591360 | elapsed time per iteration (s): 9.19 | learning rate: 1.283E-04 | global batch size: 512 | lm loss: 1.348437E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.712 | TFLOPs: 8.36 | 31: iteration 23120/ 52452 | consumed samples: 11837440 | consumed tokens: 24243077120 | elapsed time per iteration (s): 1.81 | learning rate: 1.282E-04 | global batch size: 512 | lm loss: 1.349144E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.428 | TFLOPs: 42.39 | 31: iteration 23130/ 52452 | consumed samples: 11842560 | consumed tokens: 24253562880 | elapsed time per iteration (s): 1.78 | learning rate: 1.282E-04 | global batch size: 512 | lm loss: 1.344984E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.753 | TFLOPs: 43.19 | 31: iteration 23140/ 52452 | consumed samples: 11847680 | consumed tokens: 24264048640 | elapsed time per iteration (s): 1.76 | learning rate: 1.281E-04 | global batch size: 512 | lm loss: 1.350389E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.667 | TFLOPs: 43.63 | 31: iteration 23150/ 52452 | consumed samples: 11852800 | consumed tokens: 24274534400 | elapsed time per iteration (s): 1.77 | learning rate: 1.281E-04 | global batch size: 512 | lm loss: 1.342928E+00 | grad norm: 0.189 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.958 | TFLOPs: 43.37 | 31: iteration 23160/ 52452 | consumed samples: 11857920 | consumed tokens: 24285020160 | elapsed time per iteration (s): 1.86 | learning rate: 1.280E-04 | global batch size: 512 | lm loss: 1.360875E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.667 | TFLOPs: 41.38 | 31: iteration 23170/ 52452 | consumed samples: 11863040 | consumed tokens: 24295505920 | elapsed time per iteration (s): 1.77 | learning rate: 1.279E-04 | global batch size: 512 | lm loss: 1.359802E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.654 | TFLOPs: 43.33 | 31: iteration 23180/ 52452 | consumed samples: 11868160 | consumed tokens: 24305991680 | elapsed time per iteration (s): 1.84 | learning rate: 1.279E-04 | global batch size: 512 | lm loss: 1.368124E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.357 | TFLOPs: 41.78 | 31: iteration 23190/ 52452 | consumed samples: 11873280 | consumed tokens: 24316477440 | elapsed time per iteration (s): 1.82 | learning rate: 1.278E-04 | global batch size: 512 | lm loss: 1.363118E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.005 | TFLOPs: 42.33 | 31: iteration 23200/ 52452 | consumed samples: 11878400 | consumed tokens: 24326963200 | elapsed time per iteration (s): 1.80 | learning rate: 1.278E-04 | global batch size: 512 | lm loss: 1.341910E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.836 | TFLOPs: 42.75 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23200 | lm loss value: 4.110543E+00 | lm loss PPL: 6.097981E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23200 | lm loss value: 4.065562E+00 | lm loss PPL: 5.829768E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23210/ 52452 | consumed samples: 11883520 | consumed tokens: 24337448960 | elapsed time per iteration (s): 9.35 | learning rate: 1.277E-04 | global batch size: 512 | lm loss: 1.352060E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.740 | TFLOPs: 8.22 | 31: iteration 23220/ 52452 | consumed samples: 11888640 | consumed tokens: 24347934720 | elapsed time per iteration (s): 1.87 | learning rate: 1.277E-04 | global batch size: 512 | lm loss: 1.330049E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.323 | TFLOPs: 41.17 | 31: iteration 23230/ 52452 | consumed samples: 11893760 | consumed tokens: 24358420480 | elapsed time per iteration (s): 1.78 | learning rate: 1.276E-04 | global batch size: 512 | lm loss: 1.352526E+00 | grad norm: 0.144 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.600 | TFLOPs: 43.17 | 31: iteration 23240/ 52452 | consumed samples: 11898880 | consumed tokens: 24368906240 | elapsed time per iteration (s): 1.85 | learning rate: 1.276E-04 | global batch size: 512 | lm loss: 1.358228E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.965 | TFLOPs: 41.57 | 31: iteration 23250/ 52452 | consumed samples: 11904000 | consumed tokens: 24379392000 | elapsed time per iteration (s): 1.81 | learning rate: 1.275E-04 | global batch size: 512 | lm loss: 1.338833E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.053 | TFLOPs: 42.48 | 31: iteration 23260/ 52452 | consumed samples: 11909120 | consumed tokens: 24389877760 | elapsed time per iteration (s): 1.85 | learning rate: 1.275E-04 | global batch size: 512 | lm loss: 1.343327E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.320 | TFLOPs: 41.47 | 31: iteration 23270/ 52452 | consumed samples: 11914240 | consumed tokens: 24400363520 | elapsed time per iteration (s): 1.73 | learning rate: 1.274E-04 | global batch size: 512 | lm loss: 1.344688E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.559 | TFLOPs: 44.36 | 31: iteration 23280/ 52452 | consumed samples: 11919360 | consumed tokens: 24410849280 | elapsed time per iteration (s): 1.82 | learning rate: 1.274E-04 | global batch size: 512 | lm loss: 1.358214E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.475 | TFLOPs: 42.25 | 31: iteration 23290/ 52452 | consumed samples: 11924480 | consumed tokens: 24421335040 | elapsed time per iteration (s): 1.77 | learning rate: 1.273E-04 | global batch size: 512 | lm loss: 1.347027E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.777 | TFLOPs: 43.49 | 31: iteration 23300/ 52452 | consumed samples: 11929600 | consumed tokens: 24431820800 | elapsed time per iteration (s): 1.83 | learning rate: 1.273E-04 | global batch size: 512 | lm loss: 1.359307E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.498 | TFLOPs: 42.10 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23300 | lm loss value: 4.197747E+00 | lm loss PPL: 6.653627E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23300 | lm loss value: 4.107306E+00 | lm loss PPL: 6.078275E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23310/ 52452 | consumed samples: 11934720 | consumed tokens: 24442306560 | elapsed time per iteration (s): 9.23 | learning rate: 1.272E-04 | global batch size: 512 | lm loss: 1.342824E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.469 | TFLOPs: 8.33 | 31: iteration 23320/ 52452 | consumed samples: 11939840 | consumed tokens: 24452792320 | elapsed time per iteration (s): 1.83 | learning rate: 1.271E-04 | global batch size: 512 | lm loss: 1.350090E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.324 | TFLOPs: 42.08 | 31: iteration 23330/ 52452 | consumed samples: 11944960 | consumed tokens: 24463278080 | elapsed time per iteration (s): 1.85 | learning rate: 1.271E-04 | global batch size: 512 | lm loss: 1.359249E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.426 | TFLOPs: 41.49 | 31: iteration 23340/ 52452 | consumed samples: 11950080 | consumed tokens: 24473763840 | elapsed time per iteration (s): 1.84 | learning rate: 1.270E-04 | global batch size: 512 | lm loss: 1.352289E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.097 | TFLOPs: 41.74 | 31: iteration 23350/ 52452 | consumed samples: 11955200 | consumed tokens: 24484249600 | elapsed time per iteration (s): 1.81 | learning rate: 1.270E-04 | global batch size: 512 | lm loss: 1.335525E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.948 | TFLOPs: 42.47 | 31: iteration 23360/ 52452 | consumed samples: 11960320 | consumed tokens: 24494735360 | elapsed time per iteration (s): 1.79 | learning rate: 1.269E-04 | global batch size: 512 | lm loss: 1.349355E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.700 | TFLOPs: 42.88 | 31: iteration 23370/ 52452 | consumed samples: 11965440 | consumed tokens: 24505221120 | elapsed time per iteration (s): 1.73 | learning rate: 1.269E-04 | global batch size: 512 | lm loss: 1.365724E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.753 | TFLOPs: 44.39 | 31: iteration 23380/ 52452 | consumed samples: 11970560 | consumed tokens: 24515706880 | elapsed time per iteration (s): 1.77 | learning rate: 1.268E-04 | global batch size: 512 | lm loss: 1.346466E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.947 | TFLOPs: 43.52 | 31: iteration 23390/ 52452 | consumed samples: 11975680 | consumed tokens: 24526192640 | elapsed time per iteration (s): 1.81 | learning rate: 1.268E-04 | global batch size: 512 | lm loss: 1.353633E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.100 | TFLOPs: 42.49 | 31: iteration 23400/ 52452 | consumed samples: 11980800 | consumed tokens: 24536678400 | elapsed time per iteration (s): 1.82 | learning rate: 1.267E-04 | global batch size: 512 | lm loss: 1.345000E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.768 | TFLOPs: 42.14 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23400 | lm loss value: 4.150834E+00 | lm loss PPL: 6.348693E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23400 | lm loss value: 4.056135E+00 | lm loss PPL: 5.775066E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23410/ 52452 | consumed samples: 11985920 | consumed tokens: 24547164160 | elapsed time per iteration (s): 9.12 | learning rate: 1.267E-04 | global batch size: 512 | lm loss: 1.356051E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.111 | TFLOPs: 8.42 | 31: iteration 23420/ 52452 | consumed samples: 11991040 | consumed tokens: 24557649920 | elapsed time per iteration (s): 1.82 | learning rate: 1.266E-04 | global batch size: 512 | lm loss: 1.353027E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.841 | TFLOPs: 42.30 | 31: iteration 23430/ 52452 | consumed samples: 11996160 | consumed tokens: 24568135680 | elapsed time per iteration (s): 1.76 | learning rate: 1.266E-04 | global batch size: 512 | lm loss: 1.341502E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.863 | TFLOPs: 43.66 | 31: iteration 23440/ 52452 | consumed samples: 12001280 | consumed tokens: 24578621440 | elapsed time per iteration (s): 1.76 | learning rate: 1.265E-04 | global batch size: 512 | lm loss: 1.350809E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.388 | TFLOPs: 43.59 | 31: iteration 23450/ 52452 | consumed samples: 12006400 | consumed tokens: 24589107200 | elapsed time per iteration (s): 1.79 | learning rate: 1.265E-04 | global batch size: 512 | lm loss: 1.342323E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.588 | TFLOPs: 43.02 | 31: iteration 23460/ 52452 | consumed samples: 12011520 | consumed tokens: 24599592960 | elapsed time per iteration (s): 1.74 | learning rate: 1.264E-04 | global batch size: 512 | lm loss: 1.353934E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.806 | TFLOPs: 44.10 | 31: iteration 23470/ 52452 | consumed samples: 12016640 | consumed tokens: 24610078720 | elapsed time per iteration (s): 1.79 | learning rate: 1.263E-04 | global batch size: 512 | lm loss: 1.348809E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.692 | TFLOPs: 42.88 | 31: iteration 23480/ 52452 | consumed samples: 12021760 | consumed tokens: 24620564480 | elapsed time per iteration (s): 1.82 | learning rate: 1.263E-04 | global batch size: 512 | lm loss: 1.342657E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.873 | TFLOPs: 42.16 | 31: iteration 23490/ 52452 | consumed samples: 12026880 | consumed tokens: 24631050240 | elapsed time per iteration (s): 1.76 | learning rate: 1.262E-04 | global batch size: 512 | lm loss: 1.342743E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.941 | TFLOPs: 43.67 | 31: iteration 23500/ 52452 | consumed samples: 12032000 | consumed tokens: 24641536000 | elapsed time per iteration (s): 1.77 | learning rate: 1.262E-04 | global batch size: 512 | lm loss: 1.339044E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.098 | TFLOPs: 43.39 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23500 | lm loss value: 4.106024E+00 | lm loss PPL: 6.070489E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23500 | lm loss value: 4.053474E+00 | lm loss PPL: 5.759720E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23510/ 52452 | consumed samples: 12037120 | consumed tokens: 24652021760 | elapsed time per iteration (s): 9.18 | learning rate: 1.261E-04 | global batch size: 512 | lm loss: 1.341255E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.749 | TFLOPs: 8.37 | 31: iteration 23520/ 52452 | consumed samples: 12042240 | consumed tokens: 24662507520 | elapsed time per iteration (s): 1.81 | learning rate: 1.261E-04 | global batch size: 512 | lm loss: 1.344840E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.453 | TFLOPs: 42.39 | 31: iteration 23530/ 52452 | consumed samples: 12047360 | consumed tokens: 24672993280 | elapsed time per iteration (s): 1.79 | learning rate: 1.260E-04 | global batch size: 512 | lm loss: 1.346005E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.305 | TFLOPs: 42.82 | 31: iteration 23540/ 52452 | consumed samples: 12052480 | consumed tokens: 24683479040 | elapsed time per iteration (s): 1.87 | learning rate: 1.260E-04 | global batch size: 512 | lm loss: 1.346764E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.767 | TFLOPs: 41.09 | 31: iteration 23550/ 52452 | consumed samples: 12057600 | consumed tokens: 24693964800 | elapsed time per iteration (s): 1.78 | learning rate: 1.259E-04 | global batch size: 512 | lm loss: 1.341539E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.045 | TFLOPs: 43.08 | 31: iteration 23560/ 52452 | consumed samples: 12062720 | consumed tokens: 24704450560 | elapsed time per iteration (s): 1.83 | learning rate: 1.259E-04 | global batch size: 512 | lm loss: 1.340866E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.041 | TFLOPs: 41.88 | 31: iteration 23570/ 52452 | consumed samples: 12067840 | consumed tokens: 24714936320 | elapsed time per iteration (s): 1.82 | learning rate: 1.258E-04 | global batch size: 512 | lm loss: 1.347775E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.375 | TFLOPs: 42.23 | 31: iteration 23580/ 52452 | consumed samples: 12072960 | consumed tokens: 24725422080 | elapsed time per iteration (s): 1.84 | learning rate: 1.258E-04 | global batch size: 512 | lm loss: 1.339228E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.919 | TFLOPs: 41.86 | 31: iteration 23590/ 52452 | consumed samples: 12078080 | consumed tokens: 24735907840 | elapsed time per iteration (s): 1.80 | learning rate: 1.257E-04 | global batch size: 512 | lm loss: 1.346474E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.603 | TFLOPs: 42.72 | 31: iteration 23600/ 52452 | consumed samples: 12083200 | consumed tokens: 24746393600 | elapsed time per iteration (s): 1.84 | learning rate: 1.256E-04 | global batch size: 512 | lm loss: 1.346827E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.797 | TFLOPs: 41.85 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23600 | lm loss value: 4.220072E+00 | lm loss PPL: 6.803840E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23600 | lm loss value: 4.123470E+00 | lm loss PPL: 6.177321E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23610/ 52452 | consumed samples: 12088320 | consumed tokens: 24756879360 | elapsed time per iteration (s): 9.15 | learning rate: 1.256E-04 | global batch size: 512 | lm loss: 1.347171E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.939 | TFLOPs: 8.40 | 31: iteration 23620/ 52452 | consumed samples: 12093440 | consumed tokens: 24767365120 | elapsed time per iteration (s): 1.76 | learning rate: 1.255E-04 | global batch size: 512 | lm loss: 1.346319E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.351 | TFLOPs: 43.58 | 31: iteration 23630/ 52452 | consumed samples: 12098560 | consumed tokens: 24777850880 | elapsed time per iteration (s): 1.78 | learning rate: 1.255E-04 | global batch size: 512 | lm loss: 1.337296E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.348 | TFLOPs: 43.28 | 31: iteration 23640/ 52452 | consumed samples: 12103680 | consumed tokens: 24788336640 | elapsed time per iteration (s): 1.78 | learning rate: 1.254E-04 | global batch size: 512 | lm loss: 1.343776E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.009 | TFLOPs: 43.23 | 31: iteration 23650/ 52452 | consumed samples: 12108800 | consumed tokens: 24798822400 | elapsed time per iteration (s): 1.77 | learning rate: 1.254E-04 | global batch size: 512 | lm loss: 1.342439E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.864 | TFLOPs: 43.51 | 31: iteration 23660/ 52452 | consumed samples: 12113920 | consumed tokens: 24809308160 | elapsed time per iteration (s): 1.74 | learning rate: 1.253E-04 | global batch size: 512 | lm loss: 1.359976E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.991 | TFLOPs: 44.28 | 31: iteration 23670/ 52452 | consumed samples: 12119040 | consumed tokens: 24819793920 | elapsed time per iteration (s): 1.76 | learning rate: 1.253E-04 | global batch size: 512 | lm loss: 1.342728E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.083 | TFLOPs: 43.69 | 31: iteration 23680/ 52452 | consumed samples: 12124160 | consumed tokens: 24830279680 | elapsed time per iteration (s): 1.78 | learning rate: 1.252E-04 | global batch size: 512 | lm loss: 1.339378E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.446 | TFLOPs: 43.29 | 31: iteration 23690/ 52452 | consumed samples: 12129280 | consumed tokens: 24840765440 | elapsed time per iteration (s): 1.73 | learning rate: 1.252E-04 | global batch size: 512 | lm loss: 1.346409E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.134 | TFLOPs: 44.30 | 31: iteration 23700/ 52452 | consumed samples: 12134400 | consumed tokens: 24851251200 | elapsed time per iteration (s): 1.83 | learning rate: 1.251E-04 | global batch size: 512 | lm loss: 1.348940E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.499 | TFLOPs: 41.95 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23700 | lm loss value: 4.147213E+00 | lm loss PPL: 6.325748E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23700 | lm loss value: 4.078950E+00 | lm loss PPL: 5.908340E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23710/ 52452 | consumed samples: 12139520 | consumed tokens: 24861736960 | elapsed time per iteration (s): 9.25 | learning rate: 1.251E-04 | global batch size: 512 | lm loss: 1.343109E+00 | grad norm: 0.184 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.375 | TFLOPs: 8.31 | 31: iteration 23720/ 52452 | consumed samples: 12144640 | consumed tokens: 24872222720 | elapsed time per iteration (s): 1.77 | learning rate: 1.250E-04 | global batch size: 512 | lm loss: 1.339298E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.954 | TFLOPs: 43.52 | 31: iteration 23730/ 52452 | consumed samples: 12149760 | consumed tokens: 24882708480 | elapsed time per iteration (s): 1.74 | learning rate: 1.249E-04 | global batch size: 512 | lm loss: 1.339130E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.228 | TFLOPs: 44.16 | 31: iteration 23740/ 52452 | consumed samples: 12154880 | consumed tokens: 24893194240 | elapsed time per iteration (s): 1.84 | learning rate: 1.249E-04 | global batch size: 512 | lm loss: 1.339032E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.744 | TFLOPs: 41.69 | 31: iteration 23750/ 52452 | consumed samples: 12160000 | consumed tokens: 24903680000 | elapsed time per iteration (s): 1.82 | learning rate: 1.248E-04 | global batch size: 512 | lm loss: 1.337659E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.973 | TFLOPs: 42.17 | 31: iteration 23760/ 52452 | consumed samples: 12165120 | consumed tokens: 24914165760 | elapsed time per iteration (s): 1.84 | learning rate: 1.248E-04 | global batch size: 512 | lm loss: 1.338108E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.789 | TFLOPs: 41.84 | 31: iteration 23770/ 52452 | consumed samples: 12170240 | consumed tokens: 24924651520 | elapsed time per iteration (s): 1.91 | learning rate: 1.247E-04 | global batch size: 512 | lm loss: 1.357937E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.879 | TFLOPs: 40.21 | 31: iteration 23780/ 52452 | consumed samples: 12175360 | consumed tokens: 24935137280 | elapsed time per iteration (s): 2.58 | learning rate: 1.247E-04 | global batch size: 512 | lm loss: 1.334122E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 198.473 | TFLOPs: 29.79 | 31: iteration 23790/ 52452 | consumed samples: 12180480 | consumed tokens: 24945623040 | elapsed time per iteration (s): 1.75 | learning rate: 1.246E-04 | global batch size: 512 | lm loss: 1.342398E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.208 | TFLOPs: 44.01 | 31: iteration 23800/ 52452 | consumed samples: 12185600 | consumed tokens: 24956108800 | elapsed time per iteration (s): 1.73 | learning rate: 1.246E-04 | global batch size: 512 | lm loss: 1.335262E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.616 | TFLOPs: 44.37 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23800 | lm loss value: 4.076630E+00 | lm loss PPL: 5.894646E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23800 | lm loss value: 4.032458E+00 | lm loss PPL: 5.639939E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23810/ 52452 | consumed samples: 12190720 | consumed tokens: 24966594560 | elapsed time per iteration (s): 9.17 | learning rate: 1.245E-04 | global batch size: 512 | lm loss: 1.338150E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.837 | TFLOPs: 8.38 | 31: iteration 23820/ 52452 | consumed samples: 12195840 | consumed tokens: 24977080320 | elapsed time per iteration (s): 1.77 | learning rate: 1.245E-04 | global batch size: 512 | lm loss: 1.337114E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.722 | TFLOPs: 43.49 | 31: iteration 23830/ 52452 | consumed samples: 12200960 | consumed tokens: 24987566080 | elapsed time per iteration (s): 1.73 | learning rate: 1.244E-04 | global batch size: 512 | lm loss: 1.351702E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.487 | TFLOPs: 44.35 | 31: iteration 23840/ 52452 | consumed samples: 12206080 | consumed tokens: 24998051840 | elapsed time per iteration (s): 1.79 | learning rate: 1.244E-04 | global batch size: 512 | lm loss: 1.330730E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.633 | TFLOPs: 42.87 | 31: iteration 23850/ 52452 | consumed samples: 12211200 | consumed tokens: 25008537600 | elapsed time per iteration (s): 1.74 | learning rate: 1.243E-04 | global batch size: 512 | lm loss: 1.345192E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.881 | TFLOPs: 44.26 | 31: iteration 23860/ 52452 | consumed samples: 12216320 | consumed tokens: 25019023360 | elapsed time per iteration (s): 1.77 | learning rate: 1.243E-04 | global batch size: 512 | lm loss: 1.346330E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.100 | TFLOPs: 43.39 | 31: iteration 23870/ 52452 | consumed samples: 12221440 | consumed tokens: 25029509120 | elapsed time per iteration (s): 1.74 | learning rate: 1.242E-04 | global batch size: 512 | lm loss: 1.348962E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.048 | TFLOPs: 44.29 | 31: iteration 23880/ 52452 | consumed samples: 12226560 | consumed tokens: 25039994880 | elapsed time per iteration (s): 1.74 | learning rate: 1.241E-04 | global batch size: 512 | lm loss: 1.349666E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.792 | TFLOPs: 44.25 | 31: iteration 23890/ 52452 | consumed samples: 12231680 | consumed tokens: 25050480640 | elapsed time per iteration (s): 1.80 | learning rate: 1.241E-04 | global batch size: 512 | lm loss: 1.340208E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.237 | TFLOPs: 42.66 | 31: iteration 23900/ 52452 | consumed samples: 12236800 | consumed tokens: 25060966400 | elapsed time per iteration (s): 1.81 | learning rate: 1.240E-04 | global batch size: 512 | lm loss: 1.344471E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.773 | TFLOPs: 42.44 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 23900 | lm loss value: 4.229626E+00 | lm loss PPL: 6.869152E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 23900 | lm loss value: 4.123456E+00 | lm loss PPL: 6.177233E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 23910/ 52452 | consumed samples: 12241920 | consumed tokens: 25071452160 | elapsed time per iteration (s): 9.19 | learning rate: 1.240E-04 | global batch size: 512 | lm loss: 1.340498E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.718 | TFLOPs: 8.36 | 31: iteration 23920/ 52452 | consumed samples: 12247040 | consumed tokens: 25081937920 | elapsed time per iteration (s): 1.82 | learning rate: 1.239E-04 | global batch size: 512 | lm loss: 1.344578E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.404 | TFLOPs: 42.24 | 31: iteration 23930/ 52452 | consumed samples: 12252160 | consumed tokens: 25092423680 | elapsed time per iteration (s): 1.80 | learning rate: 1.239E-04 | global batch size: 512 | lm loss: 1.329479E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.571 | TFLOPs: 42.71 | 31: iteration 23940/ 52452 | consumed samples: 12257280 | consumed tokens: 25102909440 | elapsed time per iteration (s): 1.84 | learning rate: 1.238E-04 | global batch size: 512 | lm loss: 1.333968E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.882 | TFLOPs: 41.71 | 31: iteration 23950/ 52452 | consumed samples: 12262400 | consumed tokens: 25113395200 | elapsed time per iteration (s): 1.83 | learning rate: 1.238E-04 | global batch size: 512 | lm loss: 1.340512E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.348 | TFLOPs: 41.93 | 31: iteration 23960/ 52452 | consumed samples: 12267520 | consumed tokens: 25123880960 | elapsed time per iteration (s): 1.81 | learning rate: 1.237E-04 | global batch size: 512 | lm loss: 1.337920E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.949 | TFLOPs: 42.47 | 31: iteration 23970/ 52452 | consumed samples: 12272640 | consumed tokens: 25134366720 | elapsed time per iteration (s): 1.81 | learning rate: 1.237E-04 | global batch size: 512 | lm loss: 1.346114E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.512 | TFLOPs: 42.55 | 31: iteration 23980/ 52452 | consumed samples: 12277760 | consumed tokens: 25144852480 | elapsed time per iteration (s): 1.86 | learning rate: 1.236E-04 | global batch size: 512 | lm loss: 1.338962E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.802 | TFLOPs: 41.40 | 31: iteration 23990/ 52452 | consumed samples: 12282880 | consumed tokens: 25155338240 | elapsed time per iteration (s): 1.79 | learning rate: 1.236E-04 | global batch size: 512 | lm loss: 1.348053E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.464 | TFLOPs: 43.00 | 0: [2022-12-26 01:54:23,838] [INFO] [logging.py:68:log_dist] [Rank 0] step=24000, skipped=0, lr=[0.00012349759369620987, 0.00012349759369620987, 0.00012349759369620987], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 24000/ 52452 | consumed samples: 12288000 | consumed tokens: 25165824000 | elapsed time per iteration (s): 1.92 | learning rate: 1.235E-04 | global batch size: 512 | lm loss: 1.340438E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.641 | TFLOPs: 40.02 | 0: steps: 24000 loss: 1.3358 iter time (s): 1.788 samples/sec: 286.328 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24000 | lm loss value: 4.306526E+00 | lm loss PPL: 7.418231E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24000 | lm loss value: 4.196526E+00 | lm loss PPL: 6.645504E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24010/ 52452 | consumed samples: 12293120 | consumed tokens: 25176309760 | elapsed time per iteration (s): 9.17 | learning rate: 1.234E-04 | global batch size: 512 | lm loss: 1.349311E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.860 | TFLOPs: 8.38 | 31: iteration 24020/ 52452 | consumed samples: 12298240 | consumed tokens: 25186795520 | elapsed time per iteration (s): 1.81 | learning rate: 1.234E-04 | global batch size: 512 | lm loss: 1.334070E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.653 | TFLOPs: 42.42 | 31: iteration 24030/ 52452 | consumed samples: 12303360 | consumed tokens: 25197281280 | elapsed time per iteration (s): 1.77 | learning rate: 1.233E-04 | global batch size: 512 | lm loss: 1.334016E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.590 | TFLOPs: 43.32 | 31: iteration 24040/ 52452 | consumed samples: 12308480 | consumed tokens: 25207767040 | elapsed time per iteration (s): 1.79 | learning rate: 1.233E-04 | global batch size: 512 | lm loss: 1.341809E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.677 | TFLOPs: 43.03 | 31: iteration 24050/ 52452 | consumed samples: 12313600 | consumed tokens: 25218252800 | elapsed time per iteration (s): 1.78 | learning rate: 1.232E-04 | global batch size: 512 | lm loss: 1.339232E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.994 | TFLOPs: 43.08 | 31: iteration 24060/ 52452 | consumed samples: 12318720 | consumed tokens: 25228738560 | elapsed time per iteration (s): 1.75 | learning rate: 1.232E-04 | global batch size: 512 | lm loss: 1.343428E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.766 | TFLOPs: 43.79 | 31: iteration 24070/ 52452 | consumed samples: 12323840 | consumed tokens: 25239224320 | elapsed time per iteration (s): 1.75 | learning rate: 1.231E-04 | global batch size: 512 | lm loss: 1.342922E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.090 | TFLOPs: 43.99 | 31: iteration 24080/ 52452 | consumed samples: 12328960 | consumed tokens: 25249710080 | elapsed time per iteration (s): 1.75 | learning rate: 1.231E-04 | global batch size: 512 | lm loss: 1.338891E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.944 | TFLOPs: 43.82 | 31: iteration 24090/ 52452 | consumed samples: 12334080 | consumed tokens: 25260195840 | elapsed time per iteration (s): 1.75 | learning rate: 1.230E-04 | global batch size: 512 | lm loss: 1.333759E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.202 | TFLOPs: 43.86 | 31: iteration 24100/ 52452 | consumed samples: 12339200 | consumed tokens: 25270681600 | elapsed time per iteration (s): 1.76 | learning rate: 1.230E-04 | global batch size: 512 | lm loss: 1.338770E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.311 | TFLOPs: 43.57 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24100 | lm loss value: 4.360435E+00 | lm loss PPL: 7.829115E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24100 | lm loss value: 4.251457E+00 | lm loss PPL: 7.020765E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24110/ 52452 | consumed samples: 12344320 | consumed tokens: 25281167360 | elapsed time per iteration (s): 9.10 | learning rate: 1.229E-04 | global batch size: 512 | lm loss: 1.356480E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.280 | TFLOPs: 8.45 | 31: iteration 24120/ 52452 | consumed samples: 12349440 | consumed tokens: 25291653120 | elapsed time per iteration (s): 1.76 | learning rate: 1.229E-04 | global batch size: 512 | lm loss: 1.348736E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.407 | TFLOPs: 43.74 | 31: iteration 24130/ 52452 | consumed samples: 12354560 | consumed tokens: 25302138880 | elapsed time per iteration (s): 1.77 | learning rate: 1.228E-04 | global batch size: 512 | lm loss: 1.331320E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.925 | TFLOPs: 43.52 | 31: iteration 24140/ 52452 | consumed samples: 12359680 | consumed tokens: 25312624640 | elapsed time per iteration (s): 1.76 | learning rate: 1.227E-04 | global batch size: 512 | lm loss: 1.341076E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.607 | TFLOPs: 43.77 | 31: iteration 24150/ 52452 | consumed samples: 12364800 | consumed tokens: 25323110400 | elapsed time per iteration (s): 1.76 | learning rate: 1.227E-04 | global batch size: 512 | lm loss: 1.348598E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.404 | TFLOPs: 43.59 | 31: iteration 24160/ 52452 | consumed samples: 12369920 | consumed tokens: 25333596160 | elapsed time per iteration (s): 1.82 | learning rate: 1.226E-04 | global batch size: 512 | lm loss: 1.344205E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.309 | TFLOPs: 42.22 | 31: iteration 24170/ 52452 | consumed samples: 12375040 | consumed tokens: 25344081920 | elapsed time per iteration (s): 1.81 | learning rate: 1.226E-04 | global batch size: 512 | lm loss: 1.336337E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.783 | TFLOPs: 42.44 | 31: iteration 24180/ 52452 | consumed samples: 12380160 | consumed tokens: 25354567680 | elapsed time per iteration (s): 1.78 | learning rate: 1.225E-04 | global batch size: 512 | lm loss: 1.345664E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.853 | TFLOPs: 43.06 | 31: iteration 24190/ 52452 | consumed samples: 12385280 | consumed tokens: 25365053440 | elapsed time per iteration (s): 1.76 | learning rate: 1.225E-04 | global batch size: 512 | lm loss: 1.337679E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.207 | TFLOPs: 43.56 | 31: iteration 24200/ 52452 | consumed samples: 12390400 | consumed tokens: 25375539200 | elapsed time per iteration (s): 1.74 | learning rate: 1.224E-04 | global batch size: 512 | lm loss: 1.337994E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.621 | TFLOPs: 44.22 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24200 | lm loss value: 4.143942E+00 | lm loss PPL: 6.305090E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24200 | lm loss value: 4.053241E+00 | lm loss PPL: 5.758377E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24210/ 52452 | consumed samples: 12395520 | consumed tokens: 25386024960 | elapsed time per iteration (s): 9.16 | learning rate: 1.224E-04 | global batch size: 512 | lm loss: 1.323339E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.902 | TFLOPs: 8.39 | 31: iteration 24220/ 52452 | consumed samples: 12400640 | consumed tokens: 25396510720 | elapsed time per iteration (s): 1.81 | learning rate: 1.223E-04 | global batch size: 512 | lm loss: 1.344843E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.327 | TFLOPs: 42.53 | 31: iteration 24230/ 52452 | consumed samples: 12405760 | consumed tokens: 25406996480 | elapsed time per iteration (s): 1.77 | learning rate: 1.223E-04 | global batch size: 512 | lm loss: 1.335886E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.841 | TFLOPs: 43.35 | 31: iteration 24240/ 52452 | consumed samples: 12410880 | consumed tokens: 25417482240 | elapsed time per iteration (s): 1.79 | learning rate: 1.222E-04 | global batch size: 512 | lm loss: 1.334770E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.715 | TFLOPs: 43.03 | 31: iteration 24250/ 52452 | consumed samples: 12416000 | consumed tokens: 25427968000 | elapsed time per iteration (s): 1.75 | learning rate: 1.222E-04 | global batch size: 512 | lm loss: 1.333191E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.576 | TFLOPs: 43.91 | 31: iteration 24260/ 52452 | consumed samples: 12421120 | consumed tokens: 25438453760 | elapsed time per iteration (s): 1.77 | learning rate: 1.221E-04 | global batch size: 512 | lm loss: 1.335092E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.082 | TFLOPs: 43.39 | 31: iteration 24270/ 52452 | consumed samples: 12426240 | consumed tokens: 25448939520 | elapsed time per iteration (s): 1.79 | learning rate: 1.220E-04 | global batch size: 512 | lm loss: 1.335338E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.246 | TFLOPs: 42.96 | 31: iteration 24280/ 52452 | consumed samples: 12431360 | consumed tokens: 25459425280 | elapsed time per iteration (s): 1.84 | learning rate: 1.220E-04 | global batch size: 512 | lm loss: 1.347158E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.536 | TFLOPs: 41.66 | 31: iteration 24290/ 52452 | consumed samples: 12436480 | consumed tokens: 25469911040 | elapsed time per iteration (s): 1.77 | learning rate: 1.219E-04 | global batch size: 512 | lm loss: 1.339257E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.363 | TFLOPs: 43.43 | 31: iteration 24300/ 52452 | consumed samples: 12441600 | consumed tokens: 25480396800 | elapsed time per iteration (s): 1.76 | learning rate: 1.219E-04 | global batch size: 512 | lm loss: 1.321709E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.126 | TFLOPs: 43.55 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24300 | lm loss value: 4.052700E+00 | lm loss PPL: 5.755264E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24300 | lm loss value: 3.996818E+00 | lm loss PPL: 5.442469E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24310/ 52452 | consumed samples: 12446720 | consumed tokens: 25490882560 | elapsed time per iteration (s): 9.13 | learning rate: 1.218E-04 | global batch size: 512 | lm loss: 1.331179E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.093 | TFLOPs: 8.42 | 31: iteration 24320/ 52452 | consumed samples: 12451840 | consumed tokens: 25501368320 | elapsed time per iteration (s): 1.84 | learning rate: 1.218E-04 | global batch size: 512 | lm loss: 1.333794E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.914 | TFLOPs: 41.71 | 31: iteration 24330/ 52452 | consumed samples: 12456960 | consumed tokens: 25511854080 | elapsed time per iteration (s): 1.76 | learning rate: 1.217E-04 | global batch size: 512 | lm loss: 1.338562E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.658 | TFLOPs: 43.63 | 31: iteration 24340/ 52452 | consumed samples: 12462080 | consumed tokens: 25522339840 | elapsed time per iteration (s): 1.80 | learning rate: 1.217E-04 | global batch size: 512 | lm loss: 1.328040E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.075 | TFLOPs: 42.64 | 31: iteration 24350/ 52452 | consumed samples: 12467200 | consumed tokens: 25532825600 | elapsed time per iteration (s): 1.76 | learning rate: 1.216E-04 | global batch size: 512 | lm loss: 1.333335E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.106 | TFLOPs: 43.69 | 31: iteration 24360/ 52452 | consumed samples: 12472320 | consumed tokens: 25543311360 | elapsed time per iteration (s): 1.77 | learning rate: 1.216E-04 | global batch size: 512 | lm loss: 1.338511E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.801 | TFLOPs: 43.35 | 31: iteration 24370/ 52452 | consumed samples: 12477440 | consumed tokens: 25553797120 | elapsed time per iteration (s): 1.89 | learning rate: 1.215E-04 | global batch size: 512 | lm loss: 1.349478E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.540 | TFLOPs: 40.76 | 31: iteration 24380/ 52452 | consumed samples: 12482560 | consumed tokens: 25564282880 | elapsed time per iteration (s): 1.86 | learning rate: 1.214E-04 | global batch size: 512 | lm loss: 1.338317E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.672 | TFLOPs: 41.38 | 31: iteration 24390/ 52452 | consumed samples: 12487680 | consumed tokens: 25574768640 | elapsed time per iteration (s): 1.81 | learning rate: 1.214E-04 | global batch size: 512 | lm loss: 1.341331E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.417 | TFLOPs: 42.39 | 31: iteration 24400/ 52452 | consumed samples: 12492800 | consumed tokens: 25585254400 | elapsed time per iteration (s): 1.80 | learning rate: 1.213E-04 | global batch size: 512 | lm loss: 1.333727E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.156 | TFLOPs: 42.80 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24400 | lm loss value: 4.151894E+00 | lm loss PPL: 6.355423E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24400 | lm loss value: 4.075842E+00 | lm loss PPL: 5.890005E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24410/ 52452 | consumed samples: 12497920 | consumed tokens: 25595740160 | elapsed time per iteration (s): 9.24 | learning rate: 1.213E-04 | global batch size: 512 | lm loss: 1.326821E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.435 | TFLOPs: 8.32 | 31: iteration 24420/ 52452 | consumed samples: 12503040 | consumed tokens: 25606225920 | elapsed time per iteration (s): 1.85 | learning rate: 1.212E-04 | global batch size: 512 | lm loss: 1.324632E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.920 | TFLOPs: 41.56 | 31: iteration 24430/ 52452 | consumed samples: 12508160 | consumed tokens: 25616711680 | elapsed time per iteration (s): 1.79 | learning rate: 1.212E-04 | global batch size: 512 | lm loss: 1.339741E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.574 | TFLOPs: 42.86 | 31: iteration 24440/ 52452 | consumed samples: 12513280 | consumed tokens: 25627197440 | elapsed time per iteration (s): 1.77 | learning rate: 1.211E-04 | global batch size: 512 | lm loss: 1.329124E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.657 | TFLOPs: 43.33 | 31: iteration 24450/ 52452 | consumed samples: 12518400 | consumed tokens: 25637683200 | elapsed time per iteration (s): 1.83 | learning rate: 1.211E-04 | global batch size: 512 | lm loss: 1.336985E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.287 | TFLOPs: 42.07 | 31: iteration 24460/ 52452 | consumed samples: 12523520 | consumed tokens: 25648168960 | elapsed time per iteration (s): 1.81 | learning rate: 1.210E-04 | global batch size: 512 | lm loss: 1.327644E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.109 | TFLOPs: 42.34 | 31: iteration 24470/ 52452 | consumed samples: 12528640 | consumed tokens: 25658654720 | elapsed time per iteration (s): 1.86 | learning rate: 1.210E-04 | global batch size: 512 | lm loss: 1.337239E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.928 | TFLOPs: 41.27 | 31: iteration 24480/ 52452 | consumed samples: 12533760 | consumed tokens: 25669140480 | elapsed time per iteration (s): 1.76 | learning rate: 1.209E-04 | global batch size: 512 | lm loss: 1.344207E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.779 | TFLOPs: 43.64 | 31: iteration 24490/ 52452 | consumed samples: 12538880 | consumed tokens: 25679626240 | elapsed time per iteration (s): 1.79 | learning rate: 1.209E-04 | global batch size: 512 | lm loss: 1.332287E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.379 | TFLOPs: 42.83 | 31: iteration 24500/ 52452 | consumed samples: 12544000 | consumed tokens: 25690112000 | elapsed time per iteration (s): 1.84 | learning rate: 1.208E-04 | global batch size: 512 | lm loss: 1.347640E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.632 | TFLOPs: 41.82 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24500 | lm loss value: 4.205766E+00 | lm loss PPL: 6.707194E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24500 | lm loss value: 4.114622E+00 | lm loss PPL: 6.122904E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24510/ 52452 | consumed samples: 12549120 | consumed tokens: 25700597760 | elapsed time per iteration (s): 9.18 | learning rate: 1.207E-04 | global batch size: 512 | lm loss: 1.319011E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.759 | TFLOPs: 8.37 | 31: iteration 24520/ 52452 | consumed samples: 12554240 | consumed tokens: 25711083520 | elapsed time per iteration (s): 1.74 | learning rate: 1.207E-04 | global batch size: 512 | lm loss: 1.333499E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.099 | TFLOPs: 44.29 | 31: iteration 24530/ 52452 | consumed samples: 12559360 | consumed tokens: 25721569280 | elapsed time per iteration (s): 1.89 | learning rate: 1.206E-04 | global batch size: 512 | lm loss: 1.336463E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.944 | TFLOPs: 40.67 | 31: iteration 24540/ 52452 | consumed samples: 12564480 | consumed tokens: 25732055040 | elapsed time per iteration (s): 1.84 | learning rate: 1.206E-04 | global batch size: 512 | lm loss: 1.332452E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.973 | TFLOPs: 41.72 | 31: iteration 24550/ 52452 | consumed samples: 12569600 | consumed tokens: 25742540800 | elapsed time per iteration (s): 1.78 | learning rate: 1.205E-04 | global batch size: 512 | lm loss: 1.339183E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.328 | TFLOPs: 43.13 | 31: iteration 24560/ 52452 | consumed samples: 12574720 | consumed tokens: 25753026560 | elapsed time per iteration (s): 1.84 | learning rate: 1.205E-04 | global batch size: 512 | lm loss: 1.334918E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.040 | TFLOPs: 41.73 | 31: iteration 24570/ 52452 | consumed samples: 12579840 | consumed tokens: 25763512320 | elapsed time per iteration (s): 1.78 | learning rate: 1.204E-04 | global batch size: 512 | lm loss: 1.333109E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.393 | TFLOPs: 43.14 | 31: iteration 24580/ 52452 | consumed samples: 12584960 | consumed tokens: 25773998080 | elapsed time per iteration (s): 1.78 | learning rate: 1.204E-04 | global batch size: 512 | lm loss: 1.328927E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.993 | TFLOPs: 43.08 | 31: iteration 24590/ 52452 | consumed samples: 12590080 | consumed tokens: 25784483840 | elapsed time per iteration (s): 1.85 | learning rate: 1.203E-04 | global batch size: 512 | lm loss: 1.323203E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.066 | TFLOPs: 41.59 | 31: iteration 24600/ 52452 | consumed samples: 12595200 | consumed tokens: 25794969600 | elapsed time per iteration (s): 1.81 | learning rate: 1.203E-04 | global batch size: 512 | lm loss: 1.338871E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.382 | TFLOPs: 42.53 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24600 | lm loss value: 4.223007E+00 | lm loss PPL: 6.823838E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24600 | lm loss value: 4.138256E+00 | lm loss PPL: 6.269339E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24610/ 52452 | consumed samples: 12600320 | consumed tokens: 25805455360 | elapsed time per iteration (s): 9.27 | learning rate: 1.202E-04 | global batch size: 512 | lm loss: 1.343150E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.234 | TFLOPs: 8.29 | 31: iteration 24620/ 52452 | consumed samples: 12605440 | consumed tokens: 25815941120 | elapsed time per iteration (s): 1.82 | learning rate: 1.202E-04 | global batch size: 512 | lm loss: 1.325046E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.284 | TFLOPs: 42.22 | 31: iteration 24630/ 52452 | consumed samples: 12610560 | consumed tokens: 25826426880 | elapsed time per iteration (s): 1.80 | learning rate: 1.201E-04 | global batch size: 512 | lm loss: 1.339722E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.578 | TFLOPs: 42.71 | 31: iteration 24640/ 52452 | consumed samples: 12615680 | consumed tokens: 25836912640 | elapsed time per iteration (s): 1.74 | learning rate: 1.200E-04 | global batch size: 512 | lm loss: 1.341175E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.436 | TFLOPs: 44.04 | 31: iteration 24650/ 52452 | consumed samples: 12620800 | consumed tokens: 25847398400 | elapsed time per iteration (s): 1.79 | learning rate: 1.200E-04 | global batch size: 512 | lm loss: 1.323656E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.247 | TFLOPs: 42.81 | 31: iteration 24660/ 52452 | consumed samples: 12625920 | consumed tokens: 25857884160 | elapsed time per iteration (s): 1.83 | learning rate: 1.199E-04 | global batch size: 512 | lm loss: 1.321175E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.531 | TFLOPs: 41.96 | 31: iteration 24670/ 52452 | consumed samples: 12631040 | consumed tokens: 25868369920 | elapsed time per iteration (s): 1.78 | learning rate: 1.199E-04 | global batch size: 512 | lm loss: 1.339875E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.950 | TFLOPs: 43.07 | 31: iteration 24680/ 52452 | consumed samples: 12636160 | consumed tokens: 25878855680 | elapsed time per iteration (s): 1.84 | learning rate: 1.198E-04 | global batch size: 512 | lm loss: 1.338596E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.847 | TFLOPs: 41.70 | 31: iteration 24690/ 52452 | consumed samples: 12641280 | consumed tokens: 25889341440 | elapsed time per iteration (s): 1.89 | learning rate: 1.198E-04 | global batch size: 512 | lm loss: 1.332285E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.508 | TFLOPs: 40.75 | 31: iteration 24700/ 52452 | consumed samples: 12646400 | consumed tokens: 25899827200 | elapsed time per iteration (s): 1.82 | learning rate: 1.197E-04 | global batch size: 512 | lm loss: 1.322392E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.982 | TFLOPs: 42.32 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24700 | lm loss value: 4.155592E+00 | lm loss PPL: 6.378971E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24700 | lm loss value: 4.052925E+00 | lm loss PPL: 5.756560E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24710/ 52452 | consumed samples: 12651520 | consumed tokens: 25910312960 | elapsed time per iteration (s): 9.23 | learning rate: 1.197E-04 | global batch size: 512 | lm loss: 1.333059E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.446 | TFLOPs: 8.32 | 31: iteration 24720/ 52452 | consumed samples: 12656640 | consumed tokens: 25920798720 | elapsed time per iteration (s): 1.85 | learning rate: 1.196E-04 | global batch size: 512 | lm loss: 1.332259E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.458 | TFLOPs: 41.49 | 31: iteration 24730/ 52452 | consumed samples: 12661760 | consumed tokens: 25931284480 | elapsed time per iteration (s): 1.80 | learning rate: 1.196E-04 | global batch size: 512 | lm loss: 1.322307E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.068 | TFLOPs: 42.64 | 31: iteration 24740/ 52452 | consumed samples: 12666880 | consumed tokens: 25941770240 | elapsed time per iteration (s): 1.77 | learning rate: 1.195E-04 | global batch size: 512 | lm loss: 1.346007E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.027 | TFLOPs: 43.53 | 31: iteration 24750/ 52452 | consumed samples: 12672000 | consumed tokens: 25952256000 | elapsed time per iteration (s): 1.75 | learning rate: 1.194E-04 | global batch size: 512 | lm loss: 1.324341E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.633 | TFLOPs: 43.92 | 31: iteration 24760/ 52452 | consumed samples: 12677120 | consumed tokens: 25962741760 | elapsed time per iteration (s): 1.78 | learning rate: 1.194E-04 | global batch size: 512 | lm loss: 1.336522E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.801 | TFLOPs: 43.20 | 31: iteration 24770/ 52452 | consumed samples: 12682240 | consumed tokens: 25973227520 | elapsed time per iteration (s): 1.78 | learning rate: 1.193E-04 | global batch size: 512 | lm loss: 1.332375E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.154 | TFLOPs: 43.25 | 31: iteration 24780/ 52452 | consumed samples: 12687360 | consumed tokens: 25983713280 | elapsed time per iteration (s): 1.83 | learning rate: 1.193E-04 | global batch size: 512 | lm loss: 1.339784E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.476 | TFLOPs: 42.10 | 31: iteration 24790/ 52452 | consumed samples: 12692480 | consumed tokens: 25994199040 | elapsed time per iteration (s): 1.81 | learning rate: 1.192E-04 | global batch size: 512 | lm loss: 1.338409E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.426 | TFLOPs: 42.39 | 31: iteration 24800/ 52452 | consumed samples: 12697600 | consumed tokens: 26004684800 | elapsed time per iteration (s): 1.77 | learning rate: 1.192E-04 | global batch size: 512 | lm loss: 1.322126E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.230 | TFLOPs: 43.41 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24800 | lm loss value: 4.055171E+00 | lm loss PPL: 5.769503E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24800 | lm loss value: 4.005990E+00 | lm loss PPL: 5.492618E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24810/ 52452 | consumed samples: 12702720 | consumed tokens: 26015170560 | elapsed time per iteration (s): 9.28 | learning rate: 1.191E-04 | global batch size: 512 | lm loss: 1.321124E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.181 | TFLOPs: 8.28 | 31: iteration 24820/ 52452 | consumed samples: 12707840 | consumed tokens: 26025656320 | elapsed time per iteration (s): 1.78 | learning rate: 1.191E-04 | global batch size: 512 | lm loss: 1.328190E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.440 | TFLOPs: 43.29 | 31: iteration 24830/ 52452 | consumed samples: 12712960 | consumed tokens: 26036142080 | elapsed time per iteration (s): 1.79 | learning rate: 1.190E-04 | global batch size: 512 | lm loss: 1.338092E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.799 | TFLOPs: 42.90 | 31: iteration 24840/ 52452 | consumed samples: 12718080 | consumed tokens: 26046627840 | elapsed time per iteration (s): 1.83 | learning rate: 1.190E-04 | global batch size: 512 | lm loss: 1.347391E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.138 | TFLOPs: 41.90 | 31: iteration 24850/ 52452 | consumed samples: 12723200 | consumed tokens: 26057113600 | elapsed time per iteration (s): 1.85 | learning rate: 1.189E-04 | global batch size: 512 | lm loss: 1.341498E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.591 | TFLOPs: 41.51 | 31: iteration 24860/ 52452 | consumed samples: 12728320 | consumed tokens: 26067599360 | elapsed time per iteration (s): 1.89 | learning rate: 1.189E-04 | global batch size: 512 | lm loss: 1.329458E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.362 | TFLOPs: 40.73 | 31: iteration 24870/ 52452 | consumed samples: 12733440 | consumed tokens: 26078085120 | elapsed time per iteration (s): 1.81 | learning rate: 1.188E-04 | global batch size: 512 | lm loss: 1.345183E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.367 | TFLOPs: 42.38 | 31: iteration 24880/ 52452 | consumed samples: 12738560 | consumed tokens: 26088570880 | elapsed time per iteration (s): 1.81 | learning rate: 1.187E-04 | global batch size: 512 | lm loss: 1.333103E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.923 | TFLOPs: 42.47 | 31: iteration 24890/ 52452 | consumed samples: 12743680 | consumed tokens: 26099056640 | elapsed time per iteration (s): 1.81 | learning rate: 1.187E-04 | global batch size: 512 | lm loss: 1.337992E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.208 | TFLOPs: 42.51 | 31: iteration 24900/ 52452 | consumed samples: 12748800 | consumed tokens: 26109542400 | elapsed time per iteration (s): 1.77 | learning rate: 1.186E-04 | global batch size: 512 | lm loss: 1.334531E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.268 | TFLOPs: 43.42 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 24900 | lm loss value: 4.115356E+00 | lm loss PPL: 6.127402E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 24900 | lm loss value: 4.026573E+00 | lm loss PPL: 5.606845E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 24910/ 52452 | consumed samples: 12753920 | consumed tokens: 26120028160 | elapsed time per iteration (s): 9.17 | learning rate: 1.186E-04 | global batch size: 512 | lm loss: 1.335617E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.842 | TFLOPs: 8.38 | 31: iteration 24920/ 52452 | consumed samples: 12759040 | consumed tokens: 26130513920 | elapsed time per iteration (s): 1.83 | learning rate: 1.185E-04 | global batch size: 512 | lm loss: 1.322352E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.374 | TFLOPs: 42.08 | 31: iteration 24930/ 52452 | consumed samples: 12764160 | consumed tokens: 26140999680 | elapsed time per iteration (s): 1.83 | learning rate: 1.185E-04 | global batch size: 512 | lm loss: 1.326165E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.996 | TFLOPs: 42.03 | 31: iteration 24940/ 52452 | consumed samples: 12769280 | consumed tokens: 26151485440 | elapsed time per iteration (s): 1.83 | learning rate: 1.184E-04 | global batch size: 512 | lm loss: 1.328963E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.911 | TFLOPs: 42.01 | 31: iteration 24950/ 52452 | consumed samples: 12774400 | consumed tokens: 26161971200 | elapsed time per iteration (s): 1.92 | learning rate: 1.184E-04 | global batch size: 512 | lm loss: 1.338482E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.786 | TFLOPs: 40.04 | 31: iteration 24960/ 52452 | consumed samples: 12779520 | consumed tokens: 26172456960 | elapsed time per iteration (s): 1.79 | learning rate: 1.183E-04 | global batch size: 512 | lm loss: 1.333543E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.320 | TFLOPs: 42.98 | 31: iteration 24970/ 52452 | consumed samples: 12784640 | consumed tokens: 26182942720 | elapsed time per iteration (s): 1.76 | learning rate: 1.183E-04 | global batch size: 512 | lm loss: 1.337494E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.481 | TFLOPs: 43.60 | 31: iteration 24980/ 52452 | consumed samples: 12789760 | consumed tokens: 26193428480 | elapsed time per iteration (s): 1.92 | learning rate: 1.182E-04 | global batch size: 512 | lm loss: 1.331256E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.576 | TFLOPs: 40.01 | 31: iteration 24990/ 52452 | consumed samples: 12794880 | consumed tokens: 26203914240 | elapsed time per iteration (s): 1.79 | learning rate: 1.181E-04 | global batch size: 512 | lm loss: 1.321824E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.103 | TFLOPs: 42.94 | 31: iteration 25000/ 52452 | consumed samples: 12800000 | consumed tokens: 26214400000 | elapsed time per iteration (s): 1.81 | learning rate: 1.181E-04 | global batch size: 512 | lm loss: 1.341564E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.183 | TFLOPs: 42.35 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25000 | lm loss value: 4.204094E+00 | lm loss PPL: 6.695990E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25000 | lm loss value: 4.123189E+00 | lm loss PPL: 6.175587E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25010/ 52452 | consumed samples: 12805120 | consumed tokens: 26224885760 | elapsed time per iteration (s): 9.21 | learning rate: 1.180E-04 | global batch size: 512 | lm loss: 1.345816E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.604 | TFLOPs: 8.35 | 31: iteration 25020/ 52452 | consumed samples: 12810240 | consumed tokens: 26235371520 | elapsed time per iteration (s): 1.80 | learning rate: 1.180E-04 | global batch size: 512 | lm loss: 1.334370E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.938 | TFLOPs: 42.62 | 31: iteration 25030/ 52452 | consumed samples: 12815360 | consumed tokens: 26245857280 | elapsed time per iteration (s): 1.78 | learning rate: 1.179E-04 | global batch size: 512 | lm loss: 1.318350E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.992 | TFLOPs: 43.08 | 31: iteration 25040/ 52452 | consumed samples: 12820480 | consumed tokens: 26256343040 | elapsed time per iteration (s): 1.82 | learning rate: 1.179E-04 | global batch size: 512 | lm loss: 1.329741E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.832 | TFLOPs: 42.30 | 31: iteration 25050/ 52452 | consumed samples: 12825600 | consumed tokens: 26266828800 | elapsed time per iteration (s): 1.85 | learning rate: 1.178E-04 | global batch size: 512 | lm loss: 1.339876E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.077 | TFLOPs: 41.44 | 31: iteration 25060/ 52452 | consumed samples: 12830720 | consumed tokens: 26277314560 | elapsed time per iteration (s): 1.81 | learning rate: 1.178E-04 | global batch size: 512 | lm loss: 1.327633E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.620 | TFLOPs: 42.42 | 31: iteration 25070/ 52452 | consumed samples: 12835840 | consumed tokens: 26287800320 | elapsed time per iteration (s): 1.76 | learning rate: 1.177E-04 | global batch size: 512 | lm loss: 1.343415E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.747 | TFLOPs: 43.64 | 31: iteration 25080/ 52452 | consumed samples: 12840960 | consumed tokens: 26298286080 | elapsed time per iteration (s): 1.79 | learning rate: 1.177E-04 | global batch size: 512 | lm loss: 1.328833E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.563 | TFLOPs: 42.86 | 31: iteration 25090/ 52452 | consumed samples: 12846080 | consumed tokens: 26308771840 | elapsed time per iteration (s): 1.79 | learning rate: 1.176E-04 | global batch size: 512 | lm loss: 1.324396E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.297 | TFLOPs: 42.97 | 31: iteration 25100/ 52452 | consumed samples: 12851200 | consumed tokens: 26319257600 | elapsed time per iteration (s): 1.87 | learning rate: 1.176E-04 | global batch size: 512 | lm loss: 1.328322E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.573 | TFLOPs: 41.06 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25100 | lm loss value: 4.081542E+00 | lm loss PPL: 5.923672E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25100 | lm loss value: 4.043482E+00 | lm loss PPL: 5.702455E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25110/ 52452 | consumed samples: 12856320 | consumed tokens: 26329743360 | elapsed time per iteration (s): 9.18 | learning rate: 1.175E-04 | global batch size: 512 | lm loss: 1.327396E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.756 | TFLOPs: 8.37 | 31: iteration 25120/ 52452 | consumed samples: 12861440 | consumed tokens: 26340229120 | elapsed time per iteration (s): 1.80 | learning rate: 1.174E-04 | global batch size: 512 | lm loss: 1.337540E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.125 | TFLOPs: 42.65 | 31: iteration 25130/ 52452 | consumed samples: 12866560 | consumed tokens: 26350714880 | elapsed time per iteration (s): 1.88 | learning rate: 1.174E-04 | global batch size: 512 | lm loss: 1.324348E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.487 | TFLOPs: 40.90 | 31: iteration 25140/ 52452 | consumed samples: 12871680 | consumed tokens: 26361200640 | elapsed time per iteration (s): 1.85 | learning rate: 1.173E-04 | global batch size: 512 | lm loss: 1.324986E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.105 | TFLOPs: 41.44 | 31: iteration 25150/ 52452 | consumed samples: 12876800 | consumed tokens: 26371686400 | elapsed time per iteration (s): 1.91 | learning rate: 1.173E-04 | global batch size: 512 | lm loss: 1.335399E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.322 | TFLOPs: 40.27 | 31: iteration 25160/ 52452 | consumed samples: 12881920 | consumed tokens: 26382172160 | elapsed time per iteration (s): 1.80 | learning rate: 1.172E-04 | global batch size: 512 | lm loss: 1.322014E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.477 | TFLOPs: 42.70 | 31: iteration 25170/ 52452 | consumed samples: 12887040 | consumed tokens: 26392657920 | elapsed time per iteration (s): 1.86 | learning rate: 1.172E-04 | global batch size: 512 | lm loss: 1.338760E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.694 | TFLOPs: 41.23 | 31: iteration 25180/ 52452 | consumed samples: 12892160 | consumed tokens: 26403143680 | elapsed time per iteration (s): 1.83 | learning rate: 1.171E-04 | global batch size: 512 | lm loss: 1.339674E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.503 | TFLOPs: 42.10 | 31: iteration 25190/ 52452 | consumed samples: 12897280 | consumed tokens: 26413629440 | elapsed time per iteration (s): 1.76 | learning rate: 1.171E-04 | global batch size: 512 | lm loss: 1.327759E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.555 | TFLOPs: 43.61 | 31: iteration 25200/ 52452 | consumed samples: 12902400 | consumed tokens: 26424115200 | elapsed time per iteration (s): 1.85 | learning rate: 1.170E-04 | global batch size: 512 | lm loss: 1.334904E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.451 | TFLOPs: 41.64 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25200 | lm loss value: 4.078062E+00 | lm loss PPL: 5.903096E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25200 | lm loss value: 4.016269E+00 | lm loss PPL: 5.549368E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25210/ 52452 | consumed samples: 12907520 | consumed tokens: 26434600960 | elapsed time per iteration (s): 9.13 | learning rate: 1.170E-04 | global batch size: 512 | lm loss: 1.331053E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.097 | TFLOPs: 8.42 | 31: iteration 25220/ 52452 | consumed samples: 12912640 | consumed tokens: 26445086720 | elapsed time per iteration (s): 1.76 | learning rate: 1.169E-04 | global batch size: 512 | lm loss: 1.317483E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.499 | TFLOPs: 43.60 | 31: iteration 25230/ 52452 | consumed samples: 12917760 | consumed tokens: 26455572480 | elapsed time per iteration (s): 1.74 | learning rate: 1.168E-04 | global batch size: 512 | lm loss: 1.324170E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.990 | TFLOPs: 44.13 | 31: iteration 25240/ 52452 | consumed samples: 12922880 | consumed tokens: 26466058240 | elapsed time per iteration (s): 1.76 | learning rate: 1.168E-04 | global batch size: 512 | lm loss: 1.333927E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.925 | TFLOPs: 43.67 | 31: iteration 25250/ 52452 | consumed samples: 12928000 | consumed tokens: 26476544000 | elapsed time per iteration (s): 1.78 | learning rate: 1.167E-04 | global batch size: 512 | lm loss: 1.335594E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.394 | TFLOPs: 43.29 | 31: iteration 25260/ 52452 | consumed samples: 12933120 | consumed tokens: 26487029760 | elapsed time per iteration (s): 1.78 | learning rate: 1.167E-04 | global batch size: 512 | lm loss: 1.339665E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.812 | TFLOPs: 43.20 | 31: iteration 25270/ 52452 | consumed samples: 12938240 | consumed tokens: 26497515520 | elapsed time per iteration (s): 1.75 | learning rate: 1.166E-04 | global batch size: 512 | lm loss: 1.328065E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.844 | TFLOPs: 43.95 | 31: iteration 25280/ 52452 | consumed samples: 12943360 | consumed tokens: 26508001280 | elapsed time per iteration (s): 1.73 | learning rate: 1.166E-04 | global batch size: 512 | lm loss: 1.331506E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.443 | TFLOPs: 44.49 | 31: iteration 25290/ 52452 | consumed samples: 12948480 | consumed tokens: 26518487040 | elapsed time per iteration (s): 1.75 | learning rate: 1.165E-04 | global batch size: 512 | lm loss: 1.318604E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.004 | TFLOPs: 43.83 | 31: iteration 25300/ 52452 | consumed samples: 12953600 | consumed tokens: 26528972800 | elapsed time per iteration (s): 1.75 | learning rate: 1.165E-04 | global batch size: 512 | lm loss: 1.319334E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.204 | TFLOPs: 44.01 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25300 | lm loss value: 4.165562E+00 | lm loss PPL: 6.442889E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25300 | lm loss value: 4.092951E+00 | lm loss PPL: 5.991643E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25310/ 52452 | consumed samples: 12958720 | consumed tokens: 26539458560 | elapsed time per iteration (s): 9.14 | learning rate: 1.164E-04 | global batch size: 512 | lm loss: 1.333389E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.043 | TFLOPs: 8.41 | 31: iteration 25320/ 52452 | consumed samples: 12963840 | consumed tokens: 26549944320 | elapsed time per iteration (s): 1.74 | learning rate: 1.164E-04 | global batch size: 512 | lm loss: 1.323740E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.374 | TFLOPs: 44.18 | 31: iteration 25330/ 52452 | consumed samples: 12968960 | consumed tokens: 26560430080 | elapsed time per iteration (s): 1.78 | learning rate: 1.163E-04 | global batch size: 512 | lm loss: 1.338998E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.223 | TFLOPs: 43.11 | 31: iteration 25340/ 52452 | consumed samples: 12974080 | consumed tokens: 26570915840 | elapsed time per iteration (s): 1.75 | learning rate: 1.162E-04 | global batch size: 512 | lm loss: 1.325218E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.857 | TFLOPs: 43.96 | 31: iteration 25350/ 52452 | consumed samples: 12979200 | consumed tokens: 26581401600 | elapsed time per iteration (s): 1.76 | learning rate: 1.162E-04 | global batch size: 512 | lm loss: 1.337110E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.329 | TFLOPs: 43.58 | 31: iteration 25360/ 52452 | consumed samples: 12984320 | consumed tokens: 26591887360 | elapsed time per iteration (s): 1.78 | learning rate: 1.161E-04 | global batch size: 512 | lm loss: 1.328656E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.927 | TFLOPs: 43.07 | 31: iteration 25370/ 52452 | consumed samples: 12989440 | consumed tokens: 26602373120 | elapsed time per iteration (s): 1.77 | learning rate: 1.161E-04 | global batch size: 512 | lm loss: 1.312367E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.218 | TFLOPs: 43.41 | 31: iteration 25380/ 52452 | consumed samples: 12994560 | consumed tokens: 26612858880 | elapsed time per iteration (s): 1.77 | learning rate: 1.160E-04 | global batch size: 512 | lm loss: 1.326173E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.787 | TFLOPs: 43.35 | 31: iteration 25390/ 52452 | consumed samples: 12999680 | consumed tokens: 26623344640 | elapsed time per iteration (s): 1.74 | learning rate: 1.160E-04 | global batch size: 512 | lm loss: 1.330247E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.714 | TFLOPs: 44.23 | 31: iteration 25400/ 52452 | consumed samples: 13004800 | consumed tokens: 26633830400 | elapsed time per iteration (s): 1.72 | learning rate: 1.159E-04 | global batch size: 512 | lm loss: 1.327300E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.600 | TFLOPs: 44.67 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25400 | lm loss value: 4.121696E+00 | lm loss PPL: 6.166376E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25400 | lm loss value: 4.063028E+00 | lm loss PPL: 5.815012E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25410/ 52452 | consumed samples: 13009920 | consumed tokens: 26644316160 | elapsed time per iteration (s): 9.12 | learning rate: 1.159E-04 | global batch size: 512 | lm loss: 1.326091E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.136 | TFLOPs: 8.43 | 31: iteration 25420/ 52452 | consumed samples: 13015040 | consumed tokens: 26654801920 | elapsed time per iteration (s): 1.73 | learning rate: 1.158E-04 | global batch size: 512 | lm loss: 1.317805E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.521 | TFLOPs: 44.36 | 31: iteration 25430/ 52452 | consumed samples: 13020160 | consumed tokens: 26665287680 | elapsed time per iteration (s): 1.72 | learning rate: 1.158E-04 | global batch size: 512 | lm loss: 1.335235E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.236 | TFLOPs: 44.61 | 31: iteration 25440/ 52452 | consumed samples: 13025280 | consumed tokens: 26675773440 | elapsed time per iteration (s): 1.73 | learning rate: 1.157E-04 | global batch size: 512 | lm loss: 1.327987E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.035 | TFLOPs: 44.43 | 31: iteration 25450/ 52452 | consumed samples: 13030400 | consumed tokens: 26686259200 | elapsed time per iteration (s): 1.73 | learning rate: 1.156E-04 | global batch size: 512 | lm loss: 1.334165E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.297 | TFLOPs: 44.47 | 31: iteration 25460/ 52452 | consumed samples: 13035520 | consumed tokens: 26696744960 | elapsed time per iteration (s): 1.81 | learning rate: 1.156E-04 | global batch size: 512 | lm loss: 1.325238E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.561 | TFLOPs: 42.56 | 31: iteration 25470/ 52452 | consumed samples: 13040640 | consumed tokens: 26707230720 | elapsed time per iteration (s): 1.77 | learning rate: 1.155E-04 | global batch size: 512 | lm loss: 1.326003E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.786 | TFLOPs: 43.50 | 31: iteration 25480/ 52452 | consumed samples: 13045760 | consumed tokens: 26717716480 | elapsed time per iteration (s): 1.79 | learning rate: 1.155E-04 | global batch size: 512 | lm loss: 1.328313E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.600 | TFLOPs: 42.87 | 31: iteration 25490/ 52452 | consumed samples: 13050880 | consumed tokens: 26728202240 | elapsed time per iteration (s): 1.76 | learning rate: 1.154E-04 | global batch size: 512 | lm loss: 1.306727E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.720 | TFLOPs: 43.64 | 31: iteration 25500/ 52452 | consumed samples: 13056000 | consumed tokens: 26738688000 | elapsed time per iteration (s): 1.75 | learning rate: 1.154E-04 | global batch size: 512 | lm loss: 1.323244E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.446 | TFLOPs: 43.89 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25500 | lm loss value: 4.132514E+00 | lm loss PPL: 6.233441E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25500 | lm loss value: 4.049143E+00 | lm loss PPL: 5.734831E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25510/ 52452 | consumed samples: 13061120 | consumed tokens: 26749173760 | elapsed time per iteration (s): 9.18 | learning rate: 1.153E-04 | global batch size: 512 | lm loss: 1.337503E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.780 | TFLOPs: 8.37 | 31: iteration 25520/ 52452 | consumed samples: 13066240 | consumed tokens: 26759659520 | elapsed time per iteration (s): 1.74 | learning rate: 1.153E-04 | global batch size: 512 | lm loss: 1.333243E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.876 | TFLOPs: 44.11 | 31: iteration 25530/ 52452 | consumed samples: 13071360 | consumed tokens: 26770145280 | elapsed time per iteration (s): 1.77 | learning rate: 1.152E-04 | global batch size: 512 | lm loss: 1.313940E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.060 | TFLOPs: 43.39 | 31: iteration 25540/ 52452 | consumed samples: 13076480 | consumed tokens: 26780631040 | elapsed time per iteration (s): 1.75 | learning rate: 1.152E-04 | global batch size: 512 | lm loss: 1.311060E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.125 | TFLOPs: 43.85 | 31: iteration 25550/ 52452 | consumed samples: 13081600 | consumed tokens: 26791116800 | elapsed time per iteration (s): 1.74 | learning rate: 1.151E-04 | global batch size: 512 | lm loss: 1.317646E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.245 | TFLOPs: 44.16 | 31: iteration 25560/ 52452 | consumed samples: 13086720 | consumed tokens: 26801602560 | elapsed time per iteration (s): 1.77 | learning rate: 1.151E-04 | global batch size: 512 | lm loss: 1.321274E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.827 | TFLOPs: 43.35 | 31: iteration 25570/ 52452 | consumed samples: 13091840 | consumed tokens: 26812088320 | elapsed time per iteration (s): 1.76 | learning rate: 1.150E-04 | global batch size: 512 | lm loss: 1.321002E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.539 | TFLOPs: 43.61 | 31: iteration 25580/ 52452 | consumed samples: 13096960 | consumed tokens: 26822574080 | elapsed time per iteration (s): 1.80 | learning rate: 1.149E-04 | global batch size: 512 | lm loss: 1.325721E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.970 | TFLOPs: 42.77 | 31: iteration 25590/ 52452 | consumed samples: 13102080 | consumed tokens: 26833059840 | elapsed time per iteration (s): 1.79 | learning rate: 1.149E-04 | global batch size: 512 | lm loss: 1.325259E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.254 | TFLOPs: 42.81 | 31: iteration 25600/ 52452 | consumed samples: 13107200 | consumed tokens: 26843545600 | elapsed time per iteration (s): 1.79 | learning rate: 1.148E-04 | global batch size: 512 | lm loss: 1.329399E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.044 | TFLOPs: 42.93 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25600 | lm loss value: 4.083395E+00 | lm loss PPL: 5.934658E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25600 | lm loss value: 4.024217E+00 | lm loss PPL: 5.593647E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25610/ 52452 | consumed samples: 13112320 | consumed tokens: 26854031360 | elapsed time per iteration (s): 9.20 | learning rate: 1.148E-04 | global batch size: 512 | lm loss: 1.322503E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.622 | TFLOPs: 8.35 | 31: iteration 25620/ 52452 | consumed samples: 13117440 | consumed tokens: 26864517120 | elapsed time per iteration (s): 1.84 | learning rate: 1.147E-04 | global batch size: 512 | lm loss: 1.326398E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.462 | TFLOPs: 41.80 | 31: iteration 25630/ 52452 | consumed samples: 13122560 | consumed tokens: 26875002880 | elapsed time per iteration (s): 1.78 | learning rate: 1.147E-04 | global batch size: 512 | lm loss: 1.318179E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.944 | TFLOPs: 43.07 | 31: iteration 25640/ 52452 | consumed samples: 13127680 | consumed tokens: 26885488640 | elapsed time per iteration (s): 1.79 | learning rate: 1.146E-04 | global batch size: 512 | lm loss: 1.324795E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.366 | TFLOPs: 42.98 | 31: iteration 25650/ 52452 | consumed samples: 13132800 | consumed tokens: 26895974400 | elapsed time per iteration (s): 1.74 | learning rate: 1.146E-04 | global batch size: 512 | lm loss: 1.313900E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.855 | TFLOPs: 44.26 | 31: iteration 25660/ 52452 | consumed samples: 13137920 | consumed tokens: 26906460160 | elapsed time per iteration (s): 1.73 | learning rate: 1.145E-04 | global batch size: 512 | lm loss: 1.332719E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.172 | TFLOPs: 44.45 | 31: iteration 25670/ 52452 | consumed samples: 13143040 | consumed tokens: 26916945920 | elapsed time per iteration (s): 1.75 | learning rate: 1.145E-04 | global batch size: 512 | lm loss: 1.327960E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.682 | TFLOPs: 43.93 | 31: iteration 25680/ 52452 | consumed samples: 13148160 | consumed tokens: 26927431680 | elapsed time per iteration (s): 1.75 | learning rate: 1.144E-04 | global batch size: 512 | lm loss: 1.328156E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.406 | TFLOPs: 43.89 | 31: iteration 25690/ 52452 | consumed samples: 13153280 | consumed tokens: 26937917440 | elapsed time per iteration (s): 1.83 | learning rate: 1.143E-04 | global batch size: 512 | lm loss: 1.317065E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.292 | TFLOPs: 41.92 | 31: iteration 25700/ 52452 | consumed samples: 13158400 | consumed tokens: 26948403200 | elapsed time per iteration (s): 1.92 | learning rate: 1.143E-04 | global batch size: 512 | lm loss: 1.319980E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.998 | TFLOPs: 39.92 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25700 | lm loss value: 4.189385E+00 | lm loss PPL: 6.598223E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25700 | lm loss value: 4.098220E+00 | lm loss PPL: 6.023300E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25710/ 52452 | consumed samples: 13163520 | consumed tokens: 26958888960 | elapsed time per iteration (s): 9.14 | learning rate: 1.142E-04 | global batch size: 512 | lm loss: 1.315263E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.002 | TFLOPs: 8.41 | 31: iteration 25720/ 52452 | consumed samples: 13168640 | consumed tokens: 26969374720 | elapsed time per iteration (s): 1.78 | learning rate: 1.142E-04 | global batch size: 512 | lm loss: 1.328444E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.117 | TFLOPs: 43.09 | 31: iteration 25730/ 52452 | consumed samples: 13173760 | consumed tokens: 26979860480 | elapsed time per iteration (s): 1.73 | learning rate: 1.141E-04 | global batch size: 512 | lm loss: 1.303392E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.628 | TFLOPs: 44.37 | 31: iteration 25740/ 52452 | consumed samples: 13178880 | consumed tokens: 26990346240 | elapsed time per iteration (s): 1.76 | learning rate: 1.141E-04 | global batch size: 512 | lm loss: 1.338351E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.304 | TFLOPs: 43.57 | 31: iteration 25750/ 52452 | consumed samples: 13184000 | consumed tokens: 27000832000 | elapsed time per iteration (s): 1.75 | learning rate: 1.140E-04 | global batch size: 512 | lm loss: 1.309326E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.094 | TFLOPs: 43.99 | 31: iteration 25760/ 52452 | consumed samples: 13189120 | consumed tokens: 27011317760 | elapsed time per iteration (s): 1.74 | learning rate: 1.140E-04 | global batch size: 512 | lm loss: 1.318882E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.600 | TFLOPs: 44.07 | 31: iteration 25770/ 52452 | consumed samples: 13194240 | consumed tokens: 27021803520 | elapsed time per iteration (s): 1.74 | learning rate: 1.139E-04 | global batch size: 512 | lm loss: 1.317216E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.440 | TFLOPs: 44.19 | 31: iteration 25780/ 52452 | consumed samples: 13199360 | consumed tokens: 27032289280 | elapsed time per iteration (s): 1.77 | learning rate: 1.139E-04 | global batch size: 512 | lm loss: 1.333900E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.477 | TFLOPs: 43.45 | 31: iteration 25790/ 52452 | consumed samples: 13204480 | consumed tokens: 27042775040 | elapsed time per iteration (s): 1.73 | learning rate: 1.138E-04 | global batch size: 512 | lm loss: 1.314379E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.309 | TFLOPs: 44.47 | 31: iteration 25800/ 52452 | consumed samples: 13209600 | consumed tokens: 27053260800 | elapsed time per iteration (s): 1.75 | learning rate: 1.137E-04 | global batch size: 512 | lm loss: 1.307377E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.325 | TFLOPs: 43.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25800 | lm loss value: 4.229644E+00 | lm loss PPL: 6.869276E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25800 | lm loss value: 4.128444E+00 | lm loss PPL: 6.208123E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25810/ 52452 | consumed samples: 13214720 | consumed tokens: 27063746560 | elapsed time per iteration (s): 9.13 | learning rate: 1.137E-04 | global batch size: 512 | lm loss: 1.328238E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.082 | TFLOPs: 8.42 | 31: iteration 25820/ 52452 | consumed samples: 13219840 | consumed tokens: 27074232320 | elapsed time per iteration (s): 1.73 | learning rate: 1.136E-04 | global batch size: 512 | lm loss: 1.318693E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.492 | TFLOPs: 44.35 | 31: iteration 25830/ 52452 | consumed samples: 13224960 | consumed tokens: 27084718080 | elapsed time per iteration (s): 1.74 | learning rate: 1.136E-04 | global batch size: 512 | lm loss: 1.313272E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.030 | TFLOPs: 44.13 | 31: iteration 25840/ 52452 | consumed samples: 13230080 | consumed tokens: 27095203840 | elapsed time per iteration (s): 1.72 | learning rate: 1.135E-04 | global batch size: 512 | lm loss: 1.311778E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.963 | TFLOPs: 44.57 | 31: iteration 25850/ 52452 | consumed samples: 13235200 | consumed tokens: 27105689600 | elapsed time per iteration (s): 1.72 | learning rate: 1.135E-04 | global batch size: 512 | lm loss: 1.324488E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.116 | TFLOPs: 44.60 | 31: iteration 25860/ 52452 | consumed samples: 13240320 | consumed tokens: 27116175360 | elapsed time per iteration (s): 1.75 | learning rate: 1.134E-04 | global batch size: 512 | lm loss: 1.325531E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.836 | TFLOPs: 43.80 | 31: iteration 25870/ 52452 | consumed samples: 13245440 | consumed tokens: 27126661120 | elapsed time per iteration (s): 1.73 | learning rate: 1.134E-04 | global batch size: 512 | lm loss: 1.311259E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.267 | TFLOPs: 44.32 | 31: iteration 25880/ 52452 | consumed samples: 13250560 | consumed tokens: 27137146880 | elapsed time per iteration (s): 1.75 | learning rate: 1.133E-04 | global batch size: 512 | lm loss: 1.324925E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.976 | TFLOPs: 43.82 | 31: iteration 25890/ 52452 | consumed samples: 13255680 | consumed tokens: 27147632640 | elapsed time per iteration (s): 1.75 | learning rate: 1.133E-04 | global batch size: 512 | lm loss: 1.316826E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.531 | TFLOPs: 43.91 | 31: iteration 25900/ 52452 | consumed samples: 13260800 | consumed tokens: 27158118400 | elapsed time per iteration (s): 1.74 | learning rate: 1.132E-04 | global batch size: 512 | lm loss: 1.313807E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.261 | TFLOPs: 44.17 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 25900 | lm loss value: 4.234976E+00 | lm loss PPL: 6.906004E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 25900 | lm loss value: 4.128505E+00 | lm loss PPL: 6.208505E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 25910/ 52452 | consumed samples: 13265920 | consumed tokens: 27168604160 | elapsed time per iteration (s): 9.09 | learning rate: 1.131E-04 | global batch size: 512 | lm loss: 1.313735E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.347 | TFLOPs: 8.46 | 31: iteration 25920/ 52452 | consumed samples: 13271040 | consumed tokens: 27179089920 | elapsed time per iteration (s): 1.73 | learning rate: 1.131E-04 | global batch size: 512 | lm loss: 1.332379E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.696 | TFLOPs: 44.38 | 31: iteration 25930/ 52452 | consumed samples: 13276160 | consumed tokens: 27189575680 | elapsed time per iteration (s): 1.73 | learning rate: 1.130E-04 | global batch size: 512 | lm loss: 1.321432E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.783 | TFLOPs: 44.55 | 31: iteration 25940/ 52452 | consumed samples: 13281280 | consumed tokens: 27200061440 | elapsed time per iteration (s): 1.74 | learning rate: 1.130E-04 | global batch size: 512 | lm loss: 1.319453E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.608 | TFLOPs: 44.07 | 31: iteration 25950/ 52452 | consumed samples: 13286400 | consumed tokens: 27210547200 | elapsed time per iteration (s): 1.74 | learning rate: 1.129E-04 | global batch size: 512 | lm loss: 1.319089E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.020 | TFLOPs: 44.13 | 31: iteration 25960/ 52452 | consumed samples: 13291520 | consumed tokens: 27221032960 | elapsed time per iteration (s): 1.72 | learning rate: 1.129E-04 | global batch size: 512 | lm loss: 1.322663E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.448 | TFLOPs: 44.65 | 31: iteration 25970/ 52452 | consumed samples: 13296640 | consumed tokens: 27231518720 | elapsed time per iteration (s): 1.78 | learning rate: 1.128E-04 | global batch size: 512 | lm loss: 1.318614E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.460 | TFLOPs: 43.15 | 31: iteration 25980/ 52452 | consumed samples: 13301760 | consumed tokens: 27242004480 | elapsed time per iteration (s): 1.75 | learning rate: 1.128E-04 | global batch size: 512 | lm loss: 1.329405E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.831 | TFLOPs: 43.95 | 31: iteration 25990/ 52452 | consumed samples: 13306880 | consumed tokens: 27252490240 | elapsed time per iteration (s): 1.75 | learning rate: 1.127E-04 | global batch size: 512 | lm loss: 1.314603E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.057 | TFLOPs: 43.99 | 0: [2022-12-26 03:18:33,303] [INFO] [logging.py:68:log_dist] [Rank 0] step=26000, skipped=0, lr=[0.0001126584110453081, 0.0001126584110453081, 0.0001126584110453081], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 26000/ 52452 | consumed samples: 13312000 | consumed tokens: 27262976000 | elapsed time per iteration (s): 1.73 | learning rate: 1.127E-04 | global batch size: 512 | lm loss: 1.320600E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.246 | TFLOPs: 44.31 | 0: steps: 26000 loss: 1.3077 iter time (s): 1.786 samples/sec: 286.662 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26000 | lm loss value: 4.137051E+00 | lm loss PPL: 6.261790E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26000 | lm loss value: 4.056663E+00 | lm loss PPL: 5.778118E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26010/ 52452 | consumed samples: 13317120 | consumed tokens: 27273461760 | elapsed time per iteration (s): 9.16 | learning rate: 1.126E-04 | global batch size: 512 | lm loss: 1.332088E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.914 | TFLOPs: 8.39 | 31: iteration 26020/ 52452 | consumed samples: 13322240 | consumed tokens: 27283947520 | elapsed time per iteration (s): 1.75 | learning rate: 1.125E-04 | global batch size: 512 | lm loss: 1.321881E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.765 | TFLOPs: 43.94 | 31: iteration 26030/ 52452 | consumed samples: 13327360 | consumed tokens: 27294433280 | elapsed time per iteration (s): 1.76 | learning rate: 1.125E-04 | global batch size: 512 | lm loss: 1.333996E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.978 | TFLOPs: 43.67 | 31: iteration 26040/ 52452 | consumed samples: 13332480 | consumed tokens: 27304919040 | elapsed time per iteration (s): 1.74 | learning rate: 1.124E-04 | global batch size: 512 | lm loss: 1.324089E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.043 | TFLOPs: 44.13 | 31: iteration 26050/ 52452 | consumed samples: 13337600 | consumed tokens: 27315404800 | elapsed time per iteration (s): 1.82 | learning rate: 1.124E-04 | global batch size: 512 | lm loss: 1.312012E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.133 | TFLOPs: 42.20 | 31: iteration 26060/ 52452 | consumed samples: 13342720 | consumed tokens: 27325890560 | elapsed time per iteration (s): 1.73 | learning rate: 1.123E-04 | global batch size: 512 | lm loss: 1.302172E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.407 | TFLOPs: 44.49 | 31: iteration 26070/ 52452 | consumed samples: 13347840 | consumed tokens: 27336376320 | elapsed time per iteration (s): 1.80 | learning rate: 1.123E-04 | global batch size: 512 | lm loss: 1.316122E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.399 | TFLOPs: 42.69 | 31: iteration 26080/ 52452 | consumed samples: 13352960 | consumed tokens: 27346862080 | elapsed time per iteration (s): 1.78 | learning rate: 1.122E-04 | global batch size: 512 | lm loss: 1.318584E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.864 | TFLOPs: 43.21 | 31: iteration 26090/ 52452 | consumed samples: 13358080 | consumed tokens: 27357347840 | elapsed time per iteration (s): 1.74 | learning rate: 1.122E-04 | global batch size: 512 | lm loss: 1.319826E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.684 | TFLOPs: 44.23 | 31: iteration 26100/ 52452 | consumed samples: 13363200 | consumed tokens: 27367833600 | elapsed time per iteration (s): 1.76 | learning rate: 1.121E-04 | global batch size: 512 | lm loss: 1.306041E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.957 | TFLOPs: 43.67 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26100 | lm loss value: 4.101561E+00 | lm loss PPL: 6.043456E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26100 | lm loss value: 4.044252E+00 | lm loss PPL: 5.706851E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26110/ 52452 | consumed samples: 13368320 | consumed tokens: 27378319360 | elapsed time per iteration (s): 9.19 | learning rate: 1.121E-04 | global batch size: 512 | lm loss: 1.313088E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.707 | TFLOPs: 8.36 | 31: iteration 26120/ 52452 | consumed samples: 13373440 | consumed tokens: 27388805120 | elapsed time per iteration (s): 1.77 | learning rate: 1.120E-04 | global batch size: 512 | lm loss: 1.316607E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.019 | TFLOPs: 43.53 | 31: iteration 26130/ 52452 | consumed samples: 13378560 | consumed tokens: 27399290880 | elapsed time per iteration (s): 1.73 | learning rate: 1.120E-04 | global batch size: 512 | lm loss: 1.320708E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.219 | TFLOPs: 44.46 | 31: iteration 26140/ 52452 | consumed samples: 13383680 | consumed tokens: 27409776640 | elapsed time per iteration (s): 1.74 | learning rate: 1.119E-04 | global batch size: 512 | lm loss: 1.321629E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.704 | TFLOPs: 44.08 | 31: iteration 26150/ 52452 | consumed samples: 13388800 | consumed tokens: 27420262400 | elapsed time per iteration (s): 1.80 | learning rate: 1.118E-04 | global batch size: 512 | lm loss: 1.315785E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.174 | TFLOPs: 42.65 | 31: iteration 26160/ 52452 | consumed samples: 13393920 | consumed tokens: 27430748160 | elapsed time per iteration (s): 1.77 | learning rate: 1.118E-04 | global batch size: 512 | lm loss: 1.316191E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.848 | TFLOPs: 43.35 | 31: iteration 26170/ 52452 | consumed samples: 13399040 | consumed tokens: 27441233920 | elapsed time per iteration (s): 1.87 | learning rate: 1.117E-04 | global batch size: 512 | lm loss: 1.314118E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.876 | TFLOPs: 41.11 | 31: iteration 26180/ 52452 | consumed samples: 13404160 | consumed tokens: 27451719680 | elapsed time per iteration (s): 1.75 | learning rate: 1.117E-04 | global batch size: 512 | lm loss: 1.314935E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.365 | TFLOPs: 43.88 | 31: iteration 26190/ 52452 | consumed samples: 13409280 | consumed tokens: 27462205440 | elapsed time per iteration (s): 1.79 | learning rate: 1.116E-04 | global batch size: 512 | lm loss: 1.308537E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.850 | TFLOPs: 42.90 | 31: iteration 26200/ 52452 | consumed samples: 13414400 | consumed tokens: 27472691200 | elapsed time per iteration (s): 1.77 | learning rate: 1.116E-04 | global batch size: 512 | lm loss: 1.320997E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.967 | TFLOPs: 43.52 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26200 | lm loss value: 4.099112E+00 | lm loss PPL: 6.028673E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26200 | lm loss value: 4.027051E+00 | lm loss PPL: 5.609524E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26210/ 52452 | consumed samples: 13419520 | consumed tokens: 27483176960 | elapsed time per iteration (s): 9.27 | learning rate: 1.115E-04 | global batch size: 512 | lm loss: 1.311611E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.252 | TFLOPs: 8.29 | 31: iteration 26220/ 52452 | consumed samples: 13424640 | consumed tokens: 27493662720 | elapsed time per iteration (s): 1.80 | learning rate: 1.115E-04 | global batch size: 512 | lm loss: 1.316383E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.132 | TFLOPs: 42.80 | 31: iteration 26230/ 52452 | consumed samples: 13429760 | consumed tokens: 27504148480 | elapsed time per iteration (s): 1.86 | learning rate: 1.114E-04 | global batch size: 512 | lm loss: 1.311601E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.708 | TFLOPs: 41.38 | 31: iteration 26240/ 52452 | consumed samples: 13434880 | consumed tokens: 27514634240 | elapsed time per iteration (s): 1.85 | learning rate: 1.114E-04 | global batch size: 512 | lm loss: 1.307695E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.016 | TFLOPs: 41.43 | 31: iteration 26250/ 52452 | consumed samples: 13440000 | consumed tokens: 27525120000 | elapsed time per iteration (s): 1.81 | learning rate: 1.113E-04 | global batch size: 512 | lm loss: 1.318128E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.095 | TFLOPs: 42.49 | 31: iteration 26260/ 52452 | consumed samples: 13445120 | consumed tokens: 27535605760 | elapsed time per iteration (s): 1.89 | learning rate: 1.112E-04 | global batch size: 512 | lm loss: 1.320819E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.051 | TFLOPs: 40.68 | 31: iteration 26270/ 52452 | consumed samples: 13450240 | consumed tokens: 27546091520 | elapsed time per iteration (s): 1.78 | learning rate: 1.112E-04 | global batch size: 512 | lm loss: 1.316655E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.661 | TFLOPs: 43.18 | 31: iteration 26280/ 52452 | consumed samples: 13455360 | consumed tokens: 27556577280 | elapsed time per iteration (s): 1.84 | learning rate: 1.111E-04 | global batch size: 512 | lm loss: 1.324816E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.508 | TFLOPs: 41.65 | 31: iteration 26290/ 52452 | consumed samples: 13460480 | consumed tokens: 27567063040 | elapsed time per iteration (s): 1.81 | learning rate: 1.111E-04 | global batch size: 512 | lm loss: 1.322338E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.663 | TFLOPs: 42.43 | 31: iteration 26300/ 52452 | consumed samples: 13465600 | consumed tokens: 27577548800 | elapsed time per iteration (s): 1.88 | learning rate: 1.110E-04 | global batch size: 512 | lm loss: 1.317468E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.347 | TFLOPs: 40.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26300 | lm loss value: 3.955351E+00 | lm loss PPL: 5.221401E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26300 | lm loss value: 3.924394E+00 | lm loss PPL: 5.062240E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26310/ 52452 | consumed samples: 13470720 | consumed tokens: 27588034560 | elapsed time per iteration (s): 9.20 | learning rate: 1.110E-04 | global batch size: 512 | lm loss: 1.321081E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.670 | TFLOPs: 8.36 | 31: iteration 26320/ 52452 | consumed samples: 13475840 | consumed tokens: 27598520320 | elapsed time per iteration (s): 1.81 | learning rate: 1.109E-04 | global batch size: 512 | lm loss: 1.301418E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.167 | TFLOPs: 42.35 | 31: iteration 26330/ 52452 | consumed samples: 13480960 | consumed tokens: 27609006080 | elapsed time per iteration (s): 1.75 | learning rate: 1.109E-04 | global batch size: 512 | lm loss: 1.312687E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.385 | TFLOPs: 43.89 | 31: iteration 26340/ 52452 | consumed samples: 13486080 | consumed tokens: 27619491840 | elapsed time per iteration (s): 1.79 | learning rate: 1.108E-04 | global batch size: 512 | lm loss: 1.310737E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.361 | TFLOPs: 42.98 | 31: iteration 26350/ 52452 | consumed samples: 13491200 | consumed tokens: 27629977600 | elapsed time per iteration (s): 1.80 | learning rate: 1.108E-04 | global batch size: 512 | lm loss: 1.311296E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.678 | TFLOPs: 42.58 | 31: iteration 26360/ 52452 | consumed samples: 13496320 | consumed tokens: 27640463360 | elapsed time per iteration (s): 1.79 | learning rate: 1.107E-04 | global batch size: 512 | lm loss: 1.331260E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.322 | TFLOPs: 42.83 | 31: iteration 26370/ 52452 | consumed samples: 13501440 | consumed tokens: 27650949120 | elapsed time per iteration (s): 1.83 | learning rate: 1.106E-04 | global batch size: 512 | lm loss: 1.322501E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.359 | TFLOPs: 41.93 | 31: iteration 26380/ 52452 | consumed samples: 13506560 | consumed tokens: 27661434880 | elapsed time per iteration (s): 1.84 | learning rate: 1.106E-04 | global batch size: 512 | lm loss: 1.326778E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.309 | TFLOPs: 41.77 | 31: iteration 26390/ 52452 | consumed samples: 13511680 | consumed tokens: 27671920640 | elapsed time per iteration (s): 1.81 | learning rate: 1.105E-04 | global batch size: 512 | lm loss: 1.314455E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.460 | TFLOPs: 42.40 | 31: iteration 26400/ 52452 | consumed samples: 13516800 | consumed tokens: 27682406400 | elapsed time per iteration (s): 1.79 | learning rate: 1.105E-04 | global batch size: 512 | lm loss: 1.320720E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.682 | TFLOPs: 43.03 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26400 | lm loss value: 4.134480E+00 | lm loss PPL: 6.245708E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26400 | lm loss value: 4.057313E+00 | lm loss PPL: 5.781874E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26410/ 52452 | consumed samples: 13521920 | consumed tokens: 27692892160 | elapsed time per iteration (s): 9.26 | learning rate: 1.104E-04 | global batch size: 512 | lm loss: 1.306396E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.310 | TFLOPs: 8.30 | 31: iteration 26420/ 52452 | consumed samples: 13527040 | consumed tokens: 27703377920 | elapsed time per iteration (s): 1.82 | learning rate: 1.104E-04 | global batch size: 512 | lm loss: 1.324919E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.584 | TFLOPs: 42.11 | 31: iteration 26430/ 52452 | consumed samples: 13532160 | consumed tokens: 27713863680 | elapsed time per iteration (s): 1.83 | learning rate: 1.103E-04 | global batch size: 512 | lm loss: 1.321590E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.775 | TFLOPs: 41.99 | 31: iteration 26440/ 52452 | consumed samples: 13537280 | consumed tokens: 27724349440 | elapsed time per iteration (s): 1.79 | learning rate: 1.103E-04 | global batch size: 512 | lm loss: 1.317454E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.450 | TFLOPs: 42.84 | 31: iteration 26450/ 52452 | consumed samples: 13542400 | consumed tokens: 27734835200 | elapsed time per iteration (s): 1.86 | learning rate: 1.102E-04 | global batch size: 512 | lm loss: 1.322239E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.648 | TFLOPs: 41.37 | 31: iteration 26460/ 52452 | consumed samples: 13547520 | consumed tokens: 27745320960 | elapsed time per iteration (s): 1.87 | learning rate: 1.102E-04 | global batch size: 512 | lm loss: 1.314666E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.164 | TFLOPs: 41.15 | 31: iteration 26470/ 52452 | consumed samples: 13552640 | consumed tokens: 27755806720 | elapsed time per iteration (s): 1.90 | learning rate: 1.101E-04 | global batch size: 512 | lm loss: 1.310994E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.424 | TFLOPs: 40.44 | 31: iteration 26480/ 52452 | consumed samples: 13557760 | consumed tokens: 27766292480 | elapsed time per iteration (s): 1.85 | learning rate: 1.100E-04 | global batch size: 512 | lm loss: 1.311232E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.362 | TFLOPs: 41.63 | 31: iteration 26490/ 52452 | consumed samples: 13562880 | consumed tokens: 27776778240 | elapsed time per iteration (s): 1.91 | learning rate: 1.100E-04 | global batch size: 512 | lm loss: 1.317514E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.154 | TFLOPs: 40.25 | 31: iteration 26500/ 52452 | consumed samples: 13568000 | consumed tokens: 27787264000 | elapsed time per iteration (s): 1.81 | learning rate: 1.099E-04 | global batch size: 512 | lm loss: 1.308106E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.374 | TFLOPs: 42.53 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26500 | lm loss value: 4.187002E+00 | lm loss PPL: 6.582513E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26500 | lm loss value: 4.085996E+00 | lm loss PPL: 5.950115E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26510/ 52452 | consumed samples: 13573120 | consumed tokens: 27797749760 | elapsed time per iteration (s): 9.24 | learning rate: 1.099E-04 | global batch size: 512 | lm loss: 1.308361E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.416 | TFLOPs: 8.32 | 31: iteration 26520/ 52452 | consumed samples: 13578240 | consumed tokens: 27808235520 | elapsed time per iteration (s): 1.84 | learning rate: 1.098E-04 | global batch size: 512 | lm loss: 1.317920E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.677 | TFLOPs: 41.68 | 31: iteration 26530/ 52452 | consumed samples: 13583360 | consumed tokens: 27818721280 | elapsed time per iteration (s): 1.80 | learning rate: 1.098E-04 | global batch size: 512 | lm loss: 1.320805E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.254 | TFLOPs: 42.66 | 31: iteration 26540/ 52452 | consumed samples: 13588480 | consumed tokens: 27829207040 | elapsed time per iteration (s): 1.81 | learning rate: 1.097E-04 | global batch size: 512 | lm loss: 1.321750E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.147 | TFLOPs: 42.35 | 31: iteration 26550/ 52452 | consumed samples: 13593600 | consumed tokens: 27839692800 | elapsed time per iteration (s): 1.81 | learning rate: 1.097E-04 | global batch size: 512 | lm loss: 1.302692E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.698 | TFLOPs: 42.43 | 31: iteration 26560/ 52452 | consumed samples: 13598720 | consumed tokens: 27850178560 | elapsed time per iteration (s): 1.84 | learning rate: 1.096E-04 | global batch size: 512 | lm loss: 1.306518E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.154 | TFLOPs: 41.75 | 31: iteration 26570/ 52452 | consumed samples: 13603840 | consumed tokens: 27860664320 | elapsed time per iteration (s): 1.80 | learning rate: 1.096E-04 | global batch size: 512 | lm loss: 1.301444E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.960 | TFLOPs: 42.77 | 31: iteration 26580/ 52452 | consumed samples: 13608960 | consumed tokens: 27871150080 | elapsed time per iteration (s): 1.82 | learning rate: 1.095E-04 | global batch size: 512 | lm loss: 1.323134E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.374 | TFLOPs: 42.23 | 31: iteration 26590/ 52452 | consumed samples: 13614080 | consumed tokens: 27881635840 | elapsed time per iteration (s): 1.81 | learning rate: 1.094E-04 | global batch size: 512 | lm loss: 1.300669E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.882 | TFLOPs: 42.46 | 31: iteration 26600/ 52452 | consumed samples: 13619200 | consumed tokens: 27892121600 | elapsed time per iteration (s): 1.77 | learning rate: 1.094E-04 | global batch size: 512 | lm loss: 1.308607E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.987 | TFLOPs: 43.53 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26600 | lm loss value: 4.179192E+00 | lm loss PPL: 6.531303E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26600 | lm loss value: 4.076846E+00 | lm loss PPL: 5.895923E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26610/ 52452 | consumed samples: 13624320 | consumed tokens: 27902607360 | elapsed time per iteration (s): 9.18 | learning rate: 1.093E-04 | global batch size: 512 | lm loss: 1.316161E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.754 | TFLOPs: 8.37 | 31: iteration 26620/ 52452 | consumed samples: 13629440 | consumed tokens: 27913093120 | elapsed time per iteration (s): 1.78 | learning rate: 1.093E-04 | global batch size: 512 | lm loss: 1.324419E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.037 | TFLOPs: 43.08 | 31: iteration 26630/ 52452 | consumed samples: 13634560 | consumed tokens: 27923578880 | elapsed time per iteration (s): 1.82 | learning rate: 1.092E-04 | global batch size: 512 | lm loss: 1.313899E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.557 | TFLOPs: 42.11 | 31: iteration 26640/ 52452 | consumed samples: 13639680 | consumed tokens: 27934064640 | elapsed time per iteration (s): 1.80 | learning rate: 1.092E-04 | global batch size: 512 | lm loss: 1.313122E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.745 | TFLOPs: 42.59 | 31: iteration 26650/ 52452 | consumed samples: 13644800 | consumed tokens: 27944550400 | elapsed time per iteration (s): 1.78 | learning rate: 1.091E-04 | global batch size: 512 | lm loss: 1.317161E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.149 | TFLOPs: 43.10 | 31: iteration 26660/ 52452 | consumed samples: 13649920 | consumed tokens: 27955036160 | elapsed time per iteration (s): 1.77 | learning rate: 1.091E-04 | global batch size: 512 | lm loss: 1.312761E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.775 | TFLOPs: 43.49 | 31: iteration 26670/ 52452 | consumed samples: 13655040 | consumed tokens: 27965521920 | elapsed time per iteration (s): 1.82 | learning rate: 1.090E-04 | global batch size: 512 | lm loss: 1.307919E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.735 | TFLOPs: 42.29 | 31: iteration 26680/ 52452 | consumed samples: 13660160 | consumed tokens: 27976007680 | elapsed time per iteration (s): 1.84 | learning rate: 1.090E-04 | global batch size: 512 | lm loss: 1.314707E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.732 | TFLOPs: 41.84 | 31: iteration 26690/ 52452 | consumed samples: 13665280 | consumed tokens: 27986493440 | elapsed time per iteration (s): 1.80 | learning rate: 1.089E-04 | global batch size: 512 | lm loss: 1.315015E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.707 | TFLOPs: 42.58 | 31: iteration 26700/ 52452 | consumed samples: 13670400 | consumed tokens: 27996979200 | elapsed time per iteration (s): 1.85 | learning rate: 1.088E-04 | global batch size: 512 | lm loss: 1.308524E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.824 | TFLOPs: 41.55 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26700 | lm loss value: 4.105739E+00 | lm loss PPL: 6.068758E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26700 | lm loss value: 4.033672E+00 | lm loss PPL: 5.646787E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26710/ 52452 | consumed samples: 13675520 | consumed tokens: 28007464960 | elapsed time per iteration (s): 9.11 | learning rate: 1.088E-04 | global batch size: 512 | lm loss: 1.302873E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.180 | TFLOPs: 8.43 | 31: iteration 26720/ 52452 | consumed samples: 13680640 | consumed tokens: 28017950720 | elapsed time per iteration (s): 1.91 | learning rate: 1.087E-04 | global batch size: 512 | lm loss: 1.322741E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.325 | TFLOPs: 40.27 | 31: iteration 26730/ 52452 | consumed samples: 13685760 | consumed tokens: 28028436480 | elapsed time per iteration (s): 1.78 | learning rate: 1.087E-04 | global batch size: 512 | lm loss: 1.313740E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.153 | TFLOPs: 43.10 | 31: iteration 26740/ 52452 | consumed samples: 13690880 | consumed tokens: 28038922240 | elapsed time per iteration (s): 2.05 | learning rate: 1.086E-04 | global batch size: 512 | lm loss: 1.321508E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 249.213 | TFLOPs: 37.41 | 31: iteration 26750/ 52452 | consumed samples: 13696000 | consumed tokens: 28049408000 | elapsed time per iteration (s): 1.88 | learning rate: 1.086E-04 | global batch size: 512 | lm loss: 1.323942E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.042 | TFLOPs: 40.98 | 31: iteration 26760/ 52452 | consumed samples: 13701120 | consumed tokens: 28059893760 | elapsed time per iteration (s): 2.06 | learning rate: 1.085E-04 | global batch size: 512 | lm loss: 1.310218E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 249.081 | TFLOPs: 37.39 | 31: iteration 26770/ 52452 | consumed samples: 13706240 | consumed tokens: 28070379520 | elapsed time per iteration (s): 1.87 | learning rate: 1.085E-04 | global batch size: 512 | lm loss: 1.313066E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.378 | TFLOPs: 41.18 | 31: iteration 26780/ 52452 | consumed samples: 13711360 | consumed tokens: 28080865280 | elapsed time per iteration (s): 1.83 | learning rate: 1.084E-04 | global batch size: 512 | lm loss: 1.307980E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.686 | TFLOPs: 41.98 | 31: iteration 26790/ 52452 | consumed samples: 13716480 | consumed tokens: 28091351040 | elapsed time per iteration (s): 2.02 | learning rate: 1.084E-04 | global batch size: 512 | lm loss: 1.311659E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 253.648 | TFLOPs: 38.07 | 31: iteration 26800/ 52452 | consumed samples: 13721600 | consumed tokens: 28101836800 | elapsed time per iteration (s): 1.79 | learning rate: 1.083E-04 | global batch size: 512 | lm loss: 1.312636E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.920 | TFLOPs: 42.91 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26800 | lm loss value: 3.987882E+00 | lm loss PPL: 5.394052E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26800 | lm loss value: 3.945453E+00 | lm loss PPL: 5.169977E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26810/ 52452 | consumed samples: 13726720 | consumed tokens: 28112322560 | elapsed time per iteration (s): 9.33 | learning rate: 1.082E-04 | global batch size: 512 | lm loss: 1.309243E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.894 | TFLOPs: 8.24 | 31: iteration 26820/ 52452 | consumed samples: 13731840 | consumed tokens: 28122808320 | elapsed time per iteration (s): 1.82 | learning rate: 1.082E-04 | global batch size: 512 | lm loss: 1.315278E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.240 | TFLOPs: 42.21 | 31: iteration 26830/ 52452 | consumed samples: 13736960 | consumed tokens: 28133294080 | elapsed time per iteration (s): 1.84 | learning rate: 1.081E-04 | global batch size: 512 | lm loss: 1.318089E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.541 | TFLOPs: 41.81 | 31: iteration 26840/ 52452 | consumed samples: 13742080 | consumed tokens: 28143779840 | elapsed time per iteration (s): 1.79 | learning rate: 1.081E-04 | global batch size: 512 | lm loss: 1.312148E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.080 | TFLOPs: 42.94 | 31: iteration 26850/ 52452 | consumed samples: 13747200 | consumed tokens: 28154265600 | elapsed time per iteration (s): 1.88 | learning rate: 1.080E-04 | global batch size: 512 | lm loss: 1.313885E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.640 | TFLOPs: 40.77 | 31: iteration 26860/ 52452 | consumed samples: 13752320 | consumed tokens: 28164751360 | elapsed time per iteration (s): 1.87 | learning rate: 1.080E-04 | global batch size: 512 | lm loss: 1.313391E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.233 | TFLOPs: 41.16 | 31: iteration 26870/ 52452 | consumed samples: 13757440 | consumed tokens: 28175237120 | elapsed time per iteration (s): 1.84 | learning rate: 1.079E-04 | global batch size: 512 | lm loss: 1.313554E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.267 | TFLOPs: 41.77 | 31: iteration 26880/ 52452 | consumed samples: 13762560 | consumed tokens: 28185722880 | elapsed time per iteration (s): 1.79 | learning rate: 1.079E-04 | global batch size: 512 | lm loss: 1.317464E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.350 | TFLOPs: 42.98 | 31: iteration 26890/ 52452 | consumed samples: 13767680 | consumed tokens: 28196208640 | elapsed time per iteration (s): 1.84 | learning rate: 1.078E-04 | global batch size: 512 | lm loss: 1.320531E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.670 | TFLOPs: 41.68 | 31: iteration 26900/ 52452 | consumed samples: 13772800 | consumed tokens: 28206694400 | elapsed time per iteration (s): 1.82 | learning rate: 1.078E-04 | global batch size: 512 | lm loss: 1.308554E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.898 | TFLOPs: 42.31 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 26900 | lm loss value: 4.100239E+00 | lm loss PPL: 6.035473E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 26900 | lm loss value: 4.036659E+00 | lm loss PPL: 5.663679E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 26910/ 52452 | consumed samples: 13777920 | consumed tokens: 28217180160 | elapsed time per iteration (s): 9.16 | learning rate: 1.077E-04 | global batch size: 512 | lm loss: 1.310837E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.910 | TFLOPs: 8.39 | 31: iteration 26920/ 52452 | consumed samples: 13783040 | consumed tokens: 28227665920 | elapsed time per iteration (s): 1.80 | learning rate: 1.076E-04 | global batch size: 512 | lm loss: 1.313038E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.873 | TFLOPs: 42.61 | 31: iteration 26930/ 52452 | consumed samples: 13788160 | consumed tokens: 28238151680 | elapsed time per iteration (s): 1.80 | learning rate: 1.076E-04 | global batch size: 512 | lm loss: 1.305975E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.741 | TFLOPs: 42.74 | 31: iteration 26940/ 52452 | consumed samples: 13793280 | consumed tokens: 28248637440 | elapsed time per iteration (s): 1.80 | learning rate: 1.075E-04 | global batch size: 512 | lm loss: 1.304464E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.230 | TFLOPs: 42.81 | 31: iteration 26950/ 52452 | consumed samples: 13798400 | consumed tokens: 28259123200 | elapsed time per iteration (s): 1.78 | learning rate: 1.075E-04 | global batch size: 512 | lm loss: 1.308240E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.155 | TFLOPs: 43.25 | 31: iteration 26960/ 52452 | consumed samples: 13803520 | consumed tokens: 28269608960 | elapsed time per iteration (s): 1.77 | learning rate: 1.074E-04 | global batch size: 512 | lm loss: 1.292176E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.962 | TFLOPs: 43.37 | 31: iteration 26970/ 52452 | consumed samples: 13808640 | consumed tokens: 28280094720 | elapsed time per iteration (s): 1.80 | learning rate: 1.074E-04 | global batch size: 512 | lm loss: 1.306278E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.535 | TFLOPs: 42.71 | 31: iteration 26980/ 52452 | consumed samples: 13813760 | consumed tokens: 28290580480 | elapsed time per iteration (s): 1.74 | learning rate: 1.073E-04 | global batch size: 512 | lm loss: 1.316659E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.895 | TFLOPs: 44.11 | 31: iteration 26990/ 52452 | consumed samples: 13818880 | consumed tokens: 28301066240 | elapsed time per iteration (s): 1.79 | learning rate: 1.073E-04 | global batch size: 512 | lm loss: 1.311723E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.936 | TFLOPs: 42.92 | 31: iteration 27000/ 52452 | consumed samples: 13824000 | consumed tokens: 28311552000 | elapsed time per iteration (s): 1.90 | learning rate: 1.072E-04 | global batch size: 512 | lm loss: 1.320130E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.289 | TFLOPs: 40.42 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27000 | lm loss value: 4.145646E+00 | lm loss PPL: 6.315842E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27000 | lm loss value: 4.053424E+00 | lm loss PPL: 5.759434E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27010/ 52452 | consumed samples: 13829120 | consumed tokens: 28322037760 | elapsed time per iteration (s): 9.18 | learning rate: 1.072E-04 | global batch size: 512 | lm loss: 1.316594E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.787 | TFLOPs: 8.37 | 31: iteration 27020/ 52452 | consumed samples: 13834240 | consumed tokens: 28332523520 | elapsed time per iteration (s): 1.84 | learning rate: 1.071E-04 | global batch size: 512 | lm loss: 1.304711E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.967 | TFLOPs: 41.87 | 31: iteration 27030/ 52452 | consumed samples: 13839360 | consumed tokens: 28343009280 | elapsed time per iteration (s): 1.79 | learning rate: 1.071E-04 | global batch size: 512 | lm loss: 1.312760E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.275 | TFLOPs: 42.97 | 31: iteration 27040/ 52452 | consumed samples: 13844480 | consumed tokens: 28353495040 | elapsed time per iteration (s): 1.75 | learning rate: 1.070E-04 | global batch size: 512 | lm loss: 1.306184E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.498 | TFLOPs: 43.90 | 31: iteration 27050/ 52452 | consumed samples: 13849600 | consumed tokens: 28363980800 | elapsed time per iteration (s): 1.77 | learning rate: 1.069E-04 | global batch size: 512 | lm loss: 1.311433E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.108 | TFLOPs: 43.39 | 31: iteration 27060/ 52452 | consumed samples: 13854720 | consumed tokens: 28374466560 | elapsed time per iteration (s): 1.83 | learning rate: 1.069E-04 | global batch size: 512 | lm loss: 1.303918E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.498 | TFLOPs: 41.95 | 31: iteration 27070/ 52452 | consumed samples: 13859840 | consumed tokens: 28384952320 | elapsed time per iteration (s): 1.80 | learning rate: 1.068E-04 | global batch size: 512 | lm loss: 1.300126E+00 | grad norm: 0.147 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.681 | TFLOPs: 42.73 | 31: iteration 27080/ 52452 | consumed samples: 13864960 | consumed tokens: 28395438080 | elapsed time per iteration (s): 1.74 | learning rate: 1.068E-04 | global batch size: 512 | lm loss: 1.294900E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.903 | TFLOPs: 44.11 | 31: iteration 27090/ 52452 | consumed samples: 13870080 | consumed tokens: 28405923840 | elapsed time per iteration (s): 1.82 | learning rate: 1.067E-04 | global batch size: 512 | lm loss: 1.303958E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.696 | TFLOPs: 42.13 | 31: iteration 27100/ 52452 | consumed samples: 13875200 | consumed tokens: 28416409600 | elapsed time per iteration (s): 1.75 | learning rate: 1.067E-04 | global batch size: 512 | lm loss: 1.306963E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.766 | TFLOPs: 43.79 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27100 | lm loss value: 4.135625E+00 | lm loss PPL: 6.252868E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27100 | lm loss value: 4.040182E+00 | lm loss PPL: 5.683669E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27110/ 52452 | consumed samples: 13880320 | consumed tokens: 28426895360 | elapsed time per iteration (s): 9.18 | learning rate: 1.066E-04 | global batch size: 512 | lm loss: 1.320900E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.790 | TFLOPs: 8.37 | 31: iteration 27120/ 52452 | consumed samples: 13885440 | consumed tokens: 28437381120 | elapsed time per iteration (s): 1.78 | learning rate: 1.066E-04 | global batch size: 512 | lm loss: 1.314132E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.297 | TFLOPs: 43.27 | 31: iteration 27130/ 52452 | consumed samples: 13890560 | consumed tokens: 28447866880 | elapsed time per iteration (s): 1.85 | learning rate: 1.065E-04 | global batch size: 512 | lm loss: 1.304608E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.018 | TFLOPs: 41.43 | 31: iteration 27140/ 52452 | consumed samples: 13895680 | consumed tokens: 28458352640 | elapsed time per iteration (s): 1.90 | learning rate: 1.065E-04 | global batch size: 512 | lm loss: 1.308552E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.138 | TFLOPs: 40.55 | 31: iteration 27150/ 52452 | consumed samples: 13900800 | consumed tokens: 28468838400 | elapsed time per iteration (s): 1.76 | learning rate: 1.064E-04 | global batch size: 512 | lm loss: 1.309696E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.728 | TFLOPs: 43.79 | 31: iteration 27160/ 52452 | consumed samples: 13905920 | consumed tokens: 28479324160 | elapsed time per iteration (s): 1.82 | learning rate: 1.063E-04 | global batch size: 512 | lm loss: 1.315217E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.614 | TFLOPs: 42.12 | 31: iteration 27170/ 52452 | consumed samples: 13911040 | consumed tokens: 28489809920 | elapsed time per iteration (s): 1.84 | learning rate: 1.063E-04 | global batch size: 512 | lm loss: 1.309332E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.514 | TFLOPs: 41.65 | 31: iteration 27180/ 52452 | consumed samples: 13916160 | consumed tokens: 28500295680 | elapsed time per iteration (s): 1.85 | learning rate: 1.062E-04 | global batch size: 512 | lm loss: 1.308717E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.901 | TFLOPs: 41.56 | 31: iteration 27190/ 52452 | consumed samples: 13921280 | consumed tokens: 28510781440 | elapsed time per iteration (s): 1.80 | learning rate: 1.062E-04 | global batch size: 512 | lm loss: 1.309800E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.347 | TFLOPs: 42.68 | 31: iteration 27200/ 52452 | consumed samples: 13926400 | consumed tokens: 28521267200 | elapsed time per iteration (s): 1.76 | learning rate: 1.061E-04 | global batch size: 512 | lm loss: 1.303545E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.259 | TFLOPs: 43.72 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27200 | lm loss value: 4.014624E+00 | lm loss PPL: 5.540247E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27200 | lm loss value: 3.964196E+00 | lm loss PPL: 5.267789E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27210/ 52452 | consumed samples: 13931520 | consumed tokens: 28531752960 | elapsed time per iteration (s): 9.12 | learning rate: 1.061E-04 | global batch size: 512 | lm loss: 1.294765E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.115 | TFLOPs: 8.42 | 31: iteration 27220/ 52452 | consumed samples: 13936640 | consumed tokens: 28542238720 | elapsed time per iteration (s): 1.76 | learning rate: 1.060E-04 | global batch size: 512 | lm loss: 1.303765E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.245 | TFLOPs: 43.71 | 31: iteration 27230/ 52452 | consumed samples: 13941760 | consumed tokens: 28552724480 | elapsed time per iteration (s): 1.81 | learning rate: 1.060E-04 | global batch size: 512 | lm loss: 1.296323E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.609 | TFLOPs: 42.57 | 31: iteration 27240/ 52452 | consumed samples: 13946880 | consumed tokens: 28563210240 | elapsed time per iteration (s): 1.78 | learning rate: 1.059E-04 | global batch size: 512 | lm loss: 1.296097E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.176 | TFLOPs: 43.10 | 31: iteration 27250/ 52452 | consumed samples: 13952000 | consumed tokens: 28573696000 | elapsed time per iteration (s): 1.78 | learning rate: 1.059E-04 | global batch size: 512 | lm loss: 1.291609E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.359 | TFLOPs: 43.28 | 31: iteration 27260/ 52452 | consumed samples: 13957120 | consumed tokens: 28584181760 | elapsed time per iteration (s): 1.79 | learning rate: 1.058E-04 | global batch size: 512 | lm loss: 1.313885E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.698 | TFLOPs: 43.03 | 31: iteration 27270/ 52452 | consumed samples: 13962240 | consumed tokens: 28594667520 | elapsed time per iteration (s): 1.75 | learning rate: 1.057E-04 | global batch size: 512 | lm loss: 1.295682E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.461 | TFLOPs: 43.90 | 31: iteration 27280/ 52452 | consumed samples: 13967360 | consumed tokens: 28605153280 | elapsed time per iteration (s): 1.80 | learning rate: 1.057E-04 | global batch size: 512 | lm loss: 1.315666E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.708 | TFLOPs: 42.73 | 31: iteration 27290/ 52452 | consumed samples: 13972480 | consumed tokens: 28615639040 | elapsed time per iteration (s): 1.77 | learning rate: 1.056E-04 | global batch size: 512 | lm loss: 1.310595E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.515 | TFLOPs: 43.30 | 31: iteration 27300/ 52452 | consumed samples: 13977600 | consumed tokens: 28626124800 | elapsed time per iteration (s): 1.77 | learning rate: 1.056E-04 | global batch size: 512 | lm loss: 1.296224E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.823 | TFLOPs: 43.50 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27300 | lm loss value: 4.002090E+00 | lm loss PPL: 5.471238E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27300 | lm loss value: 3.965418E+00 | lm loss PPL: 5.274232E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27310/ 52452 | consumed samples: 13982720 | consumed tokens: 28636610560 | elapsed time per iteration (s): 9.22 | learning rate: 1.055E-04 | global batch size: 512 | lm loss: 1.287892E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.557 | TFLOPs: 8.34 | 31: iteration 27320/ 52452 | consumed samples: 13987840 | consumed tokens: 28647096320 | elapsed time per iteration (s): 1.92 | learning rate: 1.055E-04 | global batch size: 512 | lm loss: 1.307150E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.655 | TFLOPs: 40.02 | 31: iteration 27330/ 52452 | consumed samples: 13992960 | consumed tokens: 28657582080 | elapsed time per iteration (s): 1.87 | learning rate: 1.054E-04 | global batch size: 512 | lm loss: 1.308029E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.130 | TFLOPs: 41.00 | 31: iteration 27340/ 52452 | consumed samples: 13998080 | consumed tokens: 28668067840 | elapsed time per iteration (s): 1.84 | learning rate: 1.054E-04 | global batch size: 512 | lm loss: 1.312700E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.360 | TFLOPs: 41.78 | 31: iteration 27350/ 52452 | consumed samples: 14003200 | consumed tokens: 28678553600 | elapsed time per iteration (s): 1.78 | learning rate: 1.053E-04 | global batch size: 512 | lm loss: 1.295490E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.166 | TFLOPs: 43.10 | 31: iteration 27360/ 52452 | consumed samples: 14008320 | consumed tokens: 28689039360 | elapsed time per iteration (s): 1.88 | learning rate: 1.053E-04 | global batch size: 512 | lm loss: 1.304549E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.859 | TFLOPs: 40.95 | 31: iteration 27370/ 52452 | consumed samples: 14013440 | consumed tokens: 28699525120 | elapsed time per iteration (s): 1.79 | learning rate: 1.052E-04 | global batch size: 512 | lm loss: 1.298854E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.397 | TFLOPs: 42.84 | 31: iteration 27380/ 52452 | consumed samples: 14018560 | consumed tokens: 28710010880 | elapsed time per iteration (s): 1.75 | learning rate: 1.051E-04 | global batch size: 512 | lm loss: 1.298710E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.936 | TFLOPs: 43.82 | 31: iteration 27390/ 52452 | consumed samples: 14023680 | consumed tokens: 28720496640 | elapsed time per iteration (s): 1.76 | learning rate: 1.051E-04 | global batch size: 512 | lm loss: 1.296281E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.973 | TFLOPs: 43.67 | 31: iteration 27400/ 52452 | consumed samples: 14028800 | consumed tokens: 28730982400 | elapsed time per iteration (s): 1.79 | learning rate: 1.050E-04 | global batch size: 512 | lm loss: 1.295052E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.519 | TFLOPs: 42.85 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27400 | lm loss value: 4.217360E+00 | lm loss PPL: 6.785411E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27400 | lm loss value: 4.115487E+00 | lm loss PPL: 6.128206E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27410/ 52452 | consumed samples: 14033920 | consumed tokens: 28741468160 | elapsed time per iteration (s): 9.12 | learning rate: 1.050E-04 | global batch size: 512 | lm loss: 1.317717E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.164 | TFLOPs: 8.43 | 31: iteration 27420/ 52452 | consumed samples: 14039040 | consumed tokens: 28751953920 | elapsed time per iteration (s): 1.75 | learning rate: 1.049E-04 | global batch size: 512 | lm loss: 1.295016E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.336 | TFLOPs: 44.03 | 31: iteration 27430/ 52452 | consumed samples: 14044160 | consumed tokens: 28762439680 | elapsed time per iteration (s): 1.74 | learning rate: 1.049E-04 | global batch size: 512 | lm loss: 1.297138E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.713 | TFLOPs: 44.08 | 31: iteration 27440/ 52452 | consumed samples: 14049280 | consumed tokens: 28772925440 | elapsed time per iteration (s): 1.83 | learning rate: 1.048E-04 | global batch size: 512 | lm loss: 1.316966E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.676 | TFLOPs: 41.98 | 31: iteration 27450/ 52452 | consumed samples: 14054400 | consumed tokens: 28783411200 | elapsed time per iteration (s): 1.77 | learning rate: 1.048E-04 | global batch size: 512 | lm loss: 1.302981E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.769 | TFLOPs: 43.34 | 31: iteration 27460/ 52452 | consumed samples: 14059520 | consumed tokens: 28793896960 | elapsed time per iteration (s): 1.78 | learning rate: 1.047E-04 | global batch size: 512 | lm loss: 1.308618E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.909 | TFLOPs: 43.21 | 31: iteration 27470/ 52452 | consumed samples: 14064640 | consumed tokens: 28804382720 | elapsed time per iteration (s): 1.76 | learning rate: 1.047E-04 | global batch size: 512 | lm loss: 1.308172E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.547 | TFLOPs: 43.76 | 31: iteration 27480/ 52452 | consumed samples: 14069760 | consumed tokens: 28814868480 | elapsed time per iteration (s): 1.73 | learning rate: 1.046E-04 | global batch size: 512 | lm loss: 1.295396E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.499 | TFLOPs: 44.35 | 31: iteration 27490/ 52452 | consumed samples: 14074880 | consumed tokens: 28825354240 | elapsed time per iteration (s): 1.73 | learning rate: 1.045E-04 | global batch size: 512 | lm loss: 1.297384E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.327 | TFLOPs: 44.48 | 31: iteration 27500/ 52452 | consumed samples: 14080000 | consumed tokens: 28835840000 | elapsed time per iteration (s): 1.77 | learning rate: 1.045E-04 | global batch size: 512 | lm loss: 1.296605E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.991 | TFLOPs: 43.53 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27500 | lm loss value: 4.149406E+00 | lm loss PPL: 6.339633E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27500 | lm loss value: 4.079131E+00 | lm loss PPL: 5.909407E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27510/ 52452 | consumed samples: 14085120 | consumed tokens: 28846325760 | elapsed time per iteration (s): 9.15 | learning rate: 1.044E-04 | global batch size: 512 | lm loss: 1.303251E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.930 | TFLOPs: 8.39 | 31: iteration 27520/ 52452 | consumed samples: 14090240 | consumed tokens: 28856811520 | elapsed time per iteration (s): 1.77 | learning rate: 1.044E-04 | global batch size: 512 | lm loss: 1.303125E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.863 | TFLOPs: 43.51 | 31: iteration 27530/ 52452 | consumed samples: 14095360 | consumed tokens: 28867297280 | elapsed time per iteration (s): 1.84 | learning rate: 1.043E-04 | global batch size: 512 | lm loss: 1.307970E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.059 | TFLOPs: 41.74 | 31: iteration 27540/ 52452 | consumed samples: 14100480 | consumed tokens: 28877783040 | elapsed time per iteration (s): 1.75 | learning rate: 1.043E-04 | global batch size: 512 | lm loss: 1.307400E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.764 | TFLOPs: 43.79 | 31: iteration 27550/ 52452 | consumed samples: 14105600 | consumed tokens: 28888268800 | elapsed time per iteration (s): 1.80 | learning rate: 1.042E-04 | global batch size: 512 | lm loss: 1.301233E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.905 | TFLOPs: 42.61 | 31: iteration 27560/ 52452 | consumed samples: 14110720 | consumed tokens: 28898754560 | elapsed time per iteration (s): 1.79 | learning rate: 1.042E-04 | global batch size: 512 | lm loss: 1.306655E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.393 | TFLOPs: 42.84 | 31: iteration 27570/ 52452 | consumed samples: 14115840 | consumed tokens: 28909240320 | elapsed time per iteration (s): 1.78 | learning rate: 1.041E-04 | global batch size: 512 | lm loss: 1.300028E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.943 | TFLOPs: 43.22 | 31: iteration 27580/ 52452 | consumed samples: 14120960 | consumed tokens: 28919726080 | elapsed time per iteration (s): 1.75 | learning rate: 1.041E-04 | global batch size: 512 | lm loss: 1.313457E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.939 | TFLOPs: 43.82 | 31: iteration 27590/ 52452 | consumed samples: 14126080 | consumed tokens: 28930211840 | elapsed time per iteration (s): 1.73 | learning rate: 1.040E-04 | global batch size: 512 | lm loss: 1.296758E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.912 | TFLOPs: 44.41 | 31: iteration 27600/ 52452 | consumed samples: 14131200 | consumed tokens: 28940697600 | elapsed time per iteration (s): 1.75 | learning rate: 1.040E-04 | global batch size: 512 | lm loss: 1.301788E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.058 | TFLOPs: 43.99 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27600 | lm loss value: 4.034539E+00 | lm loss PPL: 5.651685E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27600 | lm loss value: 4.004781E+00 | lm loss PPL: 5.485982E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27610/ 52452 | consumed samples: 14136320 | consumed tokens: 28951183360 | elapsed time per iteration (s): 9.12 | learning rate: 1.039E-04 | global batch size: 512 | lm loss: 1.299534E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.111 | TFLOPs: 8.42 | 31: iteration 27620/ 52452 | consumed samples: 14141440 | consumed tokens: 28961669120 | elapsed time per iteration (s): 1.76 | learning rate: 1.038E-04 | global batch size: 512 | lm loss: 1.311714E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.518 | TFLOPs: 43.61 | 31: iteration 27630/ 52452 | consumed samples: 14146560 | consumed tokens: 28972154880 | elapsed time per iteration (s): 1.77 | learning rate: 1.038E-04 | global batch size: 512 | lm loss: 1.304105E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.696 | TFLOPs: 43.48 | 31: iteration 27640/ 52452 | consumed samples: 14151680 | consumed tokens: 28982640640 | elapsed time per iteration (s): 1.78 | learning rate: 1.037E-04 | global batch size: 512 | lm loss: 1.296938E+00 | grad norm: 0.147 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.082 | TFLOPs: 43.24 | 31: iteration 27650/ 52452 | consumed samples: 14156800 | consumed tokens: 28993126400 | elapsed time per iteration (s): 1.76 | learning rate: 1.037E-04 | global batch size: 512 | lm loss: 1.309592E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.339 | TFLOPs: 43.73 | 31: iteration 27660/ 52452 | consumed samples: 14161920 | consumed tokens: 29003612160 | elapsed time per iteration (s): 1.77 | learning rate: 1.036E-04 | global batch size: 512 | lm loss: 1.298107E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.734 | TFLOPs: 43.34 | 31: iteration 27670/ 52452 | consumed samples: 14167040 | consumed tokens: 29014097920 | elapsed time per iteration (s): 1.78 | learning rate: 1.036E-04 | global batch size: 512 | lm loss: 1.308241E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.116 | TFLOPs: 43.09 | 31: iteration 27680/ 52452 | consumed samples: 14172160 | consumed tokens: 29024583680 | elapsed time per iteration (s): 1.78 | learning rate: 1.035E-04 | global batch size: 512 | lm loss: 1.312439E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.405 | TFLOPs: 43.14 | 31: iteration 27690/ 52452 | consumed samples: 14177280 | consumed tokens: 29035069440 | elapsed time per iteration (s): 1.79 | learning rate: 1.035E-04 | global batch size: 512 | lm loss: 1.305186E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.359 | TFLOPs: 42.83 | 31: iteration 27700/ 52452 | consumed samples: 14182400 | consumed tokens: 29045555200 | elapsed time per iteration (s): 1.75 | learning rate: 1.034E-04 | global batch size: 512 | lm loss: 1.303953E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.090 | TFLOPs: 43.84 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27700 | lm loss value: 4.250173E+00 | lm loss PPL: 7.011755E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27700 | lm loss value: 4.158265E+00 | lm loss PPL: 6.396043E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27710/ 52452 | consumed samples: 14187520 | consumed tokens: 29056040960 | elapsed time per iteration (s): 9.14 | learning rate: 1.034E-04 | global batch size: 512 | lm loss: 1.289825E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.043 | TFLOPs: 8.41 | 31: iteration 27720/ 52452 | consumed samples: 14192640 | consumed tokens: 29066526720 | elapsed time per iteration (s): 1.79 | learning rate: 1.033E-04 | global batch size: 512 | lm loss: 1.304422E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.833 | TFLOPs: 42.90 | 31: iteration 27730/ 52452 | consumed samples: 14197760 | consumed tokens: 29077012480 | elapsed time per iteration (s): 1.78 | learning rate: 1.032E-04 | global batch size: 512 | lm loss: 1.305401E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.273 | TFLOPs: 43.27 | 31: iteration 27740/ 52452 | consumed samples: 14202880 | consumed tokens: 29087498240 | elapsed time per iteration (s): 1.80 | learning rate: 1.032E-04 | global batch size: 512 | lm loss: 1.302622E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.975 | TFLOPs: 42.62 | 31: iteration 27750/ 52452 | consumed samples: 14208000 | consumed tokens: 29097984000 | elapsed time per iteration (s): 1.81 | learning rate: 1.031E-04 | global batch size: 512 | lm loss: 1.305379E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.220 | TFLOPs: 42.51 | 31: iteration 27760/ 52452 | consumed samples: 14213120 | consumed tokens: 29108469760 | elapsed time per iteration (s): 1.78 | learning rate: 1.031E-04 | global batch size: 512 | lm loss: 1.301218E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.101 | TFLOPs: 43.24 | 31: iteration 27770/ 52452 | consumed samples: 14218240 | consumed tokens: 29118955520 | elapsed time per iteration (s): 1.75 | learning rate: 1.030E-04 | global batch size: 512 | lm loss: 1.302248E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.798 | TFLOPs: 43.80 | 31: iteration 27780/ 52452 | consumed samples: 14223360 | consumed tokens: 29129441280 | elapsed time per iteration (s): 1.78 | learning rate: 1.030E-04 | global batch size: 512 | lm loss: 1.303665E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.954 | TFLOPs: 43.22 | 31: iteration 27790/ 52452 | consumed samples: 14228480 | consumed tokens: 29139927040 | elapsed time per iteration (s): 1.78 | learning rate: 1.029E-04 | global batch size: 512 | lm loss: 1.301004E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.921 | TFLOPs: 43.22 | 31: iteration 27800/ 52452 | consumed samples: 14233600 | consumed tokens: 29150412800 | elapsed time per iteration (s): 1.73 | learning rate: 1.029E-04 | global batch size: 512 | lm loss: 1.310470E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.144 | TFLOPs: 44.45 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27800 | lm loss value: 4.170066E+00 | lm loss PPL: 6.471975E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27800 | lm loss value: 4.077144E+00 | lm loss PPL: 5.897680E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27810/ 52452 | consumed samples: 14238720 | consumed tokens: 29160898560 | elapsed time per iteration (s): 9.22 | learning rate: 1.028E-04 | global batch size: 512 | lm loss: 1.298350E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.552 | TFLOPs: 8.34 | 31: iteration 27820/ 52452 | consumed samples: 14243840 | consumed tokens: 29171384320 | elapsed time per iteration (s): 1.75 | learning rate: 1.028E-04 | global batch size: 512 | lm loss: 1.307007E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.868 | TFLOPs: 43.81 | 31: iteration 27830/ 52452 | consumed samples: 14248960 | consumed tokens: 29181870080 | elapsed time per iteration (s): 1.77 | learning rate: 1.027E-04 | global batch size: 512 | lm loss: 1.299688E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.481 | TFLOPs: 43.30 | 31: iteration 27840/ 52452 | consumed samples: 14254080 | consumed tokens: 29192355840 | elapsed time per iteration (s): 1.77 | learning rate: 1.026E-04 | global batch size: 512 | lm loss: 1.314739E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.853 | TFLOPs: 43.36 | 31: iteration 27850/ 52452 | consumed samples: 14259200 | consumed tokens: 29202841600 | elapsed time per iteration (s): 1.76 | learning rate: 1.026E-04 | global batch size: 512 | lm loss: 1.297331E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.040 | TFLOPs: 43.68 | 31: iteration 27860/ 52452 | consumed samples: 14264320 | consumed tokens: 29213327360 | elapsed time per iteration (s): 1.81 | learning rate: 1.025E-04 | global batch size: 512 | lm loss: 1.300582E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.539 | TFLOPs: 42.41 | 31: iteration 27870/ 52452 | consumed samples: 14269440 | consumed tokens: 29223813120 | elapsed time per iteration (s): 1.74 | learning rate: 1.025E-04 | global batch size: 512 | lm loss: 1.299199E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.959 | TFLOPs: 44.12 | 31: iteration 27880/ 52452 | consumed samples: 14274560 | consumed tokens: 29234298880 | elapsed time per iteration (s): 1.77 | learning rate: 1.024E-04 | global batch size: 512 | lm loss: 1.299339E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.752 | TFLOPs: 43.49 | 31: iteration 27890/ 52452 | consumed samples: 14279680 | consumed tokens: 29244784640 | elapsed time per iteration (s): 1.82 | learning rate: 1.024E-04 | global batch size: 512 | lm loss: 1.305062E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.228 | TFLOPs: 42.21 | 31: iteration 27900/ 52452 | consumed samples: 14284800 | consumed tokens: 29255270400 | elapsed time per iteration (s): 1.75 | learning rate: 1.023E-04 | global batch size: 512 | lm loss: 1.293322E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.259 | TFLOPs: 44.02 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 27900 | lm loss value: 4.054846E+00 | lm loss PPL: 5.767630E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 27900 | lm loss value: 4.011719E+00 | lm loss PPL: 5.524174E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 27910/ 52452 | consumed samples: 14289920 | consumed tokens: 29265756160 | elapsed time per iteration (s): 9.20 | learning rate: 1.023E-04 | global batch size: 512 | lm loss: 1.301388E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.678 | TFLOPs: 8.36 | 31: iteration 27920/ 52452 | consumed samples: 14295040 | consumed tokens: 29276241920 | elapsed time per iteration (s): 1.75 | learning rate: 1.022E-04 | global batch size: 512 | lm loss: 1.303420E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.660 | TFLOPs: 43.93 | 31: iteration 27930/ 52452 | consumed samples: 14300160 | consumed tokens: 29286727680 | elapsed time per iteration (s): 1.77 | learning rate: 1.022E-04 | global batch size: 512 | lm loss: 1.286255E+00 | grad norm: 0.147 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.347 | TFLOPs: 43.43 | 31: iteration 27940/ 52452 | consumed samples: 14305280 | consumed tokens: 29297213440 | elapsed time per iteration (s): 1.82 | learning rate: 1.021E-04 | global batch size: 512 | lm loss: 1.311186E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.391 | TFLOPs: 42.24 | 31: iteration 27950/ 52452 | consumed samples: 14310400 | consumed tokens: 29307699200 | elapsed time per iteration (s): 1.77 | learning rate: 1.021E-04 | global batch size: 512 | lm loss: 1.300995E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.036 | TFLOPs: 43.38 | 31: iteration 27960/ 52452 | consumed samples: 14315520 | consumed tokens: 29318184960 | elapsed time per iteration (s): 1.76 | learning rate: 1.020E-04 | global batch size: 512 | lm loss: 1.291607E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.487 | TFLOPs: 43.75 | 31: iteration 27970/ 52452 | consumed samples: 14320640 | consumed tokens: 29328670720 | elapsed time per iteration (s): 1.84 | learning rate: 1.019E-04 | global batch size: 512 | lm loss: 1.302796E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.565 | TFLOPs: 41.81 | 31: iteration 27980/ 52452 | consumed samples: 14325760 | consumed tokens: 29339156480 | elapsed time per iteration (s): 1.76 | learning rate: 1.019E-04 | global batch size: 512 | lm loss: 1.293832E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.372 | TFLOPs: 43.73 | 31: iteration 27990/ 52452 | consumed samples: 14330880 | consumed tokens: 29349642240 | elapsed time per iteration (s): 1.81 | learning rate: 1.018E-04 | global batch size: 512 | lm loss: 1.294548E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.482 | TFLOPs: 42.40 | 0: [2022-12-26 04:43:13,575] [INFO] [logging.py:68:log_dist] [Rank 0] step=28000, skipped=0, lr=[0.0001017803546837164, 0.0001017803546837164, 0.0001017803546837164], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 28000/ 52452 | consumed samples: 14336000 | consumed tokens: 29360128000 | elapsed time per iteration (s): 1.79 | learning rate: 1.018E-04 | global batch size: 512 | lm loss: 1.302596E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.264 | TFLOPs: 42.82 | 0: steps: 28000 loss: 1.2971 iter time (s): 1.801 samples/sec: 284.324 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28000 | lm loss value: 4.162844E+00 | lm loss PPL: 6.425398E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28000 | lm loss value: 4.082990E+00 | lm loss PPL: 5.932256E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28010/ 52452 | consumed samples: 14341120 | consumed tokens: 29370613760 | elapsed time per iteration (s): 9.15 | learning rate: 1.017E-04 | global batch size: 512 | lm loss: 1.306989E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.952 | TFLOPs: 8.40 | 31: iteration 28020/ 52452 | consumed samples: 14346240 | consumed tokens: 29381099520 | elapsed time per iteration (s): 1.79 | learning rate: 1.017E-04 | global batch size: 512 | lm loss: 1.300491E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.874 | TFLOPs: 42.91 | 31: iteration 28030/ 52452 | consumed samples: 14351360 | consumed tokens: 29391585280 | elapsed time per iteration (s): 1.94 | learning rate: 1.016E-04 | global batch size: 512 | lm loss: 1.301063E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 264.127 | TFLOPs: 39.64 | 31: iteration 28040/ 52452 | consumed samples: 14356480 | consumed tokens: 29402071040 | elapsed time per iteration (s): 1.82 | learning rate: 1.016E-04 | global batch size: 512 | lm loss: 1.300126E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.643 | TFLOPs: 42.12 | 31: iteration 28050/ 52452 | consumed samples: 14361600 | consumed tokens: 29412556800 | elapsed time per iteration (s): 1.77 | learning rate: 1.015E-04 | global batch size: 512 | lm loss: 1.293683E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.150 | TFLOPs: 43.40 | 31: iteration 28060/ 52452 | consumed samples: 14366720 | consumed tokens: 29423042560 | elapsed time per iteration (s): 1.82 | learning rate: 1.015E-04 | global batch size: 512 | lm loss: 1.297538E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.709 | TFLOPs: 42.28 | 31: iteration 28070/ 52452 | consumed samples: 14371840 | consumed tokens: 29433528320 | elapsed time per iteration (s): 1.76 | learning rate: 1.014E-04 | global batch size: 512 | lm loss: 1.300910E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.717 | TFLOPs: 43.79 | 31: iteration 28080/ 52452 | consumed samples: 14376960 | consumed tokens: 29444014080 | elapsed time per iteration (s): 1.78 | learning rate: 1.013E-04 | global batch size: 512 | lm loss: 1.285974E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.923 | TFLOPs: 43.22 | 31: iteration 28090/ 52452 | consumed samples: 14382080 | consumed tokens: 29454499840 | elapsed time per iteration (s): 1.79 | learning rate: 1.013E-04 | global batch size: 512 | lm loss: 1.289045E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.352 | TFLOPs: 42.98 | 31: iteration 28100/ 52452 | consumed samples: 14387200 | consumed tokens: 29464985600 | elapsed time per iteration (s): 1.79 | learning rate: 1.012E-04 | global batch size: 512 | lm loss: 1.290582E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.453 | TFLOPs: 42.84 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28100 | lm loss value: 4.171079E+00 | lm loss PPL: 6.478530E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28100 | lm loss value: 4.064213E+00 | lm loss PPL: 5.821909E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28110/ 52452 | consumed samples: 14392320 | consumed tokens: 29475471360 | elapsed time per iteration (s): 9.16 | learning rate: 1.012E-04 | global batch size: 512 | lm loss: 1.295475E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.919 | TFLOPs: 8.39 | 31: iteration 28120/ 52452 | consumed samples: 14397440 | consumed tokens: 29485957120 | elapsed time per iteration (s): 1.82 | learning rate: 1.011E-04 | global batch size: 512 | lm loss: 1.287624E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.082 | TFLOPs: 42.34 | 31: iteration 28130/ 52452 | consumed samples: 14402560 | consumed tokens: 29496442880 | elapsed time per iteration (s): 1.84 | learning rate: 1.011E-04 | global batch size: 512 | lm loss: 1.288221E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.626 | TFLOPs: 41.67 | 31: iteration 28140/ 52452 | consumed samples: 14407680 | consumed tokens: 29506928640 | elapsed time per iteration (s): 1.86 | learning rate: 1.010E-04 | global batch size: 512 | lm loss: 1.291230E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.635 | TFLOPs: 41.37 | 31: iteration 28150/ 52452 | consumed samples: 14412800 | consumed tokens: 29517414400 | elapsed time per iteration (s): 1.75 | learning rate: 1.010E-04 | global batch size: 512 | lm loss: 1.286257E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.063 | TFLOPs: 43.84 | 31: iteration 28160/ 52452 | consumed samples: 14417920 | consumed tokens: 29527900160 | elapsed time per iteration (s): 1.84 | learning rate: 1.009E-04 | global batch size: 512 | lm loss: 1.296938E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.206 | TFLOPs: 41.76 | 31: iteration 28170/ 52452 | consumed samples: 14423040 | consumed tokens: 29538385920 | elapsed time per iteration (s): 1.79 | learning rate: 1.009E-04 | global batch size: 512 | lm loss: 1.300976E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.946 | TFLOPs: 42.92 | 31: iteration 28180/ 52452 | consumed samples: 14428160 | consumed tokens: 29548871680 | elapsed time per iteration (s): 1.73 | learning rate: 1.008E-04 | global batch size: 512 | lm loss: 1.274013E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.779 | TFLOPs: 44.39 | 31: iteration 28190/ 52452 | consumed samples: 14433280 | consumed tokens: 29559357440 | elapsed time per iteration (s): 1.74 | learning rate: 1.008E-04 | global batch size: 512 | lm loss: 1.280097E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.712 | TFLOPs: 44.23 | 31: iteration 28200/ 52452 | consumed samples: 14438400 | consumed tokens: 29569843200 | elapsed time per iteration (s): 1.73 | learning rate: 1.007E-04 | global batch size: 512 | lm loss: 1.292127E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.113 | TFLOPs: 44.29 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28200 | lm loss value: 4.162775E+00 | lm loss PPL: 6.424957E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28200 | lm loss value: 4.059646E+00 | lm loss PPL: 5.795380E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28210/ 52452 | consumed samples: 14443520 | consumed tokens: 29580328960 | elapsed time per iteration (s): 9.21 | learning rate: 1.006E-04 | global batch size: 512 | lm loss: 1.294711E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.580 | TFLOPs: 8.34 | 31: iteration 28220/ 52452 | consumed samples: 14448640 | consumed tokens: 29590814720 | elapsed time per iteration (s): 1.74 | learning rate: 1.006E-04 | global batch size: 512 | lm loss: 1.297812E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.449 | TFLOPs: 44.05 | 31: iteration 28230/ 52452 | consumed samples: 14453760 | consumed tokens: 29601300480 | elapsed time per iteration (s): 1.78 | learning rate: 1.005E-04 | global batch size: 512 | lm loss: 1.304016E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.942 | TFLOPs: 43.07 | 31: iteration 28240/ 52452 | consumed samples: 14458880 | consumed tokens: 29611786240 | elapsed time per iteration (s): 1.79 | learning rate: 1.005E-04 | global batch size: 512 | lm loss: 1.292583E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.753 | TFLOPs: 43.04 | 31: iteration 28250/ 52452 | consumed samples: 14464000 | consumed tokens: 29622272000 | elapsed time per iteration (s): 1.75 | learning rate: 1.004E-04 | global batch size: 512 | lm loss: 1.297168E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.835 | TFLOPs: 43.95 | 31: iteration 28260/ 52452 | consumed samples: 14469120 | consumed tokens: 29632757760 | elapsed time per iteration (s): 1.80 | learning rate: 1.004E-04 | global batch size: 512 | lm loss: 1.294041E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.663 | TFLOPs: 42.58 | 31: iteration 28270/ 52452 | consumed samples: 14474240 | consumed tokens: 29643243520 | elapsed time per iteration (s): 1.77 | learning rate: 1.003E-04 | global batch size: 512 | lm loss: 1.295107E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.758 | TFLOPs: 43.34 | 31: iteration 28280/ 52452 | consumed samples: 14479360 | consumed tokens: 29653729280 | elapsed time per iteration (s): 1.82 | learning rate: 1.003E-04 | global batch size: 512 | lm loss: 1.293159E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.631 | TFLOPs: 42.27 | 31: iteration 28290/ 52452 | consumed samples: 14484480 | consumed tokens: 29664215040 | elapsed time per iteration (s): 1.75 | learning rate: 1.002E-04 | global batch size: 512 | lm loss: 1.289629E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.374 | TFLOPs: 43.88 | 31: iteration 28300/ 52452 | consumed samples: 14489600 | consumed tokens: 29674700800 | elapsed time per iteration (s): 1.74 | learning rate: 1.002E-04 | global batch size: 512 | lm loss: 1.295389E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.726 | TFLOPs: 44.09 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28300 | lm loss value: 4.058315E+00 | lm loss PPL: 5.787670E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28300 | lm loss value: 3.988333E+00 | lm loss PPL: 5.396483E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28310/ 52452 | consumed samples: 14494720 | consumed tokens: 29685186560 | elapsed time per iteration (s): 9.17 | learning rate: 1.001E-04 | global batch size: 512 | lm loss: 1.296201E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.808 | TFLOPs: 8.38 | 31: iteration 28320/ 52452 | consumed samples: 14499840 | consumed tokens: 29695672320 | elapsed time per iteration (s): 1.75 | learning rate: 1.000E-04 | global batch size: 512 | lm loss: 1.291106E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.987 | TFLOPs: 43.98 | 31: iteration 28330/ 52452 | consumed samples: 14504960 | consumed tokens: 29706158080 | elapsed time per iteration (s): 1.77 | learning rate: 9.999E-05 | global batch size: 512 | lm loss: 1.289756E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.523 | TFLOPs: 43.31 | 31: iteration 28340/ 52452 | consumed samples: 14510080 | consumed tokens: 29716643840 | elapsed time per iteration (s): 1.77 | learning rate: 9.994E-05 | global batch size: 512 | lm loss: 1.279443E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.609 | TFLOPs: 43.47 | 31: iteration 28350/ 52452 | consumed samples: 14515200 | consumed tokens: 29727129600 | elapsed time per iteration (s): 1.77 | learning rate: 9.988E-05 | global batch size: 512 | lm loss: 1.298042E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.789 | TFLOPs: 43.50 | 31: iteration 28360/ 52452 | consumed samples: 14520320 | consumed tokens: 29737615360 | elapsed time per iteration (s): 1.80 | learning rate: 9.983E-05 | global batch size: 512 | lm loss: 1.289765E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.743 | TFLOPs: 42.74 | 31: iteration 28370/ 52452 | consumed samples: 14525440 | consumed tokens: 29748101120 | elapsed time per iteration (s): 1.73 | learning rate: 9.978E-05 | global batch size: 512 | lm loss: 1.281502E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.564 | TFLOPs: 44.36 | 31: iteration 28380/ 52452 | consumed samples: 14530560 | consumed tokens: 29758586880 | elapsed time per iteration (s): 1.76 | learning rate: 9.972E-05 | global batch size: 512 | lm loss: 1.305863E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.431 | TFLOPs: 43.59 | 31: iteration 28390/ 52452 | consumed samples: 14535680 | consumed tokens: 29769072640 | elapsed time per iteration (s): 1.79 | learning rate: 9.967E-05 | global batch size: 512 | lm loss: 1.298788E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.567 | TFLOPs: 42.86 | 31: iteration 28400/ 52452 | consumed samples: 14540800 | consumed tokens: 29779558400 | elapsed time per iteration (s): 1.75 | learning rate: 9.961E-05 | global batch size: 512 | lm loss: 1.291456E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.228 | TFLOPs: 43.86 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28400 | lm loss value: 4.122023E+00 | lm loss PPL: 6.168391E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28400 | lm loss value: 4.054214E+00 | lm loss PPL: 5.763984E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28410/ 52452 | consumed samples: 14545920 | consumed tokens: 29790044160 | elapsed time per iteration (s): 9.14 | learning rate: 9.956E-05 | global batch size: 512 | lm loss: 1.280573E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.033 | TFLOPs: 8.41 | 31: iteration 28420/ 52452 | consumed samples: 14551040 | consumed tokens: 29800529920 | elapsed time per iteration (s): 1.80 | learning rate: 9.951E-05 | global batch size: 512 | lm loss: 1.290448E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.479 | TFLOPs: 42.70 | 31: iteration 28430/ 52452 | consumed samples: 14556160 | consumed tokens: 29811015680 | elapsed time per iteration (s): 1.82 | learning rate: 9.945E-05 | global batch size: 512 | lm loss: 1.295795E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.858 | TFLOPs: 42.16 | 31: iteration 28440/ 52452 | consumed samples: 14561280 | consumed tokens: 29821501440 | elapsed time per iteration (s): 1.73 | learning rate: 9.940E-05 | global batch size: 512 | lm loss: 1.290662E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.221 | TFLOPs: 44.31 | 31: iteration 28450/ 52452 | consumed samples: 14566400 | consumed tokens: 29831987200 | elapsed time per iteration (s): 1.80 | learning rate: 9.934E-05 | global batch size: 512 | lm loss: 1.281737E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.608 | TFLOPs: 42.72 | 31: iteration 28460/ 52452 | consumed samples: 14571520 | consumed tokens: 29842472960 | elapsed time per iteration (s): 1.74 | learning rate: 9.929E-05 | global batch size: 512 | lm loss: 1.290539E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.721 | TFLOPs: 44.24 | 31: iteration 28470/ 52452 | consumed samples: 14576640 | consumed tokens: 29852958720 | elapsed time per iteration (s): 1.79 | learning rate: 9.924E-05 | global batch size: 512 | lm loss: 1.292626E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.764 | TFLOPs: 42.89 | 31: iteration 28480/ 52452 | consumed samples: 14581760 | consumed tokens: 29863444480 | elapsed time per iteration (s): 1.81 | learning rate: 9.918E-05 | global batch size: 512 | lm loss: 1.273404E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.348 | TFLOPs: 42.38 | 31: iteration 28490/ 52452 | consumed samples: 14586880 | consumed tokens: 29873930240 | elapsed time per iteration (s): 1.76 | learning rate: 9.913E-05 | global batch size: 512 | lm loss: 1.286663E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.155 | TFLOPs: 43.55 | 31: iteration 28500/ 52452 | consumed samples: 14592000 | consumed tokens: 29884416000 | elapsed time per iteration (s): 1.75 | learning rate: 9.907E-05 | global batch size: 512 | lm loss: 1.297505E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.970 | TFLOPs: 43.82 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28500 | lm loss value: 4.250025E+00 | lm loss PPL: 7.010715E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28500 | lm loss value: 4.155013E+00 | lm loss PPL: 6.375277E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28510/ 52452 | consumed samples: 14597120 | consumed tokens: 29894901760 | elapsed time per iteration (s): 9.15 | learning rate: 9.902E-05 | global batch size: 512 | lm loss: 1.286475E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.978 | TFLOPs: 8.40 | 31: iteration 28520/ 52452 | consumed samples: 14602240 | consumed tokens: 29905387520 | elapsed time per iteration (s): 1.75 | learning rate: 9.897E-05 | global batch size: 512 | lm loss: 1.289736E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.764 | TFLOPs: 43.94 | 31: iteration 28530/ 52452 | consumed samples: 14607360 | consumed tokens: 29915873280 | elapsed time per iteration (s): 1.79 | learning rate: 9.891E-05 | global batch size: 512 | lm loss: 1.293390E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.662 | TFLOPs: 42.88 | 31: iteration 28540/ 52452 | consumed samples: 14612480 | consumed tokens: 29926359040 | elapsed time per iteration (s): 1.79 | learning rate: 9.886E-05 | global batch size: 512 | lm loss: 1.291727E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.455 | TFLOPs: 43.00 | 31: iteration 28550/ 52452 | consumed samples: 14617600 | consumed tokens: 29936844800 | elapsed time per iteration (s): 1.76 | learning rate: 9.880E-05 | global batch size: 512 | lm loss: 1.294240E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.907 | TFLOPs: 43.66 | 31: iteration 28560/ 52452 | consumed samples: 14622720 | consumed tokens: 29947330560 | elapsed time per iteration (s): 1.81 | learning rate: 9.875E-05 | global batch size: 512 | lm loss: 1.295622E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.127 | TFLOPs: 42.35 | 31: iteration 28570/ 52452 | consumed samples: 14627840 | consumed tokens: 29957816320 | elapsed time per iteration (s): 1.77 | learning rate: 9.870E-05 | global batch size: 512 | lm loss: 1.288914E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.197 | TFLOPs: 43.41 | 31: iteration 28580/ 52452 | consumed samples: 14632960 | consumed tokens: 29968302080 | elapsed time per iteration (s): 1.85 | learning rate: 9.864E-05 | global batch size: 512 | lm loss: 1.288773E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.008 | TFLOPs: 41.58 | 31: iteration 28590/ 52452 | consumed samples: 14638080 | consumed tokens: 29978787840 | elapsed time per iteration (s): 1.84 | learning rate: 9.859E-05 | global batch size: 512 | lm loss: 1.294671E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.613 | TFLOPs: 41.67 | 31: iteration 28600/ 52452 | consumed samples: 14643200 | consumed tokens: 29989273600 | elapsed time per iteration (s): 1.89 | learning rate: 9.853E-05 | global batch size: 512 | lm loss: 1.295519E+00 | grad norm: 0.228 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.305 | TFLOPs: 40.72 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28600 | lm loss value: 4.200669E+00 | lm loss PPL: 6.673095E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28600 | lm loss value: 4.114905E+00 | lm loss PPL: 6.124639E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28610/ 52452 | consumed samples: 14648320 | consumed tokens: 29999759360 | elapsed time per iteration (s): 9.22 | learning rate: 9.848E-05 | global batch size: 512 | lm loss: 1.298092E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.536 | TFLOPs: 8.34 | 31: iteration 28620/ 52452 | consumed samples: 14653440 | consumed tokens: 30010245120 | elapsed time per iteration (s): 1.79 | learning rate: 9.843E-05 | global batch size: 512 | lm loss: 1.286444E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.544 | TFLOPs: 42.86 | 31: iteration 28630/ 52452 | consumed samples: 14658560 | consumed tokens: 30020730880 | elapsed time per iteration (s): 1.84 | learning rate: 9.837E-05 | global batch size: 512 | lm loss: 1.288860E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.739 | TFLOPs: 41.84 | 31: iteration 28640/ 52452 | consumed samples: 14663680 | consumed tokens: 30031216640 | elapsed time per iteration (s): 1.78 | learning rate: 9.832E-05 | global batch size: 512 | lm loss: 1.277449E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.343 | TFLOPs: 43.13 | 31: iteration 28650/ 52452 | consumed samples: 14668800 | consumed tokens: 30041702400 | elapsed time per iteration (s): 1.80 | learning rate: 9.826E-05 | global batch size: 512 | lm loss: 1.279698E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.220 | TFLOPs: 42.81 | 31: iteration 28660/ 52452 | consumed samples: 14673920 | consumed tokens: 30052188160 | elapsed time per iteration (s): 1.75 | learning rate: 9.821E-05 | global batch size: 512 | lm loss: 1.282349E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.882 | TFLOPs: 43.96 | 31: iteration 28670/ 52452 | consumed samples: 14679040 | consumed tokens: 30062673920 | elapsed time per iteration (s): 1.78 | learning rate: 9.816E-05 | global batch size: 512 | lm loss: 1.296084E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.539 | TFLOPs: 43.16 | 31: iteration 28680/ 52452 | consumed samples: 14684160 | consumed tokens: 30073159680 | elapsed time per iteration (s): 1.82 | learning rate: 9.810E-05 | global batch size: 512 | lm loss: 1.294319E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.239 | TFLOPs: 42.21 | 31: iteration 28690/ 52452 | consumed samples: 14689280 | consumed tokens: 30083645440 | elapsed time per iteration (s): 1.76 | learning rate: 9.805E-05 | global batch size: 512 | lm loss: 1.296642E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.683 | TFLOPs: 43.78 | 31: iteration 28700/ 52452 | consumed samples: 14694400 | consumed tokens: 30094131200 | elapsed time per iteration (s): 1.87 | learning rate: 9.799E-05 | global batch size: 512 | lm loss: 1.290351E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.224 | TFLOPs: 41.01 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28700 | lm loss value: 4.154287E+00 | lm loss PPL: 6.370655E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28700 | lm loss value: 4.074378E+00 | lm loss PPL: 5.881392E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28710/ 52452 | consumed samples: 14699520 | consumed tokens: 30104616960 | elapsed time per iteration (s): 9.21 | learning rate: 9.794E-05 | global batch size: 512 | lm loss: 1.293235E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.616 | TFLOPs: 8.35 | 31: iteration 28720/ 52452 | consumed samples: 14704640 | consumed tokens: 30115102720 | elapsed time per iteration (s): 1.73 | learning rate: 9.789E-05 | global batch size: 512 | lm loss: 1.279917E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.414 | TFLOPs: 44.34 | 31: iteration 28730/ 52452 | consumed samples: 14709760 | consumed tokens: 30125588480 | elapsed time per iteration (s): 1.79 | learning rate: 9.783E-05 | global batch size: 512 | lm loss: 1.304239E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.501 | TFLOPs: 42.85 | 31: iteration 28740/ 52452 | consumed samples: 14714880 | consumed tokens: 30136074240 | elapsed time per iteration (s): 1.75 | learning rate: 9.778E-05 | global batch size: 512 | lm loss: 1.287422E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.052 | TFLOPs: 43.99 | 31: iteration 28750/ 52452 | consumed samples: 14720000 | consumed tokens: 30146560000 | elapsed time per iteration (s): 1.76 | learning rate: 9.772E-05 | global batch size: 512 | lm loss: 1.288144E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.472 | TFLOPs: 43.60 | 31: iteration 28760/ 52452 | consumed samples: 14725120 | consumed tokens: 30157045760 | elapsed time per iteration (s): 1.75 | learning rate: 9.767E-05 | global batch size: 512 | lm loss: 1.292673E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.091 | TFLOPs: 43.84 | 31: iteration 28770/ 52452 | consumed samples: 14730240 | consumed tokens: 30167531520 | elapsed time per iteration (s): 1.73 | learning rate: 9.762E-05 | global batch size: 512 | lm loss: 1.296508E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.671 | TFLOPs: 44.38 | 31: iteration 28780/ 52452 | consumed samples: 14735360 | consumed tokens: 30178017280 | elapsed time per iteration (s): 1.81 | learning rate: 9.756E-05 | global batch size: 512 | lm loss: 1.308274E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.110 | TFLOPs: 42.34 | 31: iteration 28790/ 52452 | consumed samples: 14740480 | consumed tokens: 30188503040 | elapsed time per iteration (s): 1.80 | learning rate: 9.751E-05 | global batch size: 512 | lm loss: 1.291730E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.648 | TFLOPs: 42.72 | 31: iteration 28800/ 52452 | consumed samples: 14745600 | consumed tokens: 30198988800 | elapsed time per iteration (s): 1.76 | learning rate: 9.745E-05 | global batch size: 512 | lm loss: 1.286831E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.258 | TFLOPs: 43.72 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28800 | lm loss value: 4.127742E+00 | lm loss PPL: 6.203770E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28800 | lm loss value: 4.038335E+00 | lm loss PPL: 5.673180E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28810/ 52452 | consumed samples: 14750720 | consumed tokens: 30209474560 | elapsed time per iteration (s): 9.19 | learning rate: 9.740E-05 | global batch size: 512 | lm loss: 1.286394E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.737 | TFLOPs: 8.37 | 31: iteration 28820/ 52452 | consumed samples: 14755840 | consumed tokens: 30219960320 | elapsed time per iteration (s): 1.87 | learning rate: 9.735E-05 | global batch size: 512 | lm loss: 1.288356E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.291 | TFLOPs: 41.17 | 31: iteration 28830/ 52452 | consumed samples: 14760960 | consumed tokens: 30230446080 | elapsed time per iteration (s): 1.83 | learning rate: 9.729E-05 | global batch size: 512 | lm loss: 1.284603E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.359 | TFLOPs: 42.08 | 31: iteration 28840/ 52452 | consumed samples: 14766080 | consumed tokens: 30240931840 | elapsed time per iteration (s): 1.83 | learning rate: 9.724E-05 | global batch size: 512 | lm loss: 1.298328E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.033 | TFLOPs: 42.03 | 31: iteration 28850/ 52452 | consumed samples: 14771200 | consumed tokens: 30251417600 | elapsed time per iteration (s): 1.87 | learning rate: 9.718E-05 | global batch size: 512 | lm loss: 1.283352E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.027 | TFLOPs: 41.13 | 31: iteration 28860/ 52452 | consumed samples: 14776320 | consumed tokens: 30261903360 | elapsed time per iteration (s): 1.80 | learning rate: 9.713E-05 | global batch size: 512 | lm loss: 1.268827E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.221 | TFLOPs: 42.81 | 31: iteration 28870/ 52452 | consumed samples: 14781440 | consumed tokens: 30272389120 | elapsed time per iteration (s): 1.81 | learning rate: 9.708E-05 | global batch size: 512 | lm loss: 1.301299E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.475 | TFLOPs: 42.40 | 31: iteration 28880/ 52452 | consumed samples: 14786560 | consumed tokens: 30282874880 | elapsed time per iteration (s): 1.77 | learning rate: 9.702E-05 | global batch size: 512 | lm loss: 1.286982E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.336 | TFLOPs: 43.43 | 31: iteration 28890/ 52452 | consumed samples: 14791680 | consumed tokens: 30293360640 | elapsed time per iteration (s): 1.75 | learning rate: 9.697E-05 | global batch size: 512 | lm loss: 1.283575E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.462 | TFLOPs: 43.90 | 31: iteration 28900/ 52452 | consumed samples: 14796800 | consumed tokens: 30303846400 | elapsed time per iteration (s): 1.75 | learning rate: 9.691E-05 | global batch size: 512 | lm loss: 1.278687E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.019 | TFLOPs: 43.83 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 28900 | lm loss value: 4.193016E+00 | lm loss PPL: 6.622222E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 28900 | lm loss value: 4.089899E+00 | lm loss PPL: 5.973383E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 28910/ 52452 | consumed samples: 14801920 | consumed tokens: 30314332160 | elapsed time per iteration (s): 9.16 | learning rate: 9.686E-05 | global batch size: 512 | lm loss: 1.295661E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.880 | TFLOPs: 8.39 | 31: iteration 28920/ 52452 | consumed samples: 14807040 | consumed tokens: 30324817920 | elapsed time per iteration (s): 1.78 | learning rate: 9.681E-05 | global batch size: 512 | lm loss: 1.299881E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.211 | TFLOPs: 43.26 | 31: iteration 28930/ 52452 | consumed samples: 14812160 | consumed tokens: 30335303680 | elapsed time per iteration (s): 1.79 | learning rate: 9.675E-05 | global batch size: 512 | lm loss: 1.289649E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.363 | TFLOPs: 42.98 | 31: iteration 28940/ 52452 | consumed samples: 14817280 | consumed tokens: 30345789440 | elapsed time per iteration (s): 1.75 | learning rate: 9.670E-05 | global batch size: 512 | lm loss: 1.294112E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.217 | TFLOPs: 43.86 | 31: iteration 28950/ 52452 | consumed samples: 14822400 | consumed tokens: 30356275200 | elapsed time per iteration (s): 1.75 | learning rate: 9.665E-05 | global batch size: 512 | lm loss: 1.280874E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.744 | TFLOPs: 43.79 | 31: iteration 28960/ 52452 | consumed samples: 14827520 | consumed tokens: 30366760960 | elapsed time per iteration (s): 1.78 | learning rate: 9.659E-05 | global batch size: 512 | lm loss: 1.290434E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.631 | TFLOPs: 43.17 | 31: iteration 28970/ 52452 | consumed samples: 14832640 | consumed tokens: 30377246720 | elapsed time per iteration (s): 1.79 | learning rate: 9.654E-05 | global batch size: 512 | lm loss: 1.281111E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.793 | TFLOPs: 42.90 | 31: iteration 28980/ 52452 | consumed samples: 14837760 | consumed tokens: 30387732480 | elapsed time per iteration (s): 1.79 | learning rate: 9.648E-05 | global batch size: 512 | lm loss: 1.287703E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.717 | TFLOPs: 42.88 | 31: iteration 28990/ 52452 | consumed samples: 14842880 | consumed tokens: 30398218240 | elapsed time per iteration (s): 1.75 | learning rate: 9.643E-05 | global batch size: 512 | lm loss: 1.290769E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.777 | TFLOPs: 43.79 | 31: iteration 29000/ 52452 | consumed samples: 14848000 | consumed tokens: 30408704000 | elapsed time per iteration (s): 1.88 | learning rate: 9.638E-05 | global batch size: 512 | lm loss: 1.297893E+00 | grad norm: 0.183 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.845 | TFLOPs: 40.95 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29000 | lm loss value: 4.238431E+00 | lm loss PPL: 6.929903E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29000 | lm loss value: 4.129549E+00 | lm loss PPL: 6.214989E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29010/ 52452 | consumed samples: 14853120 | consumed tokens: 30419189760 | elapsed time per iteration (s): 9.17 | learning rate: 9.632E-05 | global batch size: 512 | lm loss: 1.275928E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.842 | TFLOPs: 8.38 | 31: iteration 29020/ 52452 | consumed samples: 14858240 | consumed tokens: 30429675520 | elapsed time per iteration (s): 1.79 | learning rate: 9.627E-05 | global batch size: 512 | lm loss: 1.278660E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.288 | TFLOPs: 42.82 | 31: iteration 29030/ 52452 | consumed samples: 14863360 | consumed tokens: 30440161280 | elapsed time per iteration (s): 1.84 | learning rate: 9.622E-05 | global batch size: 512 | lm loss: 1.298911E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.637 | TFLOPs: 41.82 | 31: iteration 29040/ 52452 | consumed samples: 14868480 | consumed tokens: 30450647040 | elapsed time per iteration (s): 1.74 | learning rate: 9.616E-05 | global batch size: 512 | lm loss: 1.295380E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.772 | TFLOPs: 44.24 | 31: iteration 29050/ 52452 | consumed samples: 14873600 | consumed tokens: 30461132800 | elapsed time per iteration (s): 1.78 | learning rate: 9.611E-05 | global batch size: 512 | lm loss: 1.282101E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.034 | TFLOPs: 43.23 | 31: iteration 29060/ 52452 | consumed samples: 14878720 | consumed tokens: 30471618560 | elapsed time per iteration (s): 1.80 | learning rate: 9.605E-05 | global batch size: 512 | lm loss: 1.281590E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.090 | TFLOPs: 42.64 | 31: iteration 29070/ 52452 | consumed samples: 14883840 | consumed tokens: 30482104320 | elapsed time per iteration (s): 1.82 | learning rate: 9.600E-05 | global batch size: 512 | lm loss: 1.277699E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.574 | TFLOPs: 42.11 | 31: iteration 29080/ 52452 | consumed samples: 14888960 | consumed tokens: 30492590080 | elapsed time per iteration (s): 1.77 | learning rate: 9.595E-05 | global batch size: 512 | lm loss: 1.293144E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.535 | TFLOPs: 43.46 | 31: iteration 29090/ 52452 | consumed samples: 14894080 | consumed tokens: 30503075840 | elapsed time per iteration (s): 1.98 | learning rate: 9.589E-05 | global batch size: 512 | lm loss: 1.293592E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 258.663 | TFLOPs: 38.82 | 31: iteration 29100/ 52452 | consumed samples: 14899200 | consumed tokens: 30513561600 | elapsed time per iteration (s): 1.81 | learning rate: 9.584E-05 | global batch size: 512 | lm loss: 1.292664E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.984 | TFLOPs: 42.47 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29100 | lm loss value: 4.155223E+00 | lm loss PPL: 6.376621E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29100 | lm loss value: 4.053699E+00 | lm loss PPL: 5.761014E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29110/ 52452 | consumed samples: 14904320 | consumed tokens: 30524047360 | elapsed time per iteration (s): 9.15 | learning rate: 9.578E-05 | global batch size: 512 | lm loss: 1.282765E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.954 | TFLOPs: 8.40 | 31: iteration 29120/ 52452 | consumed samples: 14909440 | consumed tokens: 30534533120 | elapsed time per iteration (s): 1.80 | learning rate: 9.573E-05 | global batch size: 512 | lm loss: 1.285770E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.305 | TFLOPs: 42.67 | 31: iteration 29130/ 52452 | consumed samples: 14914560 | consumed tokens: 30545018880 | elapsed time per iteration (s): 1.93 | learning rate: 9.568E-05 | global batch size: 512 | lm loss: 1.278966E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.386 | TFLOPs: 39.83 | 31: iteration 29140/ 52452 | consumed samples: 14919680 | consumed tokens: 30555504640 | elapsed time per iteration (s): 1.78 | learning rate: 9.562E-05 | global batch size: 512 | lm loss: 1.290480E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.185 | TFLOPs: 43.10 | 31: iteration 29150/ 52452 | consumed samples: 14924800 | consumed tokens: 30565990400 | elapsed time per iteration (s): 1.80 | learning rate: 9.557E-05 | global batch size: 512 | lm loss: 1.281447E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.002 | TFLOPs: 42.78 | 31: iteration 29160/ 52452 | consumed samples: 14929920 | consumed tokens: 30576476160 | elapsed time per iteration (s): 1.73 | learning rate: 9.552E-05 | global batch size: 512 | lm loss: 1.288400E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.017 | TFLOPs: 44.43 | 31: iteration 29170/ 52452 | consumed samples: 14935040 | consumed tokens: 30586961920 | elapsed time per iteration (s): 1.78 | learning rate: 9.546E-05 | global batch size: 512 | lm loss: 1.278608E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.133 | TFLOPs: 43.25 | 31: iteration 29180/ 52452 | consumed samples: 14940160 | consumed tokens: 30597447680 | elapsed time per iteration (s): 1.79 | learning rate: 9.541E-05 | global batch size: 512 | lm loss: 1.295565E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.732 | TFLOPs: 43.04 | 31: iteration 29190/ 52452 | consumed samples: 14945280 | consumed tokens: 30607933440 | elapsed time per iteration (s): 1.75 | learning rate: 9.535E-05 | global batch size: 512 | lm loss: 1.273680E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.829 | TFLOPs: 43.80 | 31: iteration 29200/ 52452 | consumed samples: 14950400 | consumed tokens: 30618419200 | elapsed time per iteration (s): 1.87 | learning rate: 9.530E-05 | global batch size: 512 | lm loss: 1.283333E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.294 | TFLOPs: 41.17 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29200 | lm loss value: 4.132388E+00 | lm loss PPL: 6.232656E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29200 | lm loss value: 4.044029E+00 | lm loss PPL: 5.705577E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29210/ 52452 | consumed samples: 14955520 | consumed tokens: 30628904960 | elapsed time per iteration (s): 9.10 | learning rate: 9.525E-05 | global batch size: 512 | lm loss: 1.277379E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.235 | TFLOPs: 8.44 | 31: iteration 29220/ 52452 | consumed samples: 14960640 | consumed tokens: 30639390720 | elapsed time per iteration (s): 1.78 | learning rate: 9.519E-05 | global batch size: 512 | lm loss: 1.291943E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.167 | TFLOPs: 43.10 | 31: iteration 29230/ 52452 | consumed samples: 14965760 | consumed tokens: 30649876480 | elapsed time per iteration (s): 1.74 | learning rate: 9.514E-05 | global batch size: 512 | lm loss: 1.280644E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.533 | TFLOPs: 44.06 | 31: iteration 29240/ 52452 | consumed samples: 14970880 | consumed tokens: 30660362240 | elapsed time per iteration (s): 1.80 | learning rate: 9.509E-05 | global batch size: 512 | lm loss: 1.293425E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.858 | TFLOPs: 42.61 | 31: iteration 29250/ 52452 | consumed samples: 14976000 | consumed tokens: 30670848000 | elapsed time per iteration (s): 1.77 | learning rate: 9.503E-05 | global batch size: 512 | lm loss: 1.277766E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.966 | TFLOPs: 43.37 | 31: iteration 29260/ 52452 | consumed samples: 14981120 | consumed tokens: 30681333760 | elapsed time per iteration (s): 1.79 | learning rate: 9.498E-05 | global batch size: 512 | lm loss: 1.284001E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.655 | TFLOPs: 43.03 | 31: iteration 29270/ 52452 | consumed samples: 14986240 | consumed tokens: 30691819520 | elapsed time per iteration (s): 1.76 | learning rate: 9.493E-05 | global batch size: 512 | lm loss: 1.272133E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.635 | TFLOPs: 43.62 | 31: iteration 29280/ 52452 | consumed samples: 14991360 | consumed tokens: 30702305280 | elapsed time per iteration (s): 1.89 | learning rate: 9.487E-05 | global batch size: 512 | lm loss: 1.291418E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.348 | TFLOPs: 40.73 | 31: iteration 29290/ 52452 | consumed samples: 14996480 | consumed tokens: 30712791040 | elapsed time per iteration (s): 1.79 | learning rate: 9.482E-05 | global batch size: 512 | lm loss: 1.293673E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.941 | TFLOPs: 42.92 | 31: iteration 29300/ 52452 | consumed samples: 15001600 | consumed tokens: 30723276800 | elapsed time per iteration (s): 1.78 | learning rate: 9.476E-05 | global batch size: 512 | lm loss: 1.270036E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.668 | TFLOPs: 43.18 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29300 | lm loss value: 4.099809E+00 | lm loss PPL: 6.032875E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29300 | lm loss value: 4.022911E+00 | lm loss PPL: 5.586347E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29310/ 52452 | consumed samples: 15006720 | consumed tokens: 30733762560 | elapsed time per iteration (s): 9.09 | learning rate: 9.471E-05 | global batch size: 512 | lm loss: 1.291830E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.304 | TFLOPs: 8.45 | 31: iteration 29320/ 52452 | consumed samples: 15011840 | consumed tokens: 30744248320 | elapsed time per iteration (s): 1.76 | learning rate: 9.466E-05 | global batch size: 512 | lm loss: 1.277785E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.130 | TFLOPs: 43.70 | 31: iteration 29330/ 52452 | consumed samples: 15016960 | consumed tokens: 30754734080 | elapsed time per iteration (s): 1.78 | learning rate: 9.460E-05 | global batch size: 512 | lm loss: 1.286411E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.159 | TFLOPs: 43.10 | 31: iteration 29340/ 52452 | consumed samples: 15022080 | consumed tokens: 30765219840 | elapsed time per iteration (s): 1.75 | learning rate: 9.455E-05 | global batch size: 512 | lm loss: 1.293252E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.698 | TFLOPs: 43.93 | 31: iteration 29350/ 52452 | consumed samples: 15027200 | consumed tokens: 30775705600 | elapsed time per iteration (s): 1.77 | learning rate: 9.450E-05 | global batch size: 512 | lm loss: 1.272044E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.009 | TFLOPs: 43.38 | 31: iteration 29360/ 52452 | consumed samples: 15032320 | consumed tokens: 30786191360 | elapsed time per iteration (s): 1.79 | learning rate: 9.444E-05 | global batch size: 512 | lm loss: 1.284517E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.816 | TFLOPs: 43.05 | 31: iteration 29370/ 52452 | consumed samples: 15037440 | consumed tokens: 30796677120 | elapsed time per iteration (s): 1.73 | learning rate: 9.439E-05 | global batch size: 512 | lm loss: 1.290119E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.362 | TFLOPs: 44.48 | 31: iteration 29380/ 52452 | consumed samples: 15042560 | consumed tokens: 30807162880 | elapsed time per iteration (s): 1.76 | learning rate: 9.434E-05 | global batch size: 512 | lm loss: 1.280581E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.727 | TFLOPs: 43.64 | 31: iteration 29390/ 52452 | consumed samples: 15047680 | consumed tokens: 30817648640 | elapsed time per iteration (s): 1.74 | learning rate: 9.428E-05 | global batch size: 512 | lm loss: 1.267629E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.649 | TFLOPs: 44.08 | 31: iteration 29400/ 52452 | consumed samples: 15052800 | consumed tokens: 30828134400 | elapsed time per iteration (s): 1.76 | learning rate: 9.423E-05 | global batch size: 512 | lm loss: 1.288288E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.982 | TFLOPs: 43.67 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29400 | lm loss value: 4.179555E+00 | lm loss PPL: 6.533677E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29400 | lm loss value: 4.094753E+00 | lm loss PPL: 6.002453E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29410/ 52452 | consumed samples: 15057920 | consumed tokens: 30838620160 | elapsed time per iteration (s): 9.23 | learning rate: 9.417E-05 | global batch size: 512 | lm loss: 1.265710E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.478 | TFLOPs: 8.33 | 31: iteration 29420/ 52452 | consumed samples: 15063040 | consumed tokens: 30849105920 | elapsed time per iteration (s): 1.76 | learning rate: 9.412E-05 | global batch size: 512 | lm loss: 1.280238E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.186 | TFLOPs: 43.71 | 31: iteration 29430/ 52452 | consumed samples: 15068160 | consumed tokens: 30859591680 | elapsed time per iteration (s): 1.74 | learning rate: 9.407E-05 | global batch size: 512 | lm loss: 1.300543E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.026 | TFLOPs: 44.13 | 31: iteration 29440/ 52452 | consumed samples: 15073280 | consumed tokens: 30870077440 | elapsed time per iteration (s): 1.76 | learning rate: 9.401E-05 | global batch size: 512 | lm loss: 1.278636E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.655 | TFLOPs: 43.78 | 31: iteration 29450/ 52452 | consumed samples: 15078400 | consumed tokens: 30880563200 | elapsed time per iteration (s): 1.75 | learning rate: 9.396E-05 | global batch size: 512 | lm loss: 1.281019E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.541 | TFLOPs: 43.91 | 31: iteration 29460/ 52452 | consumed samples: 15083520 | consumed tokens: 30891048960 | elapsed time per iteration (s): 1.76 | learning rate: 9.391E-05 | global batch size: 512 | lm loss: 1.290228E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.500 | TFLOPs: 43.60 | 31: iteration 29470/ 52452 | consumed samples: 15088640 | consumed tokens: 30901534720 | elapsed time per iteration (s): 1.73 | learning rate: 9.385E-05 | global batch size: 512 | lm loss: 1.284621E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.468 | TFLOPs: 44.35 | 31: iteration 29480/ 52452 | consumed samples: 15093760 | consumed tokens: 30912020480 | elapsed time per iteration (s): 1.72 | learning rate: 9.380E-05 | global batch size: 512 | lm loss: 1.282235E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.691 | TFLOPs: 44.68 | 31: iteration 29490/ 52452 | consumed samples: 15098880 | consumed tokens: 30922506240 | elapsed time per iteration (s): 1.75 | learning rate: 9.375E-05 | global batch size: 512 | lm loss: 1.282407E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.370 | TFLOPs: 43.88 | 31: iteration 29500/ 52452 | consumed samples: 15104000 | consumed tokens: 30932992000 | elapsed time per iteration (s): 1.76 | learning rate: 9.369E-05 | global batch size: 512 | lm loss: 1.282756E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.434 | TFLOPs: 43.74 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29500 | lm loss value: 4.160715E+00 | lm loss PPL: 6.411736E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29500 | lm loss value: 4.072859E+00 | lm loss PPL: 5.872463E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29510/ 52452 | consumed samples: 15109120 | consumed tokens: 30943477760 | elapsed time per iteration (s): 9.10 | learning rate: 9.364E-05 | global batch size: 512 | lm loss: 1.286150E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.252 | TFLOPs: 8.44 | 31: iteration 29520/ 52452 | consumed samples: 15114240 | consumed tokens: 30953963520 | elapsed time per iteration (s): 1.76 | learning rate: 9.358E-05 | global batch size: 512 | lm loss: 1.290564E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.918 | TFLOPs: 43.67 | 31: iteration 29530/ 52452 | consumed samples: 15119360 | consumed tokens: 30964449280 | elapsed time per iteration (s): 1.74 | learning rate: 9.353E-05 | global batch size: 512 | lm loss: 1.270011E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.671 | TFLOPs: 44.08 | 31: iteration 29540/ 52452 | consumed samples: 15124480 | consumed tokens: 30974935040 | elapsed time per iteration (s): 1.76 | learning rate: 9.348E-05 | global batch size: 512 | lm loss: 1.272679E+00 | grad norm: 0.147 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.523 | TFLOPs: 43.61 | 31: iteration 29550/ 52452 | consumed samples: 15129600 | consumed tokens: 30985420800 | elapsed time per iteration (s): 1.74 | learning rate: 9.342E-05 | global batch size: 512 | lm loss: 1.282544E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.065 | TFLOPs: 44.14 | 31: iteration 29560/ 52452 | consumed samples: 15134720 | consumed tokens: 30995906560 | elapsed time per iteration (s): 1.73 | learning rate: 9.337E-05 | global batch size: 512 | lm loss: 1.272317E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.211 | TFLOPs: 44.31 | 31: iteration 29570/ 52452 | consumed samples: 15139840 | consumed tokens: 31006392320 | elapsed time per iteration (s): 1.72 | learning rate: 9.332E-05 | global batch size: 512 | lm loss: 1.276451E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.089 | TFLOPs: 44.59 | 31: iteration 29580/ 52452 | consumed samples: 15144960 | consumed tokens: 31016878080 | elapsed time per iteration (s): 1.76 | learning rate: 9.326E-05 | global batch size: 512 | lm loss: 1.278373E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.764 | TFLOPs: 43.64 | 31: iteration 29590/ 52452 | consumed samples: 15150080 | consumed tokens: 31027363840 | elapsed time per iteration (s): 1.76 | learning rate: 9.321E-05 | global batch size: 512 | lm loss: 1.291358E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.069 | TFLOPs: 43.69 | 31: iteration 29600/ 52452 | consumed samples: 15155200 | consumed tokens: 31037849600 | elapsed time per iteration (s): 1.73 | learning rate: 9.316E-05 | global batch size: 512 | lm loss: 1.274039E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.432 | TFLOPs: 44.49 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29600 | lm loss value: 4.177039E+00 | lm loss PPL: 6.517257E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29600 | lm loss value: 4.058781E+00 | lm loss PPL: 5.790369E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29610/ 52452 | consumed samples: 15160320 | consumed tokens: 31048335360 | elapsed time per iteration (s): 9.17 | learning rate: 9.310E-05 | global batch size: 512 | lm loss: 1.278161E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.832 | TFLOPs: 8.38 | 31: iteration 29620/ 52452 | consumed samples: 15165440 | consumed tokens: 31058821120 | elapsed time per iteration (s): 1.75 | learning rate: 9.305E-05 | global batch size: 512 | lm loss: 1.283766E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.183 | TFLOPs: 43.86 | 31: iteration 29630/ 52452 | consumed samples: 15170560 | consumed tokens: 31069306880 | elapsed time per iteration (s): 1.79 | learning rate: 9.300E-05 | global batch size: 512 | lm loss: 1.277586E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.617 | TFLOPs: 42.87 | 31: iteration 29640/ 52452 | consumed samples: 15175680 | consumed tokens: 31079792640 | elapsed time per iteration (s): 1.79 | learning rate: 9.294E-05 | global batch size: 512 | lm loss: 1.280385E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.910 | TFLOPs: 42.91 | 31: iteration 29650/ 52452 | consumed samples: 15180800 | consumed tokens: 31090278400 | elapsed time per iteration (s): 1.81 | learning rate: 9.289E-05 | global batch size: 512 | lm loss: 1.284020E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.372 | TFLOPs: 42.53 | 31: iteration 29660/ 52452 | consumed samples: 15185920 | consumed tokens: 31100764160 | elapsed time per iteration (s): 1.80 | learning rate: 9.284E-05 | global batch size: 512 | lm loss: 1.290786E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.464 | TFLOPs: 42.70 | 31: iteration 29670/ 52452 | consumed samples: 15191040 | consumed tokens: 31111249920 | elapsed time per iteration (s): 1.76 | learning rate: 9.278E-05 | global batch size: 512 | lm loss: 1.295740E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.122 | TFLOPs: 43.55 | 31: iteration 29680/ 52452 | consumed samples: 15196160 | consumed tokens: 31121735680 | elapsed time per iteration (s): 1.86 | learning rate: 9.273E-05 | global batch size: 512 | lm loss: 1.286083E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.808 | TFLOPs: 41.25 | 31: iteration 29690/ 52452 | consumed samples: 15201280 | consumed tokens: 31132221440 | elapsed time per iteration (s): 1.80 | learning rate: 9.268E-05 | global batch size: 512 | lm loss: 1.272969E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.448 | TFLOPs: 42.69 | 31: iteration 29700/ 52452 | consumed samples: 15206400 | consumed tokens: 31142707200 | elapsed time per iteration (s): 1.76 | learning rate: 9.262E-05 | global batch size: 512 | lm loss: 1.280565E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.479 | TFLOPs: 43.60 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29700 | lm loss value: 4.319361E+00 | lm loss PPL: 7.514058E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29700 | lm loss value: 4.167444E+00 | lm loss PPL: 6.455027E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29710/ 52452 | consumed samples: 15211520 | consumed tokens: 31153192960 | elapsed time per iteration (s): 9.19 | learning rate: 9.257E-05 | global batch size: 512 | lm loss: 1.279119E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.714 | TFLOPs: 8.36 | 31: iteration 29720/ 52452 | consumed samples: 15216640 | consumed tokens: 31163678720 | elapsed time per iteration (s): 1.76 | learning rate: 9.252E-05 | global batch size: 512 | lm loss: 1.287481E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.371 | TFLOPs: 43.73 | 31: iteration 29730/ 52452 | consumed samples: 15221760 | consumed tokens: 31174164480 | elapsed time per iteration (s): 1.74 | learning rate: 9.246E-05 | global batch size: 512 | lm loss: 1.294060E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.797 | TFLOPs: 44.25 | 31: iteration 29740/ 52452 | consumed samples: 15226880 | consumed tokens: 31184650240 | elapsed time per iteration (s): 1.77 | learning rate: 9.241E-05 | global batch size: 512 | lm loss: 1.277136E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.624 | TFLOPs: 43.32 | 31: iteration 29750/ 52452 | consumed samples: 15232000 | consumed tokens: 31195136000 | elapsed time per iteration (s): 1.75 | learning rate: 9.236E-05 | global batch size: 512 | lm loss: 1.277102E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.763 | TFLOPs: 43.94 | 31: iteration 29760/ 52452 | consumed samples: 15237120 | consumed tokens: 31205621760 | elapsed time per iteration (s): 1.74 | learning rate: 9.230E-05 | global batch size: 512 | lm loss: 1.273627E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.971 | TFLOPs: 44.12 | 31: iteration 29770/ 52452 | consumed samples: 15242240 | consumed tokens: 31216107520 | elapsed time per iteration (s): 1.74 | learning rate: 9.225E-05 | global batch size: 512 | lm loss: 1.270973E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.669 | TFLOPs: 44.23 | 31: iteration 29780/ 52452 | consumed samples: 15247360 | consumed tokens: 31226593280 | elapsed time per iteration (s): 1.79 | learning rate: 9.220E-05 | global batch size: 512 | lm loss: 1.280617E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.746 | TFLOPs: 43.04 | 31: iteration 29790/ 52452 | consumed samples: 15252480 | consumed tokens: 31237079040 | elapsed time per iteration (s): 1.75 | learning rate: 9.214E-05 | global batch size: 512 | lm loss: 1.286984E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.660 | TFLOPs: 43.93 | 31: iteration 29800/ 52452 | consumed samples: 15257600 | consumed tokens: 31247564800 | elapsed time per iteration (s): 1.83 | learning rate: 9.209E-05 | global batch size: 512 | lm loss: 1.273277E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.901 | TFLOPs: 42.01 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29800 | lm loss value: 4.346161E+00 | lm loss PPL: 7.718162E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29800 | lm loss value: 4.157650E+00 | lm loss PPL: 6.392110E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29810/ 52452 | consumed samples: 15262720 | consumed tokens: 31258050560 | elapsed time per iteration (s): 9.21 | learning rate: 9.204E-05 | global batch size: 512 | lm loss: 1.281766E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.599 | TFLOPs: 8.35 | 31: iteration 29820/ 52452 | consumed samples: 15267840 | consumed tokens: 31268536320 | elapsed time per iteration (s): 1.75 | learning rate: 9.198E-05 | global batch size: 512 | lm loss: 1.265306E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.681 | TFLOPs: 43.93 | 31: iteration 29830/ 52452 | consumed samples: 15272960 | consumed tokens: 31279022080 | elapsed time per iteration (s): 1.75 | learning rate: 9.193E-05 | global batch size: 512 | lm loss: 1.277523E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.408 | TFLOPs: 43.89 | 31: iteration 29840/ 52452 | consumed samples: 15278080 | consumed tokens: 31289507840 | elapsed time per iteration (s): 1.77 | learning rate: 9.187E-05 | global batch size: 512 | lm loss: 1.273959E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.989 | TFLOPs: 43.38 | 31: iteration 29850/ 52452 | consumed samples: 15283200 | consumed tokens: 31299993600 | elapsed time per iteration (s): 1.82 | learning rate: 9.182E-05 | global batch size: 512 | lm loss: 1.282535E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.951 | TFLOPs: 42.32 | 31: iteration 29860/ 52452 | consumed samples: 15288320 | consumed tokens: 31310479360 | elapsed time per iteration (s): 1.81 | learning rate: 9.177E-05 | global batch size: 512 | lm loss: 1.276576E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.546 | TFLOPs: 42.41 | 31: iteration 29870/ 52452 | consumed samples: 15293440 | consumed tokens: 31320965120 | elapsed time per iteration (s): 1.84 | learning rate: 9.172E-05 | global batch size: 512 | lm loss: 1.280014E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.043 | TFLOPs: 41.73 | 31: iteration 29880/ 52452 | consumed samples: 15298560 | consumed tokens: 31331450880 | elapsed time per iteration (s): 1.77 | learning rate: 9.166E-05 | global batch size: 512 | lm loss: 1.276656E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.697 | TFLOPs: 43.33 | 31: iteration 29890/ 52452 | consumed samples: 15303680 | consumed tokens: 31341936640 | elapsed time per iteration (s): 1.76 | learning rate: 9.161E-05 | global batch size: 512 | lm loss: 1.275202E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.128 | TFLOPs: 43.55 | 31: iteration 29900/ 52452 | consumed samples: 15308800 | consumed tokens: 31352422400 | elapsed time per iteration (s): 1.86 | learning rate: 9.156E-05 | global batch size: 512 | lm loss: 1.272560E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.544 | TFLOPs: 41.21 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 29900 | lm loss value: 4.124497E+00 | lm loss PPL: 6.183672E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 29900 | lm loss value: 4.022892E+00 | lm loss PPL: 5.586240E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 29910/ 52452 | consumed samples: 15313920 | consumed tokens: 31362908160 | elapsed time per iteration (s): 9.14 | learning rate: 9.150E-05 | global batch size: 512 | lm loss: 1.271015E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.020 | TFLOPs: 8.41 | 31: iteration 29920/ 52452 | consumed samples: 15319040 | consumed tokens: 31373393920 | elapsed time per iteration (s): 1.77 | learning rate: 9.145E-05 | global batch size: 512 | lm loss: 1.282956E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.831 | TFLOPs: 43.35 | 31: iteration 29930/ 52452 | consumed samples: 15324160 | consumed tokens: 31383879680 | elapsed time per iteration (s): 1.82 | learning rate: 9.140E-05 | global batch size: 512 | lm loss: 1.278672E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.936 | TFLOPs: 42.32 | 31: iteration 29940/ 52452 | consumed samples: 15329280 | consumed tokens: 31394365440 | elapsed time per iteration (s): 1.77 | learning rate: 9.134E-05 | global batch size: 512 | lm loss: 1.273789E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.666 | TFLOPs: 43.48 | 31: iteration 29950/ 52452 | consumed samples: 15334400 | consumed tokens: 31404851200 | elapsed time per iteration (s): 1.75 | learning rate: 9.129E-05 | global batch size: 512 | lm loss: 1.268118E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.578 | TFLOPs: 43.91 | 31: iteration 29960/ 52452 | consumed samples: 15339520 | consumed tokens: 31415336960 | elapsed time per iteration (s): 1.74 | learning rate: 9.124E-05 | global batch size: 512 | lm loss: 1.269274E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.476 | TFLOPs: 44.05 | 31: iteration 29970/ 52452 | consumed samples: 15344640 | consumed tokens: 31425822720 | elapsed time per iteration (s): 1.78 | learning rate: 9.118E-05 | global batch size: 512 | lm loss: 1.274804E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.995 | TFLOPs: 43.23 | 31: iteration 29980/ 52452 | consumed samples: 15349760 | consumed tokens: 31436308480 | elapsed time per iteration (s): 1.76 | learning rate: 9.113E-05 | global batch size: 512 | lm loss: 1.265019E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.150 | TFLOPs: 43.55 | 31: iteration 29990/ 52452 | consumed samples: 15354880 | consumed tokens: 31446794240 | elapsed time per iteration (s): 1.78 | learning rate: 9.108E-05 | global batch size: 512 | lm loss: 1.273106E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.459 | TFLOPs: 43.15 | 0: [2022-12-26 06:07:15,870] [INFO] [logging.py:68:log_dist] [Rank 0] step=30000, skipped=0, lr=[9.102249347199177e-05, 9.102249347199177e-05, 9.102249347199177e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 30000/ 52452 | consumed samples: 15360000 | consumed tokens: 31457280000 | elapsed time per iteration (s): 1.77 | learning rate: 9.102E-05 | global batch size: 512 | lm loss: 1.291681E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.608 | TFLOPs: 43.32 | 0: steps: 30000 loss: 1.2852 iter time (s): 1.783 samples/sec: 287.214 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30000 | lm loss value: 4.102298E+00 | lm loss PPL: 6.047910E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30000 | lm loss value: 4.017490E+00 | lm loss PPL: 5.556149E+01 | 31: ------------------------------------------------------------------------------------------------------ 0: saving checkpoint at iteration 30000 to checkpoints_2b855brealtasky 0: [2022-12-26 06:08:29,590] [INFO] [logging.py:68:log_dist] [Rank 0] [Torch] Checkpoint global_step30000 is begin to save! 0: [2022-12-26 06:08:29,600] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_01-model_00-model_states.pt... 0: [2022-12-26 06:08:29,933] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_01-model_00-model_states.pt. 0: [2022-12-26 06:08:29,934] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_03-model_00-model_states.pt... 0: [2022-12-26 06:08:30,119] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_03-model_00-model_states.pt. 0: [2022-12-26 06:08:30,119] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_04-model_00-model_states.pt... 0: [2022-12-26 06:08:30,306] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_04-model_00-model_states.pt. 0: [2022-12-26 06:08:30,307] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_05-model_00-model_states.pt... 0: [2022-12-26 06:08:30,481] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_05-model_00-model_states.pt. 0: [2022-12-26 06:08:30,482] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_06-model_00-model_states.pt... 0: [2022-12-26 06:08:30,657] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_06-model_00-model_states.pt. 0: [2022-12-26 06:08:30,657] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_07-model_00-model_states.pt... 0: [2022-12-26 06:08:30,835] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_07-model_00-model_states.pt. 0: [2022-12-26 06:08:30,835] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_08-model_00-model_states.pt... 0: [2022-12-26 06:08:31,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_08-model_00-model_states.pt. 0: [2022-12-26 06:08:31,012] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_09-model_00-model_states.pt... 0: [2022-12-26 06:08:31,184] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_09-model_00-model_states.pt. 0: [2022-12-26 06:08:31,185] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_10-model_00-model_states.pt... 0: [2022-12-26 06:08:31,363] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_10-model_00-model_states.pt. 0: [2022-12-26 06:08:31,364] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_11-model_00-model_states.pt... 0: [2022-12-26 06:08:31,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_11-model_00-model_states.pt. 0: [2022-12-26 06:08:31,531] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_12-model_00-model_states.pt... 0: [2022-12-26 06:08:31,703] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_12-model_00-model_states.pt. 0: [2022-12-26 06:08:31,703] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_13-model_00-model_states.pt... 0: [2022-12-26 06:08:31,873] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_13-model_00-model_states.pt. 0: [2022-12-26 06:08:31,873] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_14-model_00-model_states.pt... 0: [2022-12-26 06:08:32,045] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_14-model_00-model_states.pt. 0: [2022-12-26 06:08:32,045] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_15-model_00-model_states.pt... 0: [2022-12-26 06:08:32,218] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_15-model_00-model_states.pt. 0: [2022-12-26 06:08:32,218] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_16-model_00-model_states.pt... 0: [2022-12-26 06:08:32,384] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_16-model_00-model_states.pt. 0: [2022-12-26 06:08:32,384] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_17-model_00-model_states.pt... 0: [2022-12-26 06:08:32,557] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_17-model_00-model_states.pt. 0: [2022-12-26 06:08:32,557] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_18-model_00-model_states.pt... 0: [2022-12-26 06:08:32,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_18-model_00-model_states.pt. 0: [2022-12-26 06:08:32,727] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_19-model_00-model_states.pt... 0: [2022-12-26 06:08:32,895] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_19-model_00-model_states.pt. 0: [2022-12-26 06:08:32,895] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_20-model_00-model_states.pt... 0: [2022-12-26 06:08:33,062] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_20-model_00-model_states.pt. 0: [2022-12-26 06:08:33,063] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_21-model_00-model_states.pt... 0: [2022-12-26 06:08:33,227] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_21-model_00-model_states.pt. 0: [2022-12-26 06:08:33,227] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_22-model_00-model_states.pt... 0: [2022-12-26 06:08:33,395] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_22-model_00-model_states.pt. 0: [2022-12-26 06:08:33,395] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_23-model_00-model_states.pt... 0: [2022-12-26 06:08:33,562] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_23-model_00-model_states.pt. 0: [2022-12-26 06:08:33,562] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_24-model_00-model_states.pt... 0: [2022-12-26 06:08:33,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_24-model_00-model_states.pt. 0: [2022-12-26 06:08:33,727] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_25-model_00-model_states.pt... 0: [2022-12-26 06:08:33,888] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_25-model_00-model_states.pt. 0: [2022-12-26 06:08:33,889] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_26-model_00-model_states.pt... 0: [2022-12-26 06:08:34,052] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_26-model_00-model_states.pt. 0: [2022-12-26 06:08:34,053] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_27-model_00-model_states.pt... 0: [2022-12-26 06:08:34,222] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_27-model_00-model_states.pt. 0: [2022-12-26 06:08:34,222] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_28-model_00-model_states.pt... 0: [2022-12-26 06:08:34,382] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_28-model_00-model_states.pt. 0: [2022-12-26 06:08:34,383] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_29-model_00-model_states.pt... 0: [2022-12-26 06:08:34,551] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_29-model_00-model_states.pt. 0: [2022-12-26 06:08:34,551] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_30-model_00-model_states.pt... 0: [2022-12-26 06:08:34,720] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_30-model_00-model_states.pt. 0: [2022-12-26 06:08:34,720] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_31-model_00-model_states.pt... 0: [2022-12-26 06:08:34,881] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_31-model_00-model_states.pt. 0: [2022-12-26 06:08:34,881] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_32-model_00-model_states.pt... 0: [2022-12-26 06:08:35,052] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_32-model_00-model_states.pt. 0: [2022-12-26 06:08:35,052] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_33-model_00-model_states.pt... 0: [2022-12-26 06:08:35,218] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_33-model_00-model_states.pt. 0: [2022-12-26 06:08:35,219] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_34-model_00-model_states.pt... 0: [2022-12-26 06:08:35,381] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_34-model_00-model_states.pt. 0: [2022-12-26 06:08:35,381] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_35-model_00-model_states.pt... 0: [2022-12-26 06:08:35,543] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_35-model_00-model_states.pt. 0: [2022-12-26 06:08:35,543] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_36-model_00-model_states.pt... 0: [2022-12-26 06:08:35,709] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_36-model_00-model_states.pt. 0: [2022-12-26 06:08:35,710] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/layer_38-model_00-model_states.pt... 0: [2022-12-26 06:08:35,713] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/layer_38-model_00-model_states.pt. 0: [2022-12-26 06:08:35,715] [INFO] [logging.py:68:log_dist] [Rank 0] Saving model checkpoint: checkpoints_2b855brealtasky/global_step30000/mp_rank_00_model_states.pt 0: [2022-12-26 06:08:35,715] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/mp_rank_00_model_states.pt... 0: [2022-12-26 06:08:35,722] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/mp_rank_00_model_states.pt. 0: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... 0: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... 0: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... 0: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... 0: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... 16: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt... 18: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt... 24: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt... 24: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt... 24: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt... 24: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt... 29: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt... 29: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt... 0: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... 16: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt... 16: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt... 16: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt... 14: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt... 14: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt... 14: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt... 12: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt... 3: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt... 18: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt... 18: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt... 8: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt... 8: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt... 8: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt... 5: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt... 5: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt... 5: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt... 5: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt... 5: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt... 4: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt... 4: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt... 4: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt... 4: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt... 7: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt... 9: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt... 9: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt... 13: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt... 13: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt... 13: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt... 13: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt... 24: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt... 24: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt... 24: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt... 17: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt... 17: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt... 17: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt... 17: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt... 17: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt... 28: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt... 28: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt... 28: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt... 28: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt... 28: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt... 21: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt... 21: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt... 30: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt... 26: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt... 26: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt... 26: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt... 26: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt... 26: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt... 19: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt... 19: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt... 25: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt... 25: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt... 25: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt... 25: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt... 25: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt... 31: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt... 31: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt... 31: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt... 29: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt... 29: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt... 29: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt... 11: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt... 11: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt... 11: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt... 11: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt... 10: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt... 10: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt... 10: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt... 10: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt... 10: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt... 23: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt... 23: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt... 23: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt... 23: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt... 20: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt... 20: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt... 20: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt... 20: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt... 20: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt... 2: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt... 2: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt... 2: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt... 2: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt... 27: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt... 27: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt... 22: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt... 22: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt... 22: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt... 1: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... 6: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt... 6: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt... 6: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt... 6: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt... 15: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt... 15: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt... 15: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt... 15: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt... 15: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt... 0: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... 16: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt... 16: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt... 16: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt... 14: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt... 14: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt... 14: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt... 12: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt... 12: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt... 12: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt... 12: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt... 12: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt... 3: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt... 3: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt... 18: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt... 8: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt... 8: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt... 8: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt... 8: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt... 8: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt... 5: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt... 5: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt... 4: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt... 7: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt... 7: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt... 7: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt... 7: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt... 7: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt... 9: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt... 9: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt... 9: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt... 9: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt... 13: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt... 13: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt... 13: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt... 13: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt... 24: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt... 17: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt... 17: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt... 28: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt... 28: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt... 21: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt... 21: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt... 21: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt... 21: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt... 30: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt... 30: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt... 30: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt... 30: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt... 30: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt... 26: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt... 26: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt... 26: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt... 19: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt... 19: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt... 19: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt... 19: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt... 25: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt... 31: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt... 31: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt... 31: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt... 29: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt... 29: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt... 11: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt... 11: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt... 10: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt... 10: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt... 10: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt... 23: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt... 23: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt... 20: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt... 20: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt... 20: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt... 2: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt... 2: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt... 2: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt... 2: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt... 27: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt... 27: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt... 27: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt... 27: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt... 27: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt... 22: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt... 22: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt... 22: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt... 1: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... 1: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... 6: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt... 6: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt... 6: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt... 6: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt... 15: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt... 15: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt... 15: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt... 0: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... 16: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt... 14: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt... 12: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt... 12: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt... 3: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt... 18: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt... 5: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt... 4: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt... 7: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt... 9: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt... 17: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt... 28: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt... 21: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt... 30: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt... 19: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt... 25: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt... 25: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt... 31: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt... 31: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt... 29: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt... 11: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt... 11: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt... 23: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt... 27: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt... 22: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt... 1: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... 1: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... 1: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... 1: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... 14: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt... 3: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt... 18: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt... 4: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt... 7: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt... 9: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt... 21: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt... 30: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt... 19: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt... 23: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt... 3: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt... 3: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt... 4: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt... 22: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt... 1: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... 3: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt... 18: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt... 18: [2022-12-26 06:08:35,811] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt... 0: [2022-12-26 06:08:35,960] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. 0: [2022-12-26 06:08:35,961] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt 0: [2022-12-26 06:08:35,961] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 0: [2022-12-26 06:08:35,961] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. 0: [2022-12-26 06:08:35,962] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt 0: [2022-12-26 06:08:35,962] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 0: [2022-12-26 06:08:35,979] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. 0: [2022-12-26 06:08:35,980] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. 0: [2022-12-26 06:08:35,980] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt 0: [2022-12-26 06:08:35,980] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 0: [2022-12-26 06:08:35,981] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. 0: [2022-12-26 06:08:35,981] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt 0: [2022-12-26 06:08:35,981] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 11: [2022-12-26 06:08:36,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt. 11: [2022-12-26 06:08:36,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt. 11: [2022-12-26 06:08:36,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt 11: [2022-12-26 06:08:36,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt 11: [2022-12-26 06:08:36,004] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 11: [2022-12-26 06:08:36,004] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt. 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt. 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt. 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt. 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt. 17: [2022-12-26 06:08:36,005] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt 17: [2022-12-26 06:08:36,005] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt 17: [2022-12-26 06:08:36,005] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt 17: [2022-12-26 06:08:36,005] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt 17: [2022-12-26 06:08:36,005] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 17: [2022-12-26 06:08:36,005] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 26: [2022-12-26 06:08:35,994] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt. 26: [2022-12-26 06:08:35,994] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt. 26: [2022-12-26 06:08:35,994] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt 26: [2022-12-26 06:08:35,994] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt 26: [2022-12-26 06:08:35,994] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 26: [2022-12-26 06:08:35,994] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 26: [2022-12-26 06:08:36,003] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt. 26: [2022-12-26 06:08:36,003] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt 26: [2022-12-26 06:08:36,003] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 26: [2022-12-26 06:08:36,006] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt. 26: [2022-12-26 06:08:36,006] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt 26: [2022-12-26 06:08:36,006] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 26: [2022-12-26 06:08:36,014] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt. 26: [2022-12-26 06:08:36,014] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt 26: [2022-12-26 06:08:36,014] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt. 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt. 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt. 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt. 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt. 22: [2022-12-26 06:08:36,021] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt 22: [2022-12-26 06:08:36,021] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt 22: [2022-12-26 06:08:36,021] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt 22: [2022-12-26 06:08:36,021] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt 22: [2022-12-26 06:08:36,021] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 22: [2022-12-26 06:08:36,021] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt. 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt. 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt. 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt. 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt. 30: [2022-12-26 06:08:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt 30: [2022-12-26 06:08:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt 30: [2022-12-26 06:08:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt 30: [2022-12-26 06:08:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 30: [2022-12-26 06:08:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 30: [2022-12-26 06:08:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 11: [2022-12-26 06:08:36,009] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt. 11: [2022-12-26 06:08:36,009] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt 11: [2022-12-26 06:08:36,009] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 11: [2022-12-26 06:08:36,009] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt. 11: [2022-12-26 06:08:36,009] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt 11: [2022-12-26 06:08:36,009] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 21: [2022-12-26 06:08:36,031] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt. 21: [2022-12-26 06:08:36,031] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt. 21: [2022-12-26 06:08:36,031] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt 21: [2022-12-26 06:08:36,031] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt 21: [2022-12-26 06:08:36,031] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 21: [2022-12-26 06:08:36,031] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 21: [2022-12-26 06:08:36,036] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt. 21: [2022-12-26 06:08:36,036] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt 21: [2022-12-26 06:08:36,036] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 21: [2022-12-26 06:08:36,036] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt. 21: [2022-12-26 06:08:36,036] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt 21: [2022-12-26 06:08:36,036] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 2: [2022-12-26 06:08:36,037] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt. 2: [2022-12-26 06:08:36,037] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt. 2: [2022-12-26 06:08:36,037] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt. 2: [2022-12-26 06:08:36,037] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt 2: [2022-12-26 06:08:36,037] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt 2: [2022-12-26 06:08:36,037] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt 2: [2022-12-26 06:08:36,037] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 2: [2022-12-26 06:08:36,037] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 2: [2022-12-26 06:08:36,037] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 2: [2022-12-26 06:08:36,041] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt. 2: [2022-12-26 06:08:36,041] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt. 2: [2022-12-26 06:08:36,041] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt 2: [2022-12-26 06:08:36,041] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt 2: [2022-12-26 06:08:36,041] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 2: [2022-12-26 06:08:36,041] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt. 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt. 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt. 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt. 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt. 27: [2022-12-26 06:08:36,046] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt 27: [2022-12-26 06:08:36,046] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt 27: [2022-12-26 06:08:36,046] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt 27: [2022-12-26 06:08:36,046] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt 27: [2022-12-26 06:08:36,046] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 27: [2022-12-26 06:08:36,046] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 0: [2022-12-26 06:08:36,048] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. 0: [2022-12-26 06:08:36,048] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt 0: [2022-12-26 06:08:36,048] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 0: [2022-12-26 06:08:36,048] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. 0: [2022-12-26 06:08:36,049] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt 0: [2022-12-26 06:08:36,048] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. 0: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 0: [2022-12-26 06:08:36,049] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt 0: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt. 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt. 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt. 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt. 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt. 10: [2022-12-26 06:08:36,049] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt 10: [2022-12-26 06:08:36,049] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt 10: [2022-12-26 06:08:36,049] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt 10: [2022-12-26 06:08:36,049] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt 10: [2022-12-26 06:08:36,049] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 10: [2022-12-26 06:08:36,049] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 21: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt. 21: [2022-12-26 06:08:36,053] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt 21: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt. 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt. 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt. 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt. 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt. 9: [2022-12-26 06:08:36,053] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt 9: [2022-12-26 06:08:36,053] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt 9: [2022-12-26 06:08:36,053] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 9: [2022-12-26 06:08:36,053] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt 9: [2022-12-26 06:08:36,053] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 9: [2022-12-26 06:08:36,053] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt. 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt. 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt. 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt. 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt. 23: [2022-12-26 06:08:36,066] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt 23: [2022-12-26 06:08:36,066] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt 23: [2022-12-26 06:08:36,066] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt 23: [2022-12-26 06:08:36,066] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt 23: [2022-12-26 06:08:36,066] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 23: [2022-12-26 06:08:36,066] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. 1: [2022-12-26 06:08:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt 1: [2022-12-26 06:08:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt 1: [2022-12-26 06:08:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 1: [2022-12-26 06:08:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 1: [2022-12-26 06:08:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 1: [2022-12-26 06:08:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 26: [2022-12-26 06:08:36,060] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt. 31: [2022-12-26 06:08:36,070] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt. 31: [2022-12-26 06:08:36,070] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt. 31: [2022-12-26 06:08:36,070] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt. 26: [2022-12-26 06:08:36,060] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt 26: [2022-12-26 06:08:36,060] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt. 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt. 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt. 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt. 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt. 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt. 13: [2022-12-26 06:08:36,073] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt 13: [2022-12-26 06:08:36,073] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 13: [2022-12-26 06:08:36,073] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt 13: [2022-12-26 06:08:36,073] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt 13: [2022-12-26 06:08:36,073] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt 13: [2022-12-26 06:08:36,073] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 13: [2022-12-26 06:08:36,073] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 31: [2022-12-26 06:08:36,070] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt 31: [2022-12-26 06:08:36,070] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt 31: [2022-12-26 06:08:36,070] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt 31: [2022-12-26 06:08:36,070] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 31: [2022-12-26 06:08:36,070] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 31: [2022-12-26 06:08:36,070] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 31: [2022-12-26 06:08:36,072] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt. 31: [2022-12-26 06:08:36,072] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt 31: [2022-12-26 06:08:36,072] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 31: [2022-12-26 06:08:36,072] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt. 31: [2022-12-26 06:08:36,072] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt 31: [2022-12-26 06:08:36,072] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 31: [2022-12-26 06:08:36,072] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt. 31: [2022-12-26 06:08:36,072] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt 31: [2022-12-26 06:08:36,072] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 11: [2022-12-26 06:08:36,082] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt. 11: [2022-12-26 06:08:36,082] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt 11: [2022-12-26 06:08:36,082] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 19: [2022-12-26 06:08:36,089] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt. 19: [2022-12-26 06:08:36,089] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt. 19: [2022-12-26 06:08:36,089] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt. 19: [2022-12-26 06:08:36,089] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt. 19: [2022-12-26 06:08:36,089] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt. 19: [2022-12-26 06:08:36,089] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt. 19: [2022-12-26 06:08:36,090] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt 19: [2022-12-26 06:08:36,090] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt 19: [2022-12-26 06:08:36,090] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt 19: [2022-12-26 06:08:36,090] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt 19: [2022-12-26 06:08:36,090] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt 19: [2022-12-26 06:08:36,090] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 19: [2022-12-26 06:08:36,090] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 19: [2022-12-26 06:08:36,090] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 19: [2022-12-26 06:08:36,090] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt 19: [2022-12-26 06:08:36,090] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 19: [2022-12-26 06:08:36,090] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 19: [2022-12-26 06:08:36,090] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 8: [2022-12-26 06:08:36,090] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt. 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt. 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt. 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt. 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt. 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt. 8: [2022-12-26 06:08:36,091] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt 8: [2022-12-26 06:08:36,091] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 8: [2022-12-26 06:08:36,091] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt 8: [2022-12-26 06:08:36,091] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt 8: [2022-12-26 06:08:36,091] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt 8: [2022-12-26 06:08:36,091] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 8: [2022-12-26 06:08:36,091] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt. 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt. 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt. 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt. 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt. 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt. 18: [2022-12-26 06:08:36,094] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt 18: [2022-12-26 06:08:36,094] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt 18: [2022-12-26 06:08:36,094] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt 18: [2022-12-26 06:08:36,094] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt 18: [2022-12-26 06:08:36,094] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt 18: [2022-12-26 06:08:36,094] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 18: [2022-12-26 06:08:36,094] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 5: [2022-12-26 06:08:36,077] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt. 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt. 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt. 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt. 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt. 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt. 3: [2022-12-26 06:08:36,088] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt 3: [2022-12-26 06:08:36,088] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt 3: [2022-12-26 06:08:36,088] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt 3: [2022-12-26 06:08:36,088] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt 3: [2022-12-26 06:08:36,088] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 3: [2022-12-26 06:08:36,088] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 5: [2022-12-26 06:08:36,077] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt. 5: [2022-12-26 06:08:36,078] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt 5: [2022-12-26 06:08:36,078] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 5: [2022-12-26 06:08:36,078] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt 5: [2022-12-26 06:08:36,078] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 5: [2022-12-26 06:08:36,081] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt. 5: [2022-12-26 06:08:36,081] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt 5: [2022-12-26 06:08:36,082] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 5: [2022-12-26 06:08:36,082] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt. 5: [2022-12-26 06:08:36,082] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt. 5: [2022-12-26 06:08:36,082] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt 5: [2022-12-26 06:08:36,082] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt 5: [2022-12-26 06:08:36,082] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 5: [2022-12-26 06:08:36,082] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt. 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt. 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt. 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt. 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt. 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt. 28: [2022-12-26 06:08:36,112] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt 28: [2022-12-26 06:08:36,112] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt 28: [2022-12-26 06:08:36,112] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt 28: [2022-12-26 06:08:36,112] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 28: [2022-12-26 06:08:36,112] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 28: [2022-12-26 06:08:36,112] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 28: [2022-12-26 06:08:36,112] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt. 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt. 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt. 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt. 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt. 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt. 7: [2022-12-26 06:08:36,114] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt 7: [2022-12-26 06:08:36,114] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt 7: [2022-12-26 06:08:36,114] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt 7: [2022-12-26 06:08:36,114] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 7: [2022-12-26 06:08:36,114] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt 7: [2022-12-26 06:08:36,114] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 7: [2022-12-26 06:08:36,114] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt. 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt. 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt. 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt. 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt. 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt. 6: [2022-12-26 06:08:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt 6: [2022-12-26 06:08:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt 6: [2022-12-26 06:08:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt 6: [2022-12-26 06:08:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 6: [2022-12-26 06:08:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt 6: [2022-12-26 06:08:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 6: [2022-12-26 06:08:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 4: [2022-12-26 06:08:36,128] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt. 4: [2022-12-26 06:08:36,128] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt. 4: [2022-12-26 06:08:36,128] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt. 4: [2022-12-26 06:08:36,128] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt 4: [2022-12-26 06:08:36,128] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt 4: [2022-12-26 06:08:36,128] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt 4: [2022-12-26 06:08:36,128] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 4: [2022-12-26 06:08:36,128] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 4: [2022-12-26 06:08:36,128] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 25: [2022-12-26 06:08:36,131] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt. 25: [2022-12-26 06:08:36,131] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt. 25: [2022-12-26 06:08:36,131] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt. 25: [2022-12-26 06:08:36,131] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt. 25: [2022-12-26 06:08:36,131] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt. 25: [2022-12-26 06:08:36,131] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt. 25: [2022-12-26 06:08:36,132] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt 25: [2022-12-26 06:08:36,132] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt 25: [2022-12-26 06:08:36,132] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt 25: [2022-12-26 06:08:36,132] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt 25: [2022-12-26 06:08:36,132] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt 25: [2022-12-26 06:08:36,132] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 25: [2022-12-26 06:08:36,132] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 25: [2022-12-26 06:08:36,132] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 25: [2022-12-26 06:08:36,132] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 25: [2022-12-26 06:08:36,132] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 25: [2022-12-26 06:08:36,132] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt 25: [2022-12-26 06:08:36,132] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt. 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt. 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt. 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt. 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt. 15: [2022-12-26 06:08:36,136] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt 15: [2022-12-26 06:08:36,136] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt 15: [2022-12-26 06:08:36,136] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt 15: [2022-12-26 06:08:36,136] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt 15: [2022-12-26 06:08:36,136] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 15: [2022-12-26 06:08:36,136] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 4: [2022-12-26 06:08:36,139] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt. 4: [2022-12-26 06:08:36,139] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt 4: [2022-12-26 06:08:36,139] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 4: [2022-12-26 06:08:36,139] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt. 4: [2022-12-26 06:08:36,139] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt 4: [2022-12-26 06:08:36,139] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt. 4: [2022-12-26 06:08:36,139] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 4: [2022-12-26 06:08:36,139] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt 4: [2022-12-26 06:08:36,139] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 3: [2022-12-26 06:08:36,124] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt. 3: [2022-12-26 06:08:36,124] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt 3: [2022-12-26 06:08:36,124] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 11: [2022-12-26 06:08:36,147] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt. 11: [2022-12-26 06:08:36,147] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt 11: [2022-12-26 06:08:36,147] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt. 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt. 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt. 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt. 12: [2022-12-26 06:08:36,148] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt 12: [2022-12-26 06:08:36,148] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt 12: [2022-12-26 06:08:36,148] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 12: [2022-12-26 06:08:36,148] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt. 12: [2022-12-26 06:08:36,148] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt 12: [2022-12-26 06:08:36,148] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 15: [2022-12-26 06:08:36,150] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt. 15: [2022-12-26 06:08:36,150] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt 15: [2022-12-26 06:08:36,150] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 12: [2022-12-26 06:08:36,153] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt. 12: [2022-12-26 06:08:36,153] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt. 12: [2022-12-26 06:08:36,153] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt. 12: [2022-12-26 06:08:36,153] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt 12: [2022-12-26 06:08:36,153] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt 12: [2022-12-26 06:08:36,153] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt 12: [2022-12-26 06:08:36,153] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 12: [2022-12-26 06:08:36,153] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 12: [2022-12-26 06:08:36,153] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt. 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt. 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt. 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt. 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt. 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt. 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt. 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt. 16: [2022-12-26 06:08:36,184] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt 16: [2022-12-26 06:08:36,184] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt 16: [2022-12-26 06:08:36,184] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt 16: [2022-12-26 06:08:36,184] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt 16: [2022-12-26 06:08:36,184] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt 16: [2022-12-26 06:08:36,184] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt 16: [2022-12-26 06:08:36,184] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt 16: [2022-12-26 06:08:36,184] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 16: [2022-12-26 06:08:36,184] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 1: [2022-12-26 06:08:36,194] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. 1: [2022-12-26 06:08:36,194] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt 1: [2022-12-26 06:08:36,194] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 0: [2022-12-26 06:08:36,194] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt 0: [2022-12-26 06:08:36,194] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt. 29: [2022-12-26 06:08:36,197] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt. 29: [2022-12-26 06:08:36,197] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt. 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt. 29: [2022-12-26 06:08:36,197] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt. 29: [2022-12-26 06:08:36,197] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt. 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 29: [2022-12-26 06:08:36,197] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt. 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 29: [2022-12-26 06:08:36,197] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt 29: [2022-12-26 06:08:36,197] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 29: [2022-12-26 06:08:36,197] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 24: [2022-12-26 06:08:36,206] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt. 24: [2022-12-26 06:08:36,206] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt 24: [2022-12-26 06:08:36,206] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt. 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt. 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt. 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt. 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt. 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt. 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt. 24: [2022-12-26 06:08:36,207] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt 24: [2022-12-26 06:08:36,207] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt 24: [2022-12-26 06:08:36,207] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt 24: [2022-12-26 06:08:36,207] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt 24: [2022-12-26 06:08:36,207] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt 24: [2022-12-26 06:08:36,207] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt 24: [2022-12-26 06:08:36,207] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 24: [2022-12-26 06:08:36,207] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 20: [2022-12-26 06:08:36,208] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt. 20: [2022-12-26 06:08:36,208] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt. 20: [2022-12-26 06:08:36,208] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt. 20: [2022-12-26 06:08:36,208] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt. 20: [2022-12-26 06:08:36,208] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt. 20: [2022-12-26 06:08:36,208] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt. 20: [2022-12-26 06:08:36,208] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt. 20: [2022-12-26 06:08:36,208] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt 20: [2022-12-26 06:08:36,209] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt 20: [2022-12-26 06:08:36,209] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt 20: [2022-12-26 06:08:36,209] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt 20: [2022-12-26 06:08:36,209] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt 20: [2022-12-26 06:08:36,209] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt 20: [2022-12-26 06:08:36,209] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt 20: [2022-12-26 06:08:36,209] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 20: [2022-12-26 06:08:36,209] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 20: [2022-12-26 06:08:36,209] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 20: [2022-12-26 06:08:36,209] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 20: [2022-12-26 06:08:36,209] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 20: [2022-12-26 06:08:36,209] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 20: [2022-12-26 06:08:36,209] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 23: [2022-12-26 06:08:36,217] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt. 23: [2022-12-26 06:08:36,217] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt 23: [2022-12-26 06:08:36,217] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 2: [2022-12-26 06:08:36,220] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt. 2: [2022-12-26 06:08:36,220] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt 2: [2022-12-26 06:08:36,220] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt. 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt. 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt. 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt. 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt. 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt. 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt. 14: [2022-12-26 06:08:36,225] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt 14: [2022-12-26 06:08:36,225] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt. 14: [2022-12-26 06:08:36,225] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt 14: [2022-12-26 06:08:36,225] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt 14: [2022-12-26 06:08:36,225] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt 14: [2022-12-26 06:08:36,225] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt 14: [2022-12-26 06:08:36,225] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 14: [2022-12-26 06:08:36,225] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 14: [2022-12-26 06:08:36,225] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 10: [2022-12-26 06:08:36,228] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt. 10: [2022-12-26 06:08:36,228] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt 10: [2022-12-26 06:08:36,228] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 6: [2022-12-26 06:08:36,231] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt. 6: [2022-12-26 06:08:36,231] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt 6: [2022-12-26 06:08:36,231] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 22: [2022-12-26 06:08:36,233] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt. 22: [2022-12-26 06:08:36,233] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt 22: [2022-12-26 06:08:36,233] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 9: [2022-12-26 06:08:36,233] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt. 9: [2022-12-26 06:08:36,234] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt 9: [2022-12-26 06:08:36,234] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 21: [2022-12-26 06:08:36,236] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt. 21: [2022-12-26 06:08:36,237] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt 21: [2022-12-26 06:08:36,237] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 17: [2022-12-26 06:08:36,239] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt. 17: [2022-12-26 06:08:36,239] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt 17: [2022-12-26 06:08:36,239] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 30: [2022-12-26 06:08:36,239] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt. 30: [2022-12-26 06:08:36,239] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt 30: [2022-12-26 06:08:36,239] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 27: [2022-12-26 06:08:36,246] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt. 27: [2022-12-26 06:08:36,246] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt 27: [2022-12-26 06:08:36,247] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 5: [2022-12-26 06:08:36,240] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt. 5: [2022-12-26 06:08:36,240] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt 5: [2022-12-26 06:08:36,240] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 19: [2022-12-26 06:08:36,249] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt. 19: [2022-12-26 06:08:36,249] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt 19: [2022-12-26 06:08:36,249] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 13: [2022-12-26 06:08:36,254] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt. 13: [2022-12-26 06:08:36,254] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt 13: [2022-12-26 06:08:36,254] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 28: [2022-12-26 06:08:36,256] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt. 7: [2022-12-26 06:08:36,256] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt. 28: [2022-12-26 06:08:36,256] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt 28: [2022-12-26 06:08:36,256] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 7: [2022-12-26 06:08:36,256] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt 7: [2022-12-26 06:08:36,256] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 8: [2022-12-26 06:08:36,257] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt. 8: [2022-12-26 06:08:36,257] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt 8: [2022-12-26 06:08:36,257] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 11: [2022-12-26 06:08:36,259] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt. 11: [2022-12-26 06:08:36,259] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt 11: [2022-12-26 06:08:36,259] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 31: [2022-12-26 06:08:36,260] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt. 20: [2022-12-26 06:08:36,266] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt. 20: [2022-12-26 06:08:36,266] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt 20: [2022-12-26 06:08:36,266] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 31: [2022-12-26 06:08:36,260] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt 31: [2022-12-26 06:08:36,260] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 25: [2022-12-26 06:08:36,269] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt. 25: [2022-12-26 06:08:36,269] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt 25: [2022-12-26 06:08:36,269] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 18: [2022-12-26 06:08:36,270] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt. 18: [2022-12-26 06:08:36,270] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt 18: [2022-12-26 06:08:36,270] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 26: [2022-12-26 06:08:36,276] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt. 26: [2022-12-26 06:08:36,276] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt 26: [2022-12-26 06:08:36,277] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 4: [2022-12-26 06:08:36,278] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt. 4: [2022-12-26 06:08:36,278] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt 4: [2022-12-26 06:08:36,278] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 15: [2022-12-26 06:08:36,282] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt. 15: [2022-12-26 06:08:36,282] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt 15: [2022-12-26 06:08:36,282] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 3: [2022-12-26 06:08:36,291] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt. 3: [2022-12-26 06:08:36,291] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt 3: [2022-12-26 06:08:36,291] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 29: [2022-12-26 06:08:36,295] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt. 29: [2022-12-26 06:08:36,295] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt 29: [2022-12-26 06:08:36,295] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 23: [2022-12-26 06:08:36,297] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt. 23: [2022-12-26 06:08:36,297] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt 23: [2022-12-26 06:08:36,297] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 6: [2022-12-26 06:08:36,297] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt. 6: [2022-12-26 06:08:36,297] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt 6: [2022-12-26 06:08:36,297] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 23: [2022-12-26 06:08:36,297] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt. 23: [2022-12-26 06:08:36,297] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt 23: [2022-12-26 06:08:36,298] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 7: [2022-12-26 06:08:36,298] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt. 7: [2022-12-26 06:08:36,298] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt 7: [2022-12-26 06:08:36,298] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 25: [2022-12-26 06:08:36,299] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt. 25: [2022-12-26 06:08:36,299] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt 25: [2022-12-26 06:08:36,299] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 26: [2022-12-26 06:08:36,300] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt. 26: [2022-12-26 06:08:36,300] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt 26: [2022-12-26 06:08:36,300] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 19: [2022-12-26 06:08:36,300] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt. 19: [2022-12-26 06:08:36,300] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt 19: [2022-12-26 06:08:36,300] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 2: [2022-12-26 06:08:36,301] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt. 2: [2022-12-26 06:08:36,302] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt 2: [2022-12-26 06:08:36,302] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 10: [2022-12-26 06:08:36,302] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt. 10: [2022-12-26 06:08:36,303] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt 10: [2022-12-26 06:08:36,303] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 18: [2022-12-26 06:08:36,304] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt. 27: [2022-12-26 06:08:36,304] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt. 27: [2022-12-26 06:08:36,304] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt 18: [2022-12-26 06:08:36,304] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt 27: [2022-12-26 06:08:36,304] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 18: [2022-12-26 06:08:36,304] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 4: [2022-12-26 06:08:36,305] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt. 4: [2022-12-26 06:08:36,305] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt 4: [2022-12-26 06:08:36,305] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 31: [2022-12-26 06:08:36,305] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt. 31: [2022-12-26 06:08:36,305] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt 31: [2022-12-26 06:08:36,305] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 11: [2022-12-26 06:08:36,305] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt. 11: [2022-12-26 06:08:36,305] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt 11: [2022-12-26 06:08:36,306] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 30: [2022-12-26 06:08:36,306] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt. 30: [2022-12-26 06:08:36,306] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt 30: [2022-12-26 06:08:36,306] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 21: [2022-12-26 06:08:36,306] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt. 21: [2022-12-26 06:08:36,306] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt 21: [2022-12-26 06:08:36,306] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 22: [2022-12-26 06:08:36,306] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt. 22: [2022-12-26 06:08:36,306] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt 22: [2022-12-26 06:08:36,306] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 22: [2022-12-26 06:08:36,306] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt. 22: [2022-12-26 06:08:36,307] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt 22: [2022-12-26 06:08:36,307] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 28: [2022-12-26 06:08:36,307] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt. 28: [2022-12-26 06:08:36,308] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt 28: [2022-12-26 06:08:36,308] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 13: [2022-12-26 06:08:36,308] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt. 13: [2022-12-26 06:08:36,308] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt 13: [2022-12-26 06:08:36,308] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 17: [2022-12-26 06:08:36,308] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt. 17: [2022-12-26 06:08:36,308] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt 17: [2022-12-26 06:08:36,308] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 3: [2022-12-26 06:08:36,308] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt. 3: [2022-12-26 06:08:36,308] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt 15: [2022-12-26 06:08:36,308] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt. 15: [2022-12-26 06:08:36,308] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt 3: [2022-12-26 06:08:36,308] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 15: [2022-12-26 06:08:36,308] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 21: [2022-12-26 06:08:36,309] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt. 30: [2022-12-26 06:08:36,309] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt. 21: [2022-12-26 06:08:36,309] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt 30: [2022-12-26 06:08:36,309] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt 30: [2022-12-26 06:08:36,309] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 21: [2022-12-26 06:08:36,309] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 2: [2022-12-26 06:08:36,309] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt. 2: [2022-12-26 06:08:36,309] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt 2: [2022-12-26 06:08:36,310] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 5: [2022-12-26 06:08:36,310] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt. 5: [2022-12-26 06:08:36,310] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt. 5: [2022-12-26 06:08:36,310] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt 5: [2022-12-26 06:08:36,310] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt 5: [2022-12-26 06:08:36,310] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 5: [2022-12-26 06:08:36,310] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 8: [2022-12-26 06:08:36,310] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt. 8: [2022-12-26 06:08:36,310] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt 8: [2022-12-26 06:08:36,310] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 1: [2022-12-26 06:08:36,312] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. 1: [2022-12-26 06:08:36,312] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. 1: [2022-12-26 06:08:36,312] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt 1: [2022-12-26 06:08:36,312] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt 1: [2022-12-26 06:08:36,312] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 1: [2022-12-26 06:08:36,312] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 27: [2022-12-26 06:08:36,313] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt. 27: [2022-12-26 06:08:36,313] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt 27: [2022-12-26 06:08:36,313] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 17: [2022-12-26 06:08:36,313] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt. 17: [2022-12-26 06:08:36,313] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt 17: [2022-12-26 06:08:36,313] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 9: [2022-12-26 06:08:36,316] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt. 9: [2022-12-26 06:08:36,316] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt. 9: [2022-12-26 06:08:36,316] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt 9: [2022-12-26 06:08:36,316] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt 9: [2022-12-26 06:08:36,316] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 9: [2022-12-26 06:08:36,316] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 10: [2022-12-26 06:08:36,317] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt. 10: [2022-12-26 06:08:36,317] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step30000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt 10: [2022-12-26 06:08:36,317] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step30000 is ready now! 0: successfully saved checkpoint at iteration 30000 to checkpoints_2b855brealtasky 31: time (ms) | save-checkpoint: 6750.54 31: iteration 30010/ 52452 | consumed samples: 15365120 | consumed tokens: 31467765760 | elapsed time per iteration (s): 9.80 | learning rate: 9.097E-05 | global batch size: 512 | lm loss: 1.271440E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 52.222 | TFLOPs: 7.84 | 31: iteration 30020/ 52452 | consumed samples: 15370240 | consumed tokens: 31478251520 | elapsed time per iteration (s): 1.79 | learning rate: 9.092E-05 | global batch size: 512 | lm loss: 1.280777E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.531 | TFLOPs: 42.86 | 31: iteration 30030/ 52452 | consumed samples: 15375360 | consumed tokens: 31488737280 | elapsed time per iteration (s): 1.84 | learning rate: 9.086E-05 | global batch size: 512 | lm loss: 1.276760E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.782 | TFLOPs: 41.84 | 31: iteration 30040/ 52452 | consumed samples: 15380480 | consumed tokens: 31499223040 | elapsed time per iteration (s): 1.78 | learning rate: 9.081E-05 | global batch size: 512 | lm loss: 1.268621E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.948 | TFLOPs: 43.22 | 31: iteration 30050/ 52452 | consumed samples: 15385600 | consumed tokens: 31509708800 | elapsed time per iteration (s): 1.74 | learning rate: 9.076E-05 | global batch size: 512 | lm loss: 1.278510E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.142 | TFLOPs: 44.15 | 31: iteration 30060/ 52452 | consumed samples: 15390720 | consumed tokens: 31520194560 | elapsed time per iteration (s): 1.77 | learning rate: 9.070E-05 | global batch size: 512 | lm loss: 1.266389E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.455 | TFLOPs: 43.45 | 31: iteration 30070/ 52452 | consumed samples: 15395840 | consumed tokens: 31530680320 | elapsed time per iteration (s): 1.76 | learning rate: 9.065E-05 | global batch size: 512 | lm loss: 1.266713E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.454 | TFLOPs: 43.60 | 31: iteration 30080/ 52452 | consumed samples: 15400960 | consumed tokens: 31541166080 | elapsed time per iteration (s): 1.78 | learning rate: 9.060E-05 | global batch size: 512 | lm loss: 1.271072E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.570 | TFLOPs: 43.16 | 31: iteration 30090/ 52452 | consumed samples: 15406080 | consumed tokens: 31551651840 | elapsed time per iteration (s): 1.78 | learning rate: 9.054E-05 | global batch size: 512 | lm loss: 1.274982E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.347 | TFLOPs: 43.28 | 31: iteration 30100/ 52452 | consumed samples: 15411200 | consumed tokens: 31562137600 | elapsed time per iteration (s): 1.78 | learning rate: 9.049E-05 | global batch size: 512 | lm loss: 1.269497E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.886 | TFLOPs: 43.21 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30100 | lm loss value: 4.161458E+00 | lm loss PPL: 6.416498E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30100 | lm loss value: 4.051394E+00 | lm loss PPL: 5.747752E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 30110/ 52452 | consumed samples: 15416320 | consumed tokens: 31572623360 | elapsed time per iteration (s): 9.21 | learning rate: 9.044E-05 | global batch size: 512 | lm loss: 1.274196E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.618 | TFLOPs: 8.35 | 31: iteration 30120/ 52452 | consumed samples: 15421440 | consumed tokens: 31583109120 | elapsed time per iteration (s): 1.76 | learning rate: 9.038E-05 | global batch size: 512 | lm loss: 1.274973E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.235 | TFLOPs: 43.56 | 31: iteration 30130/ 52452 | consumed samples: 15426560 | consumed tokens: 31593594880 | elapsed time per iteration (s): 1.79 | learning rate: 9.033E-05 | global batch size: 512 | lm loss: 1.274345E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.569 | TFLOPs: 42.86 | 31: iteration 30140/ 52452 | consumed samples: 15431680 | consumed tokens: 31604080640 | elapsed time per iteration (s): 1.77 | learning rate: 9.028E-05 | global batch size: 512 | lm loss: 1.276896E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.867 | TFLOPs: 43.51 | 31: iteration 30150/ 52452 | consumed samples: 15436800 | consumed tokens: 31614566400 | elapsed time per iteration (s): 1.76 | learning rate: 9.022E-05 | global batch size: 512 | lm loss: 1.266687E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.038 | TFLOPs: 43.68 | 31: iteration 30160/ 52452 | consumed samples: 15441920 | consumed tokens: 31625052160 | elapsed time per iteration (s): 1.82 | learning rate: 9.017E-05 | global batch size: 512 | lm loss: 1.272293E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.046 | TFLOPs: 42.33 | 31: iteration 30170/ 52452 | consumed samples: 15447040 | consumed tokens: 31635537920 | elapsed time per iteration (s): 1.79 | learning rate: 9.012E-05 | global batch size: 512 | lm loss: 1.266266E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.819 | TFLOPs: 42.90 | 31: iteration 30180/ 52452 | consumed samples: 15452160 | consumed tokens: 31646023680 | elapsed time per iteration (s): 1.77 | learning rate: 9.007E-05 | global batch size: 512 | lm loss: 1.277109E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.863 | TFLOPs: 43.36 | 31: iteration 30190/ 52452 | consumed samples: 15457280 | consumed tokens: 31656509440 | elapsed time per iteration (s): 1.76 | learning rate: 9.001E-05 | global batch size: 512 | lm loss: 1.263786E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.354 | TFLOPs: 43.73 | 31: iteration 30200/ 52452 | consumed samples: 15462400 | consumed tokens: 31666995200 | elapsed time per iteration (s): 1.79 | learning rate: 8.996E-05 | global batch size: 512 | lm loss: 1.284448E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.271 | TFLOPs: 42.97 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30200 | lm loss value: 4.000529E+00 | lm loss PPL: 5.462703E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30200 | lm loss value: 3.955659E+00 | lm loss PPL: 5.223012E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 30210/ 52452 | consumed samples: 15467520 | consumed tokens: 31677480960 | elapsed time per iteration (s): 9.18 | learning rate: 8.991E-05 | global batch size: 512 | lm loss: 1.268567E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.795 | TFLOPs: 8.37 | 31: iteration 30220/ 52452 | consumed samples: 15472640 | consumed tokens: 31687966720 | elapsed time per iteration (s): 1.76 | learning rate: 8.985E-05 | global batch size: 512 | lm loss: 1.277575E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.833 | TFLOPs: 43.65 | 31: iteration 30230/ 52452 | consumed samples: 15477760 | consumed tokens: 31698452480 | elapsed time per iteration (s): 1.73 | learning rate: 8.980E-05 | global batch size: 512 | lm loss: 1.277777E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.257 | TFLOPs: 44.32 | 31: iteration 30240/ 52452 | consumed samples: 15482880 | consumed tokens: 31708938240 | elapsed time per iteration (s): 1.73 | learning rate: 8.975E-05 | global batch size: 512 | lm loss: 1.282715E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.032 | TFLOPs: 44.43 | 31: iteration 30250/ 52452 | consumed samples: 15488000 | consumed tokens: 31719424000 | elapsed time per iteration (s): 1.74 | learning rate: 8.969E-05 | global batch size: 512 | lm loss: 1.271209E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.323 | TFLOPs: 44.18 | 31: iteration 30260/ 52452 | consumed samples: 15493120 | consumed tokens: 31729909760 | elapsed time per iteration (s): 1.75 | learning rate: 8.964E-05 | global batch size: 512 | lm loss: 1.274054E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.881 | TFLOPs: 43.81 | 31: iteration 30270/ 52452 | consumed samples: 15498240 | consumed tokens: 31740395520 | elapsed time per iteration (s): 1.77 | learning rate: 8.959E-05 | global batch size: 512 | lm loss: 1.267169E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.038 | TFLOPs: 43.38 | 31: iteration 30280/ 52452 | consumed samples: 15503360 | consumed tokens: 31750881280 | elapsed time per iteration (s): 1.77 | learning rate: 8.953E-05 | global batch size: 512 | lm loss: 1.266014E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.510 | TFLOPs: 43.30 | 31: iteration 30290/ 52452 | consumed samples: 15508480 | consumed tokens: 31761367040 | elapsed time per iteration (s): 1.78 | learning rate: 8.948E-05 | global batch size: 512 | lm loss: 1.293854E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.228 | TFLOPs: 43.11 | 31: iteration 30300/ 52452 | consumed samples: 15513600 | consumed tokens: 31771852800 | elapsed time per iteration (s): 1.76 | learning rate: 8.943E-05 | global batch size: 512 | lm loss: 1.259113E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.579 | TFLOPs: 43.76 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30300 | lm loss value: 4.160722E+00 | lm loss PPL: 6.411782E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30300 | lm loss value: 4.064845E+00 | lm loss PPL: 5.825588E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 30310/ 52452 | consumed samples: 15518720 | consumed tokens: 31782338560 | elapsed time per iteration (s): 9.19 | learning rate: 8.938E-05 | global batch size: 512 | lm loss: 1.278683E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.727 | TFLOPs: 8.36 | 31: iteration 30320/ 52452 | consumed samples: 15523840 | consumed tokens: 31792824320 | elapsed time per iteration (s): 1.77 | learning rate: 8.932E-05 | global batch size: 512 | lm loss: 1.262948E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.625 | TFLOPs: 43.32 | 31: iteration 30330/ 52452 | consumed samples: 15528960 | consumed tokens: 31803310080 | elapsed time per iteration (s): 1.77 | learning rate: 8.927E-05 | global batch size: 512 | lm loss: 1.264736E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.699 | TFLOPs: 43.48 | 31: iteration 30340/ 52452 | consumed samples: 15534080 | consumed tokens: 31813795840 | elapsed time per iteration (s): 1.78 | learning rate: 8.922E-05 | global batch size: 512 | lm loss: 1.277583E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.957 | TFLOPs: 43.22 | 31: iteration 30350/ 52452 | consumed samples: 15539200 | consumed tokens: 31824281600 | elapsed time per iteration (s): 1.79 | learning rate: 8.916E-05 | global batch size: 512 | lm loss: 1.264381E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.430 | TFLOPs: 42.84 | 31: iteration 30360/ 52452 | consumed samples: 15544320 | consumed tokens: 31834767360 | elapsed time per iteration (s): 1.76 | learning rate: 8.911E-05 | global batch size: 512 | lm loss: 1.272087E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.275 | TFLOPs: 43.72 | 31: iteration 30370/ 52452 | consumed samples: 15549440 | consumed tokens: 31845253120 | elapsed time per iteration (s): 1.80 | learning rate: 8.906E-05 | global batch size: 512 | lm loss: 1.282848E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.408 | TFLOPs: 42.69 | 31: iteration 30380/ 52452 | consumed samples: 15554560 | consumed tokens: 31855738880 | elapsed time per iteration (s): 1.78 | learning rate: 8.901E-05 | global batch size: 512 | lm loss: 1.273582E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.329 | TFLOPs: 43.28 | 31: iteration 30390/ 52452 | consumed samples: 15559680 | consumed tokens: 31866224640 | elapsed time per iteration (s): 1.76 | learning rate: 8.895E-05 | global batch size: 512 | lm loss: 1.282166E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.556 | TFLOPs: 43.61 | 31: iteration 30400/ 52452 | consumed samples: 15564800 | consumed tokens: 31876710400 | elapsed time per iteration (s): 1.80 | learning rate: 8.890E-05 | global batch size: 512 | lm loss: 1.266588E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.650 | TFLOPs: 42.72 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30400 | lm loss value: 4.271733E+00 | lm loss PPL: 7.164568E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30400 | lm loss value: 4.128555E+00 | lm loss PPL: 6.208813E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 30410/ 52452 | consumed samples: 15569920 | consumed tokens: 31887196160 | elapsed time per iteration (s): 9.15 | learning rate: 8.885E-05 | global batch size: 512 | lm loss: 1.262545E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.939 | TFLOPs: 8.40 | 31: iteration 30420/ 52452 | consumed samples: 15575040 | consumed tokens: 31897681920 | elapsed time per iteration (s): 1.74 | learning rate: 8.879E-05 | global batch size: 512 | lm loss: 1.271043E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.679 | TFLOPs: 44.08 | 31: iteration 30430/ 52452 | consumed samples: 15580160 | consumed tokens: 31908167680 | elapsed time per iteration (s): 1.78 | learning rate: 8.874E-05 | global batch size: 512 | lm loss: 1.270982E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.338 | TFLOPs: 43.13 | 31: iteration 30440/ 52452 | consumed samples: 15585280 | consumed tokens: 31918653440 | elapsed time per iteration (s): 1.76 | learning rate: 8.869E-05 | global batch size: 512 | lm loss: 1.275850E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.110 | TFLOPs: 43.54 | 31: iteration 30450/ 52452 | consumed samples: 15590400 | consumed tokens: 31929139200 | elapsed time per iteration (s): 1.75 | learning rate: 8.863E-05 | global batch size: 512 | lm loss: 1.264798E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.939 | TFLOPs: 43.82 | 31: iteration 30460/ 52452 | consumed samples: 15595520 | consumed tokens: 31939624960 | elapsed time per iteration (s): 1.78 | learning rate: 8.858E-05 | global batch size: 512 | lm loss: 1.273897E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.095 | TFLOPs: 43.09 | 31: iteration 30470/ 52452 | consumed samples: 15600640 | consumed tokens: 31950110720 | elapsed time per iteration (s): 1.80 | learning rate: 8.853E-05 | global batch size: 512 | lm loss: 1.270444E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.925 | TFLOPs: 42.77 | 31: iteration 30480/ 52452 | consumed samples: 15605760 | consumed tokens: 31960596480 | elapsed time per iteration (s): 1.76 | learning rate: 8.848E-05 | global batch size: 512 | lm loss: 1.281792E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.542 | TFLOPs: 43.76 | 31: iteration 30490/ 52452 | consumed samples: 15610880 | consumed tokens: 31971082240 | elapsed time per iteration (s): 1.76 | learning rate: 8.842E-05 | global batch size: 512 | lm loss: 1.279477E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.644 | TFLOPs: 43.62 | 31: iteration 30500/ 52452 | consumed samples: 15616000 | consumed tokens: 31981568000 | elapsed time per iteration (s): 1.73 | learning rate: 8.837E-05 | global batch size: 512 | lm loss: 1.271597E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.504 | TFLOPs: 44.50 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30500 | lm loss value: 3.956156E+00 | lm loss PPL: 5.225608E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30500 | lm loss value: 3.923685E+00 | lm loss PPL: 5.058649E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 30510/ 52452 | consumed samples: 15621120 | consumed tokens: 31992053760 | elapsed time per iteration (s): 9.25 | learning rate: 8.832E-05 | global batch size: 512 | lm loss: 1.275737E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.340 | TFLOPs: 8.31 | 31: iteration 30520/ 52452 | consumed samples: 15626240 | consumed tokens: 32002539520 | elapsed time per iteration (s): 1.76 | learning rate: 8.826E-05 | global batch size: 512 | lm loss: 1.267119E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.832 | TFLOPs: 43.65 | 31: iteration 30530/ 52452 | consumed samples: 15631360 | consumed tokens: 32013025280 | elapsed time per iteration (s): 1.79 | learning rate: 8.821E-05 | global batch size: 512 | lm loss: 1.280132E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.621 | TFLOPs: 43.02 | 31: iteration 30540/ 52452 | consumed samples: 15636480 | consumed tokens: 32023511040 | elapsed time per iteration (s): 1.78 | learning rate: 8.816E-05 | global batch size: 512 | lm loss: 1.262043E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.409 | TFLOPs: 43.14 | 31: iteration 30550/ 52452 | consumed samples: 15641600 | consumed tokens: 32033996800 | elapsed time per iteration (s): 1.76 | learning rate: 8.811E-05 | global batch size: 512 | lm loss: 1.291262E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.651 | TFLOPs: 43.63 | 31: iteration 30560/ 52452 | consumed samples: 15646720 | consumed tokens: 32044482560 | elapsed time per iteration (s): 1.74 | learning rate: 8.805E-05 | global batch size: 512 | lm loss: 1.270120E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.945 | TFLOPs: 44.27 | 31: iteration 30570/ 52452 | consumed samples: 15651840 | consumed tokens: 32054968320 | elapsed time per iteration (s): 1.73 | learning rate: 8.800E-05 | global batch size: 512 | lm loss: 1.267014E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.550 | TFLOPs: 44.36 | 31: iteration 30580/ 52452 | consumed samples: 15656960 | consumed tokens: 32065454080 | elapsed time per iteration (s): 1.79 | learning rate: 8.795E-05 | global batch size: 512 | lm loss: 1.282860E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.714 | TFLOPs: 43.03 | 31: iteration 30590/ 52452 | consumed samples: 15662080 | consumed tokens: 32075939840 | elapsed time per iteration (s): 1.75 | learning rate: 8.789E-05 | global batch size: 512 | lm loss: 1.268345E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.407 | TFLOPs: 44.04 | 31: iteration 30600/ 52452 | consumed samples: 15667200 | consumed tokens: 32086425600 | elapsed time per iteration (s): 1.78 | learning rate: 8.784E-05 | global batch size: 512 | lm loss: 1.261990E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.193 | TFLOPs: 43.26 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30600 | lm loss value: 4.125101E+00 | lm loss PPL: 6.187406E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30600 | lm loss value: 4.049742E+00 | lm loss PPL: 5.738266E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 30610/ 52452 | consumed samples: 15672320 | consumed tokens: 32096911360 | elapsed time per iteration (s): 9.19 | learning rate: 8.779E-05 | global batch size: 512 | lm loss: 1.257034E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.727 | TFLOPs: 8.36 | 31: iteration 30620/ 52452 | consumed samples: 15677440 | consumed tokens: 32107397120 | elapsed time per iteration (s): 1.75 | learning rate: 8.774E-05 | global batch size: 512 | lm loss: 1.264174E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.842 | TFLOPs: 43.80 | 31: iteration 30630/ 52452 | consumed samples: 15682560 | consumed tokens: 32117882880 | elapsed time per iteration (s): 1.80 | learning rate: 8.768E-05 | global batch size: 512 | lm loss: 1.266626E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.473 | TFLOPs: 42.70 | 31: iteration 30640/ 52452 | consumed samples: 15687680 | consumed tokens: 32128368640 | elapsed time per iteration (s): 1.81 | learning rate: 8.763E-05 | global batch size: 512 | lm loss: 1.266486E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.882 | TFLOPs: 42.46 | 31: iteration 30650/ 52452 | consumed samples: 15692800 | consumed tokens: 32138854400 | elapsed time per iteration (s): 1.79 | learning rate: 8.758E-05 | global batch size: 512 | lm loss: 1.261359E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.477 | TFLOPs: 43.00 | 31: iteration 30660/ 52452 | consumed samples: 15697920 | consumed tokens: 32149340160 | elapsed time per iteration (s): 1.77 | learning rate: 8.753E-05 | global batch size: 512 | lm loss: 1.274198E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.131 | TFLOPs: 43.40 | 31: iteration 30670/ 52452 | consumed samples: 15703040 | consumed tokens: 32159825920 | elapsed time per iteration (s): 1.80 | learning rate: 8.747E-05 | global batch size: 512 | lm loss: 1.267750E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.550 | TFLOPs: 42.71 | 31: iteration 30680/ 52452 | consumed samples: 15708160 | consumed tokens: 32170311680 | elapsed time per iteration (s): 1.80 | learning rate: 8.742E-05 | global batch size: 512 | lm loss: 1.284188E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.934 | TFLOPs: 42.62 | 31: iteration 30690/ 52452 | consumed samples: 15713280 | consumed tokens: 32180797440 | elapsed time per iteration (s): 1.85 | learning rate: 8.737E-05 | global batch size: 512 | lm loss: 1.259173E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.032 | TFLOPs: 41.58 | 31: iteration 30700/ 52452 | consumed samples: 15718400 | consumed tokens: 32191283200 | elapsed time per iteration (s): 1.85 | learning rate: 8.731E-05 | global batch size: 512 | lm loss: 1.270781E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.268 | TFLOPs: 41.62 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30700 | lm loss value: 4.152423E+00 | lm loss PPL: 6.358788E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30700 | lm loss value: 4.053478E+00 | lm loss PPL: 5.759745E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 30710/ 52452 | consumed samples: 15723520 | consumed tokens: 32201768960 | elapsed time per iteration (s): 9.20 | learning rate: 8.726E-05 | global batch size: 512 | lm loss: 1.272814E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.675 | TFLOPs: 8.36 | 31: iteration 30720/ 52452 | consumed samples: 15728640 | consumed tokens: 32212254720 | elapsed time per iteration (s): 1.79 | learning rate: 8.721E-05 | global batch size: 512 | lm loss: 1.277891E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.284 | TFLOPs: 42.82 | 31: iteration 30730/ 52452 | consumed samples: 15733760 | consumed tokens: 32222740480 | elapsed time per iteration (s): 1.81 | learning rate: 8.716E-05 | global batch size: 512 | lm loss: 1.266306E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.996 | TFLOPs: 42.48 | 31: iteration 30740/ 52452 | consumed samples: 15738880 | consumed tokens: 32233226240 | elapsed time per iteration (s): 1.78 | learning rate: 8.710E-05 | global batch size: 512 | lm loss: 1.268457E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.571 | TFLOPs: 43.16 | 31: iteration 30750/ 52452 | consumed samples: 15744000 | consumed tokens: 32243712000 | elapsed time per iteration (s): 1.81 | learning rate: 8.705E-05 | global batch size: 512 | lm loss: 1.266020E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.311 | TFLOPs: 42.52 | 31: iteration 30760/ 52452 | consumed samples: 15749120 | consumed tokens: 32254197760 | elapsed time per iteration (s): 1.77 | learning rate: 8.700E-05 | global batch size: 512 | lm loss: 1.274157E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.489 | TFLOPs: 43.30 | 31: iteration 30770/ 52452 | consumed samples: 15754240 | consumed tokens: 32264683520 | elapsed time per iteration (s): 1.77 | learning rate: 8.695E-05 | global batch size: 512 | lm loss: 1.275329E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.623 | TFLOPs: 43.47 | 31: iteration 30780/ 52452 | consumed samples: 15759360 | consumed tokens: 32275169280 | elapsed time per iteration (s): 1.79 | learning rate: 8.689E-05 | global batch size: 512 | lm loss: 1.272166E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.120 | TFLOPs: 42.95 | 31: iteration 30790/ 52452 | consumed samples: 15764480 | consumed tokens: 32285655040 | elapsed time per iteration (s): 1.82 | learning rate: 8.684E-05 | global batch size: 512 | lm loss: 1.265258E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.128 | TFLOPs: 42.20 | 31: iteration 30800/ 52452 | consumed samples: 15769600 | consumed tokens: 32296140800 | elapsed time per iteration (s): 1.75 | learning rate: 8.679E-05 | global batch size: 512 | lm loss: 1.266730E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.469 | TFLOPs: 43.90 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30800 | lm loss value: 4.139665E+00 | lm loss PPL: 6.278176E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30800 | lm loss value: 4.046441E+00 | lm loss PPL: 5.719352E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 30810/ 52452 | consumed samples: 15774720 | consumed tokens: 32306626560 | elapsed time per iteration (s): 9.15 | learning rate: 8.674E-05 | global batch size: 512 | lm loss: 1.285217E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.984 | TFLOPs: 8.40 | 31: iteration 30820/ 52452 | consumed samples: 15779840 | consumed tokens: 32317112320 | elapsed time per iteration (s): 1.74 | learning rate: 8.668E-05 | global batch size: 512 | lm loss: 1.272276E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.459 | TFLOPs: 44.05 | 31: iteration 30830/ 52452 | consumed samples: 15784960 | consumed tokens: 32327598080 | elapsed time per iteration (s): 1.78 | learning rate: 8.663E-05 | global batch size: 512 | lm loss: 1.258922E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.338 | TFLOPs: 43.13 | 31: iteration 30840/ 52452 | consumed samples: 15790080 | consumed tokens: 32338083840 | elapsed time per iteration (s): 1.75 | learning rate: 8.658E-05 | global batch size: 512 | lm loss: 1.263467E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.443 | TFLOPs: 43.89 | 31: iteration 30850/ 52452 | consumed samples: 15795200 | consumed tokens: 32348569600 | elapsed time per iteration (s): 1.80 | learning rate: 8.653E-05 | global batch size: 512 | lm loss: 1.271202E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.006 | TFLOPs: 42.78 | 31: iteration 30860/ 52452 | consumed samples: 15800320 | consumed tokens: 32359055360 | elapsed time per iteration (s): 1.81 | learning rate: 8.647E-05 | global batch size: 512 | lm loss: 1.274811E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.243 | TFLOPs: 42.36 | 31: iteration 30870/ 52452 | consumed samples: 15805440 | consumed tokens: 32369541120 | elapsed time per iteration (s): 1.79 | learning rate: 8.642E-05 | global batch size: 512 | lm loss: 1.254047E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.474 | TFLOPs: 42.85 | 31: iteration 30880/ 52452 | consumed samples: 15810560 | consumed tokens: 32380026880 | elapsed time per iteration (s): 1.77 | learning rate: 8.637E-05 | global batch size: 512 | lm loss: 1.268263E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.649 | TFLOPs: 43.32 | 31: iteration 30890/ 52452 | consumed samples: 15815680 | consumed tokens: 32390512640 | elapsed time per iteration (s): 1.74 | learning rate: 8.632E-05 | global batch size: 512 | lm loss: 1.269029E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.356 | TFLOPs: 44.18 | 31: iteration 30900/ 52452 | consumed samples: 15820800 | consumed tokens: 32400998400 | elapsed time per iteration (s): 1.76 | learning rate: 8.626E-05 | global batch size: 512 | lm loss: 1.277138E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.730 | TFLOPs: 43.64 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 30900 | lm loss value: 4.087131E+00 | lm loss PPL: 5.956874E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 30900 | lm loss value: 4.020193E+00 | lm loss PPL: 5.571184E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 30910/ 52452 | consumed samples: 15825920 | consumed tokens: 32411484160 | elapsed time per iteration (s): 9.16 | learning rate: 8.621E-05 | global batch size: 512 | lm loss: 1.267545E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.879 | TFLOPs: 8.39 | 31: iteration 30920/ 52452 | consumed samples: 15831040 | consumed tokens: 32421969920 | elapsed time per iteration (s): 1.78 | learning rate: 8.616E-05 | global batch size: 512 | lm loss: 1.266442E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.261 | TFLOPs: 43.27 | 31: iteration 30930/ 52452 | consumed samples: 15836160 | consumed tokens: 32432455680 | elapsed time per iteration (s): 1.76 | learning rate: 8.611E-05 | global batch size: 512 | lm loss: 1.268861E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.097 | TFLOPs: 43.54 | 31: iteration 30940/ 52452 | consumed samples: 15841280 | consumed tokens: 32442941440 | elapsed time per iteration (s): 1.79 | learning rate: 8.605E-05 | global batch size: 512 | lm loss: 1.261790E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.534 | TFLOPs: 43.01 | 31: iteration 30950/ 52452 | consumed samples: 15846400 | consumed tokens: 32453427200 | elapsed time per iteration (s): 1.75 | learning rate: 8.600E-05 | global batch size: 512 | lm loss: 1.273112E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.701 | TFLOPs: 43.93 | 31: iteration 30960/ 52452 | consumed samples: 15851520 | consumed tokens: 32463912960 | elapsed time per iteration (s): 1.75 | learning rate: 8.595E-05 | global batch size: 512 | lm loss: 1.265602E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.265 | TFLOPs: 44.02 | 31: iteration 30970/ 52452 | consumed samples: 15856640 | consumed tokens: 32474398720 | elapsed time per iteration (s): 1.76 | learning rate: 8.590E-05 | global batch size: 512 | lm loss: 1.265615E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.945 | TFLOPs: 43.67 | 31: iteration 30980/ 52452 | consumed samples: 15861760 | consumed tokens: 32484884480 | elapsed time per iteration (s): 1.75 | learning rate: 8.584E-05 | global batch size: 512 | lm loss: 1.266575E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.741 | TFLOPs: 43.79 | 31: iteration 30990/ 52452 | consumed samples: 15866880 | consumed tokens: 32495370240 | elapsed time per iteration (s): 2.67 | learning rate: 8.579E-05 | global batch size: 512 | lm loss: 1.256278E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 191.777 | TFLOPs: 28.78 | 31: iteration 31000/ 52452 | consumed samples: 15872000 | consumed tokens: 32505856000 | elapsed time per iteration (s): 1.75 | learning rate: 8.574E-05 | global batch size: 512 | lm loss: 1.271296E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.316 | TFLOPs: 44.03 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31000 | lm loss value: 4.129251E+00 | lm loss PPL: 6.213134E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31000 | lm loss value: 4.039928E+00 | lm loss PPL: 5.682225E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31010/ 52452 | consumed samples: 15877120 | consumed tokens: 32516341760 | elapsed time per iteration (s): 9.19 | learning rate: 8.569E-05 | global batch size: 512 | lm loss: 1.271100E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.740 | TFLOPs: 8.37 | 31: iteration 31020/ 52452 | consumed samples: 15882240 | consumed tokens: 32526827520 | elapsed time per iteration (s): 1.75 | learning rate: 8.563E-05 | global batch size: 512 | lm loss: 1.267715E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.696 | TFLOPs: 43.93 | 31: iteration 31030/ 52452 | consumed samples: 15887360 | consumed tokens: 32537313280 | elapsed time per iteration (s): 1.74 | learning rate: 8.558E-05 | global batch size: 512 | lm loss: 1.284603E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.826 | TFLOPs: 44.25 | 31: iteration 31040/ 52452 | consumed samples: 15892480 | consumed tokens: 32547799040 | elapsed time per iteration (s): 1.75 | learning rate: 8.553E-05 | global batch size: 512 | lm loss: 1.253365E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.130 | TFLOPs: 43.85 | 31: iteration 31050/ 52452 | consumed samples: 15897600 | consumed tokens: 32558284800 | elapsed time per iteration (s): 1.82 | learning rate: 8.548E-05 | global batch size: 512 | lm loss: 1.270357E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.434 | TFLOPs: 42.24 | 31: iteration 31060/ 52452 | consumed samples: 15902720 | consumed tokens: 32568770560 | elapsed time per iteration (s): 1.73 | learning rate: 8.542E-05 | global batch size: 512 | lm loss: 1.265711E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.142 | TFLOPs: 44.45 | 31: iteration 31070/ 52452 | consumed samples: 15907840 | consumed tokens: 32579256320 | elapsed time per iteration (s): 1.76 | learning rate: 8.537E-05 | global batch size: 512 | lm loss: 1.263300E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.099 | TFLOPs: 43.69 | 31: iteration 31080/ 52452 | consumed samples: 15912960 | consumed tokens: 32589742080 | elapsed time per iteration (s): 1.76 | learning rate: 8.532E-05 | global batch size: 512 | lm loss: 1.263645E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.794 | TFLOPs: 43.65 | 31: iteration 31090/ 52452 | consumed samples: 15918080 | consumed tokens: 32600227840 | elapsed time per iteration (s): 1.74 | learning rate: 8.527E-05 | global batch size: 512 | lm loss: 1.271340E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.521 | TFLOPs: 44.21 | 31: iteration 31100/ 52452 | consumed samples: 15923200 | consumed tokens: 32610713600 | elapsed time per iteration (s): 1.73 | learning rate: 8.521E-05 | global batch size: 512 | lm loss: 1.263054E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.639 | TFLOPs: 44.37 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31100 | lm loss value: 4.152987E+00 | lm loss PPL: 6.362373E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31100 | lm loss value: 4.058660E+00 | lm loss PPL: 5.789668E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31110/ 52452 | consumed samples: 15928320 | consumed tokens: 32621199360 | elapsed time per iteration (s): 9.14 | learning rate: 8.516E-05 | global batch size: 512 | lm loss: 1.253338E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.008 | TFLOPs: 8.41 | 31: iteration 31120/ 52452 | consumed samples: 15933440 | consumed tokens: 32631685120 | elapsed time per iteration (s): 1.74 | learning rate: 8.511E-05 | global batch size: 512 | lm loss: 1.280688E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.811 | TFLOPs: 44.25 | 31: iteration 31130/ 52452 | consumed samples: 15938560 | consumed tokens: 32642170880 | elapsed time per iteration (s): 1.74 | learning rate: 8.506E-05 | global batch size: 512 | lm loss: 1.256334E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.327 | TFLOPs: 44.18 | 31: iteration 31140/ 52452 | consumed samples: 15943680 | consumed tokens: 32652656640 | elapsed time per iteration (s): 1.77 | learning rate: 8.500E-05 | global batch size: 512 | lm loss: 1.269792E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.524 | TFLOPs: 43.31 | 31: iteration 31150/ 52452 | consumed samples: 15948800 | consumed tokens: 32663142400 | elapsed time per iteration (s): 1.80 | learning rate: 8.495E-05 | global batch size: 512 | lm loss: 1.288596E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.204 | TFLOPs: 42.81 | 31: iteration 31160/ 52452 | consumed samples: 15953920 | consumed tokens: 32673628160 | elapsed time per iteration (s): 1.72 | learning rate: 8.490E-05 | global batch size: 512 | lm loss: 1.261039E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.694 | TFLOPs: 44.68 | 31: iteration 31170/ 52452 | consumed samples: 15959040 | consumed tokens: 32684113920 | elapsed time per iteration (s): 1.74 | learning rate: 8.485E-05 | global batch size: 512 | lm loss: 1.259981E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.396 | TFLOPs: 44.19 | 31: iteration 31180/ 52452 | consumed samples: 15964160 | consumed tokens: 32694599680 | elapsed time per iteration (s): 1.74 | learning rate: 8.480E-05 | global batch size: 512 | lm loss: 1.260822E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.090 | TFLOPs: 44.29 | 31: iteration 31190/ 52452 | consumed samples: 15969280 | consumed tokens: 32705085440 | elapsed time per iteration (s): 1.75 | learning rate: 8.474E-05 | global batch size: 512 | lm loss: 1.268003E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.699 | TFLOPs: 43.93 | 31: iteration 31200/ 52452 | consumed samples: 15974400 | consumed tokens: 32715571200 | elapsed time per iteration (s): 1.82 | learning rate: 8.469E-05 | global batch size: 512 | lm loss: 1.258704E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.174 | TFLOPs: 42.20 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31200 | lm loss value: 4.177945E+00 | lm loss PPL: 6.523167E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31200 | lm loss value: 4.079510E+00 | lm loss PPL: 5.911651E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31210/ 52452 | consumed samples: 15979520 | consumed tokens: 32726056960 | elapsed time per iteration (s): 9.22 | learning rate: 8.464E-05 | global batch size: 512 | lm loss: 1.256559E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.553 | TFLOPs: 8.34 | 31: iteration 31220/ 52452 | consumed samples: 15984640 | consumed tokens: 32736542720 | elapsed time per iteration (s): 1.79 | learning rate: 8.459E-05 | global batch size: 512 | lm loss: 1.265375E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.205 | TFLOPs: 42.96 | 31: iteration 31230/ 52452 | consumed samples: 15989760 | consumed tokens: 32747028480 | elapsed time per iteration (s): 1.74 | learning rate: 8.453E-05 | global batch size: 512 | lm loss: 1.265629E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.865 | TFLOPs: 44.11 | 31: iteration 31240/ 52452 | consumed samples: 15994880 | consumed tokens: 32757514240 | elapsed time per iteration (s): 1.75 | learning rate: 8.448E-05 | global batch size: 512 | lm loss: 1.274156E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.252 | TFLOPs: 44.02 | 31: iteration 31250/ 52452 | consumed samples: 16000000 | consumed tokens: 32768000000 | elapsed time per iteration (s): 1.80 | learning rate: 8.443E-05 | global batch size: 512 | lm loss: 1.270238E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.209 | TFLOPs: 42.81 | 31: iteration 31260/ 52452 | consumed samples: 16005120 | consumed tokens: 32778485760 | elapsed time per iteration (s): 1.75 | learning rate: 8.438E-05 | global batch size: 512 | lm loss: 1.267513E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.820 | TFLOPs: 43.80 | 31: iteration 31270/ 52452 | consumed samples: 16010240 | consumed tokens: 32788971520 | elapsed time per iteration (s): 1.77 | learning rate: 8.433E-05 | global batch size: 512 | lm loss: 1.277088E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.980 | TFLOPs: 43.37 | 31: iteration 31280/ 52452 | consumed samples: 16015360 | consumed tokens: 32799457280 | elapsed time per iteration (s): 1.75 | learning rate: 8.427E-05 | global batch size: 512 | lm loss: 1.261953E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.285 | TFLOPs: 44.02 | 31: iteration 31290/ 52452 | consumed samples: 16020480 | consumed tokens: 32809943040 | elapsed time per iteration (s): 1.73 | learning rate: 8.422E-05 | global batch size: 512 | lm loss: 1.263068E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.469 | TFLOPs: 44.50 | 31: iteration 31300/ 52452 | consumed samples: 16025600 | consumed tokens: 32820428800 | elapsed time per iteration (s): 1.74 | learning rate: 8.417E-05 | global batch size: 512 | lm loss: 1.259819E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.744 | TFLOPs: 44.09 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31300 | lm loss value: 4.196469E+00 | lm loss PPL: 6.645127E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31300 | lm loss value: 4.073676E+00 | lm loss PPL: 5.877259E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31310/ 52452 | consumed samples: 16030720 | consumed tokens: 32830914560 | elapsed time per iteration (s): 9.22 | learning rate: 8.412E-05 | global batch size: 512 | lm loss: 1.278238E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.529 | TFLOPs: 8.33 | 31: iteration 31320/ 52452 | consumed samples: 16035840 | consumed tokens: 32841400320 | elapsed time per iteration (s): 1.79 | learning rate: 8.406E-05 | global batch size: 512 | lm loss: 1.254076E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.335 | TFLOPs: 42.83 | 31: iteration 31330/ 52452 | consumed samples: 16040960 | consumed tokens: 32851886080 | elapsed time per iteration (s): 1.78 | learning rate: 8.401E-05 | global batch size: 512 | lm loss: 1.275264E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.170 | TFLOPs: 43.10 | 31: iteration 31340/ 52452 | consumed samples: 16046080 | consumed tokens: 32862371840 | elapsed time per iteration (s): 1.81 | learning rate: 8.396E-05 | global batch size: 512 | lm loss: 1.269114E+00 | grad norm: 0.147 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.202 | TFLOPs: 42.51 | 31: iteration 31350/ 52452 | consumed samples: 16051200 | consumed tokens: 32872857600 | elapsed time per iteration (s): 1.72 | learning rate: 8.391E-05 | global batch size: 512 | lm loss: 1.256629E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.337 | TFLOPs: 44.63 | 31: iteration 31360/ 52452 | consumed samples: 16056320 | consumed tokens: 32883343360 | elapsed time per iteration (s): 1.78 | learning rate: 8.386E-05 | global batch size: 512 | lm loss: 1.254748E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.464 | TFLOPs: 43.15 | 31: iteration 31370/ 52452 | consumed samples: 16061440 | consumed tokens: 32893829120 | elapsed time per iteration (s): 1.78 | learning rate: 8.380E-05 | global batch size: 512 | lm loss: 1.270018E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.165 | TFLOPs: 43.25 | 31: iteration 31380/ 52452 | consumed samples: 16066560 | consumed tokens: 32904314880 | elapsed time per iteration (s): 1.75 | learning rate: 8.375E-05 | global batch size: 512 | lm loss: 1.268687E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.304 | TFLOPs: 43.87 | 31: iteration 31390/ 52452 | consumed samples: 16071680 | consumed tokens: 32914800640 | elapsed time per iteration (s): 1.74 | learning rate: 8.370E-05 | global batch size: 512 | lm loss: 1.262936E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.134 | TFLOPs: 44.15 | 31: iteration 31400/ 52452 | consumed samples: 16076800 | consumed tokens: 32925286400 | elapsed time per iteration (s): 1.75 | learning rate: 8.365E-05 | global batch size: 512 | lm loss: 1.258900E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.353 | TFLOPs: 43.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31400 | lm loss value: 4.287925E+00 | lm loss PPL: 7.281520E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31400 | lm loss value: 4.146459E+00 | lm loss PPL: 6.320975E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31410/ 52452 | consumed samples: 16081920 | consumed tokens: 32935772160 | elapsed time per iteration (s): 9.13 | learning rate: 8.360E-05 | global batch size: 512 | lm loss: 1.259099E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.108 | TFLOPs: 8.42 | 31: iteration 31420/ 52452 | consumed samples: 16087040 | consumed tokens: 32946257920 | elapsed time per iteration (s): 1.76 | learning rate: 8.354E-05 | global batch size: 512 | lm loss: 1.263121E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.672 | TFLOPs: 43.63 | 31: iteration 31430/ 52452 | consumed samples: 16092160 | consumed tokens: 32956743680 | elapsed time per iteration (s): 1.75 | learning rate: 8.349E-05 | global batch size: 512 | lm loss: 1.255364E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.872 | TFLOPs: 43.81 | 31: iteration 31440/ 52452 | consumed samples: 16097280 | consumed tokens: 32967229440 | elapsed time per iteration (s): 1.75 | learning rate: 8.344E-05 | global batch size: 512 | lm loss: 1.268621E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.431 | TFLOPs: 43.89 | 31: iteration 31450/ 52452 | consumed samples: 16102400 | consumed tokens: 32977715200 | elapsed time per iteration (s): 1.84 | learning rate: 8.339E-05 | global batch size: 512 | lm loss: 1.263665E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.662 | TFLOPs: 41.83 | 31: iteration 31460/ 52452 | consumed samples: 16107520 | consumed tokens: 32988200960 | elapsed time per iteration (s): 1.77 | learning rate: 8.334E-05 | global batch size: 512 | lm loss: 1.259117E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.994 | TFLOPs: 43.38 | 31: iteration 31470/ 52452 | consumed samples: 16112640 | consumed tokens: 32998686720 | elapsed time per iteration (s): 1.77 | learning rate: 8.328E-05 | global batch size: 512 | lm loss: 1.268430E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.602 | TFLOPs: 43.32 | 31: iteration 31480/ 52452 | consumed samples: 16117760 | consumed tokens: 33009172480 | elapsed time per iteration (s): 1.73 | learning rate: 8.323E-05 | global batch size: 512 | lm loss: 1.255056E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.392 | TFLOPs: 44.49 | 31: iteration 31490/ 52452 | consumed samples: 16122880 | consumed tokens: 33019658240 | elapsed time per iteration (s): 1.75 | learning rate: 8.318E-05 | global batch size: 512 | lm loss: 1.270245E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.263 | TFLOPs: 43.87 | 31: iteration 31500/ 52452 | consumed samples: 16128000 | consumed tokens: 33030144000 | elapsed time per iteration (s): 1.81 | learning rate: 8.313E-05 | global batch size: 512 | lm loss: 1.262477E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.317 | TFLOPs: 42.37 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31500 | lm loss value: 4.199834E+00 | lm loss PPL: 6.667529E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31500 | lm loss value: 4.078706E+00 | lm loss PPL: 5.906897E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31510/ 52452 | consumed samples: 16133120 | consumed tokens: 33040629760 | elapsed time per iteration (s): 9.20 | learning rate: 8.308E-05 | global batch size: 512 | lm loss: 1.266219E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.647 | TFLOPs: 8.35 | 31: iteration 31520/ 52452 | consumed samples: 16138240 | consumed tokens: 33051115520 | elapsed time per iteration (s): 1.79 | learning rate: 8.302E-05 | global batch size: 512 | lm loss: 1.259319E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.053 | TFLOPs: 42.93 | 31: iteration 31530/ 52452 | consumed samples: 16143360 | consumed tokens: 33061601280 | elapsed time per iteration (s): 1.78 | learning rate: 8.297E-05 | global batch size: 512 | lm loss: 1.255383E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.866 | TFLOPs: 43.06 | 31: iteration 31540/ 52452 | consumed samples: 16148480 | consumed tokens: 33072087040 | elapsed time per iteration (s): 1.74 | learning rate: 8.292E-05 | global batch size: 512 | lm loss: 1.266399E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.573 | TFLOPs: 44.21 | 31: iteration 31550/ 52452 | consumed samples: 16153600 | consumed tokens: 33082572800 | elapsed time per iteration (s): 1.76 | learning rate: 8.287E-05 | global batch size: 512 | lm loss: 1.259219E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.868 | TFLOPs: 43.66 | 31: iteration 31560/ 52452 | consumed samples: 16158720 | consumed tokens: 33093058560 | elapsed time per iteration (s): 2.20 | learning rate: 8.282E-05 | global batch size: 512 | lm loss: 1.266225E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 232.282 | TFLOPs: 34.86 | 31: iteration 31570/ 52452 | consumed samples: 16163840 | consumed tokens: 33103544320 | elapsed time per iteration (s): 1.78 | learning rate: 8.276E-05 | global batch size: 512 | lm loss: 1.269666E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.314 | TFLOPs: 43.12 | 31: iteration 31580/ 52452 | consumed samples: 16168960 | consumed tokens: 33114030080 | elapsed time per iteration (s): 1.76 | learning rate: 8.271E-05 | global batch size: 512 | lm loss: 1.262561E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.281 | TFLOPs: 43.72 | 31: iteration 31590/ 52452 | consumed samples: 16174080 | consumed tokens: 33124515840 | elapsed time per iteration (s): 1.93 | learning rate: 8.266E-05 | global batch size: 512 | lm loss: 1.277200E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.280 | TFLOPs: 39.82 | 31: iteration 31600/ 52452 | consumed samples: 16179200 | consumed tokens: 33135001600 | elapsed time per iteration (s): 1.79 | learning rate: 8.261E-05 | global batch size: 512 | lm loss: 1.264800E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.717 | TFLOPs: 42.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31600 | lm loss value: 4.056925E+00 | lm loss PPL: 5.779630E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31600 | lm loss value: 4.001184E+00 | lm loss PPL: 5.466283E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31610/ 52452 | consumed samples: 16184320 | consumed tokens: 33145487360 | elapsed time per iteration (s): 9.21 | learning rate: 8.256E-05 | global batch size: 512 | lm loss: 1.263851E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.619 | TFLOPs: 8.35 | 31: iteration 31620/ 52452 | consumed samples: 16189440 | consumed tokens: 33155973120 | elapsed time per iteration (s): 1.78 | learning rate: 8.250E-05 | global batch size: 512 | lm loss: 1.257065E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.948 | TFLOPs: 43.22 | 31: iteration 31630/ 52452 | consumed samples: 16194560 | consumed tokens: 33166458880 | elapsed time per iteration (s): 1.73 | learning rate: 8.245E-05 | global batch size: 512 | lm loss: 1.254826E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.106 | TFLOPs: 44.29 | 31: iteration 31640/ 52452 | consumed samples: 16199680 | consumed tokens: 33176944640 | elapsed time per iteration (s): 1.77 | learning rate: 8.240E-05 | global batch size: 512 | lm loss: 1.275992E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.801 | TFLOPs: 43.50 | 31: iteration 31650/ 52452 | consumed samples: 16204800 | consumed tokens: 33187430400 | elapsed time per iteration (s): 1.77 | learning rate: 8.235E-05 | global batch size: 512 | lm loss: 1.256630E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.865 | TFLOPs: 43.51 | 31: iteration 31660/ 52452 | consumed samples: 16209920 | consumed tokens: 33197916160 | elapsed time per iteration (s): 1.77 | learning rate: 8.230E-05 | global batch size: 512 | lm loss: 1.256915E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.477 | TFLOPs: 43.45 | 31: iteration 31670/ 52452 | consumed samples: 16215040 | consumed tokens: 33208401920 | elapsed time per iteration (s): 1.79 | learning rate: 8.225E-05 | global batch size: 512 | lm loss: 1.261810E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.365 | TFLOPs: 42.98 | 31: iteration 31680/ 52452 | consumed samples: 16220160 | consumed tokens: 33218887680 | elapsed time per iteration (s): 1.83 | learning rate: 8.219E-05 | global batch size: 512 | lm loss: 1.269917E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.346 | TFLOPs: 41.93 | 31: iteration 31690/ 52452 | consumed samples: 16225280 | consumed tokens: 33229373440 | elapsed time per iteration (s): 1.76 | learning rate: 8.214E-05 | global batch size: 512 | lm loss: 1.249867E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.695 | TFLOPs: 43.63 | 31: iteration 31700/ 52452 | consumed samples: 16230400 | consumed tokens: 33239859200 | elapsed time per iteration (s): 1.79 | learning rate: 8.209E-05 | global batch size: 512 | lm loss: 1.261913E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.327 | TFLOPs: 42.83 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31700 | lm loss value: 4.142843E+00 | lm loss PPL: 6.298161E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31700 | lm loss value: 4.045140E+00 | lm loss PPL: 5.711917E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31710/ 52452 | consumed samples: 16235520 | consumed tokens: 33250344960 | elapsed time per iteration (s): 9.21 | learning rate: 8.204E-05 | global batch size: 512 | lm loss: 1.248500E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.617 | TFLOPs: 8.35 | 31: iteration 31720/ 52452 | consumed samples: 16240640 | consumed tokens: 33260830720 | elapsed time per iteration (s): 1.82 | learning rate: 8.199E-05 | global batch size: 512 | lm loss: 1.254242E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.920 | TFLOPs: 42.31 | 31: iteration 31730/ 52452 | consumed samples: 16245760 | consumed tokens: 33271316480 | elapsed time per iteration (s): 1.79 | learning rate: 8.194E-05 | global batch size: 512 | lm loss: 1.256283E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.582 | TFLOPs: 43.01 | 31: iteration 31740/ 52452 | consumed samples: 16250880 | consumed tokens: 33281802240 | elapsed time per iteration (s): 1.85 | learning rate: 8.188E-05 | global batch size: 512 | lm loss: 1.263567E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.456 | TFLOPs: 41.64 | 31: iteration 31750/ 52452 | consumed samples: 16256000 | consumed tokens: 33292288000 | elapsed time per iteration (s): 1.83 | learning rate: 8.183E-05 | global batch size: 512 | lm loss: 1.260094E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.781 | TFLOPs: 41.99 | 31: iteration 31760/ 52452 | consumed samples: 16261120 | consumed tokens: 33302773760 | elapsed time per iteration (s): 1.84 | learning rate: 8.178E-05 | global batch size: 512 | lm loss: 1.254665E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.075 | TFLOPs: 41.74 | 31: iteration 31770/ 52452 | consumed samples: 16266240 | consumed tokens: 33313259520 | elapsed time per iteration (s): 1.79 | learning rate: 8.173E-05 | global batch size: 512 | lm loss: 1.271610E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.449 | TFLOPs: 42.99 | 31: iteration 31780/ 52452 | consumed samples: 16271360 | consumed tokens: 33323745280 | elapsed time per iteration (s): 1.81 | learning rate: 8.168E-05 | global batch size: 512 | lm loss: 1.250640E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.524 | TFLOPs: 42.56 | 31: iteration 31790/ 52452 | consumed samples: 16276480 | consumed tokens: 33334231040 | elapsed time per iteration (s): 1.86 | learning rate: 8.162E-05 | global batch size: 512 | lm loss: 1.262271E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.780 | TFLOPs: 41.39 | 31: iteration 31800/ 52452 | consumed samples: 16281600 | consumed tokens: 33344716800 | elapsed time per iteration (s): 1.79 | learning rate: 8.157E-05 | global batch size: 512 | lm loss: 1.256595E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.302 | TFLOPs: 42.97 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31800 | lm loss value: 4.158000E+00 | lm loss PPL: 6.394348E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31800 | lm loss value: 4.059430E+00 | lm loss PPL: 5.794128E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31810/ 52452 | consumed samples: 16286720 | consumed tokens: 33355202560 | elapsed time per iteration (s): 9.27 | learning rate: 8.152E-05 | global batch size: 512 | lm loss: 1.261883E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.226 | TFLOPs: 8.29 | 31: iteration 31820/ 52452 | consumed samples: 16291840 | consumed tokens: 33365688320 | elapsed time per iteration (s): 1.81 | learning rate: 8.147E-05 | global batch size: 512 | lm loss: 1.261728E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.174 | TFLOPs: 42.35 | 31: iteration 31830/ 52452 | consumed samples: 16296960 | consumed tokens: 33376174080 | elapsed time per iteration (s): 1.81 | learning rate: 8.142E-05 | global batch size: 512 | lm loss: 1.264300E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.328 | TFLOPs: 42.38 | 31: iteration 31840/ 52452 | consumed samples: 16302080 | consumed tokens: 33386659840 | elapsed time per iteration (s): 1.84 | learning rate: 8.137E-05 | global batch size: 512 | lm loss: 1.262947E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.259 | TFLOPs: 41.77 | 31: iteration 31850/ 52452 | consumed samples: 16307200 | consumed tokens: 33397145600 | elapsed time per iteration (s): 1.78 | learning rate: 8.132E-05 | global batch size: 512 | lm loss: 1.249704E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.809 | TFLOPs: 43.20 | 31: iteration 31860/ 52452 | consumed samples: 16312320 | consumed tokens: 33407631360 | elapsed time per iteration (s): 1.78 | learning rate: 8.126E-05 | global batch size: 512 | lm loss: 1.261071E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.313 | TFLOPs: 43.27 | 31: iteration 31870/ 52452 | consumed samples: 16317440 | consumed tokens: 33418117120 | elapsed time per iteration (s): 1.79 | learning rate: 8.121E-05 | global batch size: 512 | lm loss: 1.245938E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.356 | TFLOPs: 42.83 | 31: iteration 31880/ 52452 | consumed samples: 16322560 | consumed tokens: 33428602880 | elapsed time per iteration (s): 1.74 | learning rate: 8.116E-05 | global batch size: 512 | lm loss: 1.261353E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.703 | TFLOPs: 44.08 | 31: iteration 31890/ 52452 | consumed samples: 16327680 | consumed tokens: 33439088640 | elapsed time per iteration (s): 1.78 | learning rate: 8.111E-05 | global batch size: 512 | lm loss: 1.257400E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.339 | TFLOPs: 43.13 | 31: iteration 31900/ 52452 | consumed samples: 16332800 | consumed tokens: 33449574400 | elapsed time per iteration (s): 1.76 | learning rate: 8.106E-05 | global batch size: 512 | lm loss: 1.262384E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.974 | TFLOPs: 43.67 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 31900 | lm loss value: 4.194408E+00 | lm loss PPL: 6.631446E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 31900 | lm loss value: 4.069057E+00 | lm loss PPL: 5.850174E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 31910/ 52452 | consumed samples: 16337920 | consumed tokens: 33460060160 | elapsed time per iteration (s): 9.16 | learning rate: 8.101E-05 | global batch size: 512 | lm loss: 1.269141E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.898 | TFLOPs: 8.39 | 31: iteration 31920/ 52452 | consumed samples: 16343040 | consumed tokens: 33470545920 | elapsed time per iteration (s): 1.89 | learning rate: 8.095E-05 | global batch size: 512 | lm loss: 1.246208E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.354 | TFLOPs: 40.58 | 31: iteration 31930/ 52452 | consumed samples: 16348160 | consumed tokens: 33481031680 | elapsed time per iteration (s): 1.92 | learning rate: 8.090E-05 | global batch size: 512 | lm loss: 1.260759E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.215 | TFLOPs: 40.11 | 31: iteration 31940/ 52452 | consumed samples: 16353280 | consumed tokens: 33491517440 | elapsed time per iteration (s): 1.79 | learning rate: 8.085E-05 | global batch size: 512 | lm loss: 1.249117E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.520 | TFLOPs: 42.86 | 31: iteration 31950/ 52452 | consumed samples: 16358400 | consumed tokens: 33502003200 | elapsed time per iteration (s): 1.85 | learning rate: 8.080E-05 | global batch size: 512 | lm loss: 1.248518E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.806 | TFLOPs: 41.55 | 31: iteration 31960/ 52452 | consumed samples: 16363520 | consumed tokens: 33512488960 | elapsed time per iteration (s): 1.78 | learning rate: 8.075E-05 | global batch size: 512 | lm loss: 1.268768E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.298 | TFLOPs: 43.27 | 31: iteration 31970/ 52452 | consumed samples: 16368640 | consumed tokens: 33522974720 | elapsed time per iteration (s): 1.77 | learning rate: 8.070E-05 | global batch size: 512 | lm loss: 1.246767E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.708 | TFLOPs: 43.33 | 31: iteration 31980/ 52452 | consumed samples: 16373760 | consumed tokens: 33533460480 | elapsed time per iteration (s): 1.83 | learning rate: 8.065E-05 | global batch size: 512 | lm loss: 1.258108E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.344 | TFLOPs: 42.08 | 31: iteration 31990/ 52452 | consumed samples: 16378880 | consumed tokens: 33543946240 | elapsed time per iteration (s): 1.77 | learning rate: 8.059E-05 | global batch size: 512 | lm loss: 1.255118E+00 | grad norm: 0.146 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.802 | TFLOPs: 43.35 | 0: [2022-12-26 07:31:32,072] [INFO] [logging.py:68:log_dist] [Rank 0] step=32000, skipped=0, lr=[8.054213866759568e-05, 8.054213866759568e-05, 8.054213866759568e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 32000/ 52452 | consumed samples: 16384000 | consumed tokens: 33554432000 | elapsed time per iteration (s): 1.78 | learning rate: 8.054E-05 | global batch size: 512 | lm loss: 1.243481E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.026 | TFLOPs: 43.08 | 0: steps: 32000 loss: 1.2537 iter time (s): 1.786 samples/sec: 286.618 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32000 | lm loss value: 4.094697E+00 | lm loss PPL: 6.002112E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32000 | lm loss value: 4.018249E+00 | lm loss PPL: 5.560366E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32010/ 52452 | consumed samples: 16389120 | consumed tokens: 33564917760 | elapsed time per iteration (s): 9.20 | learning rate: 8.049E-05 | global batch size: 512 | lm loss: 1.263798E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.641 | TFLOPs: 8.35 | 31: iteration 32020/ 52452 | consumed samples: 16394240 | consumed tokens: 33575403520 | elapsed time per iteration (s): 1.82 | learning rate: 8.044E-05 | global batch size: 512 | lm loss: 1.255578E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.069 | TFLOPs: 42.34 | 31: iteration 32030/ 52452 | consumed samples: 16399360 | consumed tokens: 33585889280 | elapsed time per iteration (s): 1.89 | learning rate: 8.039E-05 | global batch size: 512 | lm loss: 1.252963E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.818 | TFLOPs: 40.65 | 31: iteration 32040/ 52452 | consumed samples: 16404480 | consumed tokens: 33596375040 | elapsed time per iteration (s): 1.74 | learning rate: 8.034E-05 | global batch size: 512 | lm loss: 1.251450E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.557 | TFLOPs: 44.06 | 31: iteration 32050/ 52452 | consumed samples: 16409600 | consumed tokens: 33606860800 | elapsed time per iteration (s): 1.78 | learning rate: 8.029E-05 | global batch size: 512 | lm loss: 1.255252E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.751 | TFLOPs: 43.19 | 31: iteration 32060/ 52452 | consumed samples: 16414720 | consumed tokens: 33617346560 | elapsed time per iteration (s): 1.82 | learning rate: 8.023E-05 | global batch size: 512 | lm loss: 1.252958E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.166 | TFLOPs: 42.20 | 31: iteration 32070/ 52452 | consumed samples: 16419840 | consumed tokens: 33627832320 | elapsed time per iteration (s): 1.81 | learning rate: 8.018E-05 | global batch size: 512 | lm loss: 1.259120E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.282 | TFLOPs: 42.52 | 31: iteration 32080/ 52452 | consumed samples: 16424960 | consumed tokens: 33638318080 | elapsed time per iteration (s): 1.77 | learning rate: 8.013E-05 | global batch size: 512 | lm loss: 1.261746E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.526 | TFLOPs: 43.46 | 31: iteration 32090/ 52452 | consumed samples: 16430080 | consumed tokens: 33648803840 | elapsed time per iteration (s): 1.80 | learning rate: 8.008E-05 | global batch size: 512 | lm loss: 1.249301E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.175 | TFLOPs: 42.80 | 31: iteration 32100/ 52452 | consumed samples: 16435200 | consumed tokens: 33659289600 | elapsed time per iteration (s): 1.77 | learning rate: 8.003E-05 | global batch size: 512 | lm loss: 1.256820E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.758 | TFLOPs: 43.49 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32100 | lm loss value: 4.275436E+00 | lm loss PPL: 7.191148E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32100 | lm loss value: 4.133353E+00 | lm loss PPL: 6.238674E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32110/ 52452 | consumed samples: 16440320 | consumed tokens: 33669775360 | elapsed time per iteration (s): 9.19 | learning rate: 7.998E-05 | global batch size: 512 | lm loss: 1.267039E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.696 | TFLOPs: 8.36 | 31: iteration 32120/ 52452 | consumed samples: 16445440 | consumed tokens: 33680261120 | elapsed time per iteration (s): 1.91 | learning rate: 7.993E-05 | global batch size: 512 | lm loss: 1.257854E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.681 | TFLOPs: 40.33 | 31: iteration 32130/ 52452 | consumed samples: 16450560 | consumed tokens: 33690746880 | elapsed time per iteration (s): 1.75 | learning rate: 7.987E-05 | global batch size: 512 | lm loss: 1.253945E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.491 | TFLOPs: 43.90 | 31: iteration 32140/ 52452 | consumed samples: 16455680 | consumed tokens: 33701232640 | elapsed time per iteration (s): 1.77 | learning rate: 7.982E-05 | global batch size: 512 | lm loss: 1.247538E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.456 | TFLOPs: 43.45 | 31: iteration 32150/ 52452 | consumed samples: 16460800 | consumed tokens: 33711718400 | elapsed time per iteration (s): 1.77 | learning rate: 7.977E-05 | global batch size: 512 | lm loss: 1.255824E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.163 | TFLOPs: 43.40 | 31: iteration 32160/ 52452 | consumed samples: 16465920 | consumed tokens: 33722204160 | elapsed time per iteration (s): 1.83 | learning rate: 7.972E-05 | global batch size: 512 | lm loss: 1.259098E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.382 | TFLOPs: 42.08 | 31: iteration 32170/ 52452 | consumed samples: 16471040 | consumed tokens: 33732689920 | elapsed time per iteration (s): 1.73 | learning rate: 7.967E-05 | global batch size: 512 | lm loss: 1.251150E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.305 | TFLOPs: 44.32 | 31: iteration 32180/ 52452 | consumed samples: 16476160 | consumed tokens: 33743175680 | elapsed time per iteration (s): 1.77 | learning rate: 7.962E-05 | global batch size: 512 | lm loss: 1.258390E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.374 | TFLOPs: 43.43 | 31: iteration 32190/ 52452 | consumed samples: 16481280 | consumed tokens: 33753661440 | elapsed time per iteration (s): 1.82 | learning rate: 7.957E-05 | global batch size: 512 | lm loss: 1.247532E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.723 | TFLOPs: 42.14 | 31: iteration 32200/ 52452 | consumed samples: 16486400 | consumed tokens: 33764147200 | elapsed time per iteration (s): 1.76 | learning rate: 7.952E-05 | global batch size: 512 | lm loss: 1.243466E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.314 | TFLOPs: 43.57 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32200 | lm loss value: 4.170547E+00 | lm loss PPL: 6.475089E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32200 | lm loss value: 4.075847E+00 | lm loss PPL: 5.890036E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32210/ 52452 | consumed samples: 16491520 | consumed tokens: 33774632960 | elapsed time per iteration (s): 9.22 | learning rate: 7.946E-05 | global batch size: 512 | lm loss: 1.263857E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.541 | TFLOPs: 8.34 | 31: iteration 32220/ 52452 | consumed samples: 16496640 | consumed tokens: 33785118720 | elapsed time per iteration (s): 1.76 | learning rate: 7.941E-05 | global batch size: 512 | lm loss: 1.263401E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.042 | TFLOPs: 43.68 | 31: iteration 32230/ 52452 | consumed samples: 16501760 | consumed tokens: 33795604480 | elapsed time per iteration (s): 1.78 | learning rate: 7.936E-05 | global batch size: 512 | lm loss: 1.253452E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.985 | TFLOPs: 43.22 | 31: iteration 32240/ 52452 | consumed samples: 16506880 | consumed tokens: 33806090240 | elapsed time per iteration (s): 1.80 | learning rate: 7.931E-05 | global batch size: 512 | lm loss: 1.257318E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.389 | TFLOPs: 42.69 | 31: iteration 32250/ 52452 | consumed samples: 16512000 | consumed tokens: 33816576000 | elapsed time per iteration (s): 1.76 | learning rate: 7.926E-05 | global batch size: 512 | lm loss: 1.255615E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.128 | TFLOPs: 43.55 | 31: iteration 32260/ 52452 | consumed samples: 16517120 | consumed tokens: 33827061760 | elapsed time per iteration (s): 1.84 | learning rate: 7.921E-05 | global batch size: 512 | lm loss: 1.252673E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.474 | TFLOPs: 41.80 | 31: iteration 32270/ 52452 | consumed samples: 16522240 | consumed tokens: 33837547520 | elapsed time per iteration (s): 1.77 | learning rate: 7.916E-05 | global batch size: 512 | lm loss: 1.250669E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.036 | TFLOPs: 43.38 | 31: iteration 32280/ 52452 | consumed samples: 16527360 | consumed tokens: 33848033280 | elapsed time per iteration (s): 1.82 | learning rate: 7.911E-05 | global batch size: 512 | lm loss: 1.260597E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.862 | TFLOPs: 42.31 | 31: iteration 32290/ 52452 | consumed samples: 16532480 | consumed tokens: 33858519040 | elapsed time per iteration (s): 1.84 | learning rate: 7.905E-05 | global batch size: 512 | lm loss: 1.261053E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.034 | TFLOPs: 41.73 | 31: iteration 32300/ 52452 | consumed samples: 16537600 | consumed tokens: 33869004800 | elapsed time per iteration (s): 1.77 | learning rate: 7.900E-05 | global batch size: 512 | lm loss: 1.245059E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.583 | TFLOPs: 43.31 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32300 | lm loss value: 4.008545E+00 | lm loss PPL: 5.506669E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32300 | lm loss value: 3.954150E+00 | lm loss PPL: 5.215134E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32310/ 52452 | consumed samples: 16542720 | consumed tokens: 33879490560 | elapsed time per iteration (s): 9.15 | learning rate: 7.895E-05 | global batch size: 512 | lm loss: 1.247396E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.980 | TFLOPs: 8.40 | 31: iteration 32320/ 52452 | consumed samples: 16547840 | consumed tokens: 33889976320 | elapsed time per iteration (s): 1.75 | learning rate: 7.890E-05 | global batch size: 512 | lm loss: 1.234735E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.452 | TFLOPs: 43.90 | 31: iteration 32330/ 52452 | consumed samples: 16552960 | consumed tokens: 33900462080 | elapsed time per iteration (s): 1.82 | learning rate: 7.885E-05 | global batch size: 512 | lm loss: 1.251208E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.556 | TFLOPs: 42.26 | 31: iteration 32340/ 52452 | consumed samples: 16558080 | consumed tokens: 33910947840 | elapsed time per iteration (s): 1.75 | learning rate: 7.880E-05 | global batch size: 512 | lm loss: 1.252262E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.976 | TFLOPs: 43.97 | 31: iteration 32350/ 52452 | consumed samples: 16563200 | consumed tokens: 33921433600 | elapsed time per iteration (s): 1.77 | learning rate: 7.875E-05 | global batch size: 512 | lm loss: 1.242386E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.936 | TFLOPs: 43.52 | 31: iteration 32360/ 52452 | consumed samples: 16568320 | consumed tokens: 33931919360 | elapsed time per iteration (s): 1.81 | learning rate: 7.870E-05 | global batch size: 512 | lm loss: 1.248245E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.110 | TFLOPs: 42.34 | 31: iteration 32370/ 52452 | consumed samples: 16573440 | consumed tokens: 33942405120 | elapsed time per iteration (s): 1.84 | learning rate: 7.865E-05 | global batch size: 512 | lm loss: 1.246936E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.610 | TFLOPs: 41.82 | 31: iteration 32380/ 52452 | consumed samples: 16578560 | consumed tokens: 33952890880 | elapsed time per iteration (s): 1.83 | learning rate: 7.859E-05 | global batch size: 512 | lm loss: 1.236509E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.053 | TFLOPs: 41.88 | 31: iteration 32390/ 52452 | consumed samples: 16583680 | consumed tokens: 33963376640 | elapsed time per iteration (s): 1.79 | learning rate: 7.854E-05 | global batch size: 512 | lm loss: 1.264188E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.615 | TFLOPs: 42.87 | 31: iteration 32400/ 52452 | consumed samples: 16588800 | consumed tokens: 33973862400 | elapsed time per iteration (s): 1.78 | learning rate: 7.849E-05 | global batch size: 512 | lm loss: 1.248196E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.915 | TFLOPs: 43.21 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32400 | lm loss value: 4.105745E+00 | lm loss PPL: 6.068796E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32400 | lm loss value: 4.024553E+00 | lm loss PPL: 5.595531E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32410/ 52452 | consumed samples: 16593920 | consumed tokens: 33984348160 | elapsed time per iteration (s): 9.11 | learning rate: 7.844E-05 | global batch size: 512 | lm loss: 1.264555E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.173 | TFLOPs: 8.43 | 31: iteration 32420/ 52452 | consumed samples: 16599040 | consumed tokens: 33994833920 | elapsed time per iteration (s): 1.76 | learning rate: 7.839E-05 | global batch size: 512 | lm loss: 1.243181E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.380 | TFLOPs: 43.73 | 31: iteration 32430/ 52452 | consumed samples: 16604160 | consumed tokens: 34005319680 | elapsed time per iteration (s): 1.76 | learning rate: 7.834E-05 | global batch size: 512 | lm loss: 1.249044E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.420 | TFLOPs: 43.59 | 31: iteration 32440/ 52452 | consumed samples: 16609280 | consumed tokens: 34015805440 | elapsed time per iteration (s): 1.76 | learning rate: 7.829E-05 | global batch size: 512 | lm loss: 1.245684E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.522 | TFLOPs: 43.76 | 31: iteration 32450/ 52452 | consumed samples: 16614400 | consumed tokens: 34026291200 | elapsed time per iteration (s): 1.75 | learning rate: 7.824E-05 | global batch size: 512 | lm loss: 1.255245E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.410 | TFLOPs: 43.89 | 31: iteration 32460/ 52452 | consumed samples: 16619520 | consumed tokens: 34036776960 | elapsed time per iteration (s): 1.87 | learning rate: 7.819E-05 | global batch size: 512 | lm loss: 1.268672E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.303 | TFLOPs: 41.17 | 31: iteration 32470/ 52452 | consumed samples: 16624640 | consumed tokens: 34047262720 | elapsed time per iteration (s): 1.76 | learning rate: 7.814E-05 | global batch size: 512 | lm loss: 1.241777E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.228 | TFLOPs: 43.56 | 31: iteration 32480/ 52452 | consumed samples: 16629760 | consumed tokens: 34057748480 | elapsed time per iteration (s): 1.85 | learning rate: 7.809E-05 | global batch size: 512 | lm loss: 1.258928E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.106 | TFLOPs: 41.44 | 31: iteration 32490/ 52452 | consumed samples: 16634880 | consumed tokens: 34068234240 | elapsed time per iteration (s): 1.77 | learning rate: 7.803E-05 | global batch size: 512 | lm loss: 1.235063E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.767 | TFLOPs: 43.34 | 31: iteration 32500/ 52452 | consumed samples: 16640000 | consumed tokens: 34078720000 | elapsed time per iteration (s): 1.82 | learning rate: 7.798E-05 | global batch size: 512 | lm loss: 1.265728E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.707 | TFLOPs: 42.28 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32500 | lm loss value: 4.060029E+00 | lm loss PPL: 5.797599E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32500 | lm loss value: 3.990627E+00 | lm loss PPL: 5.408880E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32510/ 52452 | consumed samples: 16645120 | consumed tokens: 34089205760 | elapsed time per iteration (s): 9.27 | learning rate: 7.793E-05 | global batch size: 512 | lm loss: 1.252638E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.219 | TFLOPs: 8.29 | 31: iteration 32520/ 52452 | consumed samples: 16650240 | consumed tokens: 34099691520 | elapsed time per iteration (s): 1.77 | learning rate: 7.788E-05 | global batch size: 512 | lm loss: 1.250185E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.814 | TFLOPs: 43.50 | 31: iteration 32530/ 52452 | consumed samples: 16655360 | consumed tokens: 34110177280 | elapsed time per iteration (s): 1.78 | learning rate: 7.783E-05 | global batch size: 512 | lm loss: 1.254531E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.167 | TFLOPs: 43.10 | 31: iteration 32540/ 52452 | consumed samples: 16660480 | consumed tokens: 34120663040 | elapsed time per iteration (s): 1.80 | learning rate: 7.778E-05 | global batch size: 512 | lm loss: 1.253722E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.412 | TFLOPs: 42.69 | 31: iteration 32550/ 52452 | consumed samples: 16665600 | consumed tokens: 34131148800 | elapsed time per iteration (s): 1.76 | learning rate: 7.773E-05 | global batch size: 512 | lm loss: 1.251303E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.672 | TFLOPs: 43.78 | 31: iteration 32560/ 52452 | consumed samples: 16670720 | consumed tokens: 34141634560 | elapsed time per iteration (s): 1.76 | learning rate: 7.768E-05 | global batch size: 512 | lm loss: 1.244718E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.580 | TFLOPs: 43.61 | 31: iteration 32570/ 52452 | consumed samples: 16675840 | consumed tokens: 34152120320 | elapsed time per iteration (s): 1.79 | learning rate: 7.763E-05 | global batch size: 512 | lm loss: 1.252084E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.592 | TFLOPs: 43.02 | 31: iteration 32580/ 52452 | consumed samples: 16680960 | consumed tokens: 34162606080 | elapsed time per iteration (s): 1.73 | learning rate: 7.758E-05 | global batch size: 512 | lm loss: 1.255960E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.326 | TFLOPs: 44.33 | 31: iteration 32590/ 52452 | consumed samples: 16686080 | consumed tokens: 34173091840 | elapsed time per iteration (s): 1.76 | learning rate: 7.753E-05 | global batch size: 512 | lm loss: 1.251511E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.190 | TFLOPs: 43.56 | 31: iteration 32600/ 52452 | consumed samples: 16691200 | consumed tokens: 34183577600 | elapsed time per iteration (s): 1.77 | learning rate: 7.748E-05 | global batch size: 512 | lm loss: 1.247050E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.970 | TFLOPs: 43.37 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32600 | lm loss value: 4.214667E+00 | lm loss PPL: 6.767162E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32600 | lm loss value: 4.096213E+00 | lm loss PPL: 6.011223E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32610/ 52452 | consumed samples: 16696320 | consumed tokens: 34194063360 | elapsed time per iteration (s): 9.45 | learning rate: 7.742E-05 | global batch size: 512 | lm loss: 1.250987E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.170 | TFLOPs: 8.13 | 31: iteration 32620/ 52452 | consumed samples: 16701440 | consumed tokens: 34204549120 | elapsed time per iteration (s): 1.79 | learning rate: 7.737E-05 | global batch size: 512 | lm loss: 1.257004E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.184 | TFLOPs: 42.95 | 31: iteration 32630/ 52452 | consumed samples: 16706560 | consumed tokens: 34215034880 | elapsed time per iteration (s): 1.85 | learning rate: 7.732E-05 | global batch size: 512 | lm loss: 1.263923E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.775 | TFLOPs: 41.54 | 31: iteration 32640/ 52452 | consumed samples: 16711680 | consumed tokens: 34225520640 | elapsed time per iteration (s): 1.88 | learning rate: 7.727E-05 | global batch size: 512 | lm loss: 1.247720E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.229 | TFLOPs: 40.86 | 31: iteration 32650/ 52452 | consumed samples: 16716800 | consumed tokens: 34236006400 | elapsed time per iteration (s): 1.85 | learning rate: 7.722E-05 | global batch size: 512 | lm loss: 1.251138E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.620 | TFLOPs: 41.52 | 31: iteration 32660/ 52452 | consumed samples: 16721920 | consumed tokens: 34246492160 | elapsed time per iteration (s): 1.79 | learning rate: 7.717E-05 | global batch size: 512 | lm loss: 1.243679E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.095 | TFLOPs: 42.94 | 31: iteration 32670/ 52452 | consumed samples: 16727040 | consumed tokens: 34256977920 | elapsed time per iteration (s): 1.80 | learning rate: 7.712E-05 | global batch size: 512 | lm loss: 1.249175E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.192 | TFLOPs: 42.81 | 31: iteration 32680/ 52452 | consumed samples: 16732160 | consumed tokens: 34267463680 | elapsed time per iteration (s): 1.80 | learning rate: 7.707E-05 | global batch size: 512 | lm loss: 1.259912E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.554 | TFLOPs: 42.71 | 31: iteration 32690/ 52452 | consumed samples: 16737280 | consumed tokens: 34277949440 | elapsed time per iteration (s): 1.79 | learning rate: 7.702E-05 | global batch size: 512 | lm loss: 1.244529E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.215 | TFLOPs: 42.96 | 31: iteration 32700/ 52452 | consumed samples: 16742400 | consumed tokens: 34288435200 | elapsed time per iteration (s): 1.77 | learning rate: 7.697E-05 | global batch size: 512 | lm loss: 1.247370E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.270 | TFLOPs: 43.42 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32700 | lm loss value: 4.229328E+00 | lm loss PPL: 6.867108E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32700 | lm loss value: 4.127186E+00 | lm loss PPL: 6.200319E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32710/ 52452 | consumed samples: 16747520 | consumed tokens: 34298920960 | elapsed time per iteration (s): 9.11 | learning rate: 7.692E-05 | global batch size: 512 | lm loss: 1.244128E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.187 | TFLOPs: 8.43 | 31: iteration 32720/ 52452 | consumed samples: 16752640 | consumed tokens: 34309406720 | elapsed time per iteration (s): 1.77 | learning rate: 7.687E-05 | global batch size: 512 | lm loss: 1.255527E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.129 | TFLOPs: 43.40 | 31: iteration 32730/ 52452 | consumed samples: 16757760 | consumed tokens: 34319892480 | elapsed time per iteration (s): 1.79 | learning rate: 7.682E-05 | global batch size: 512 | lm loss: 1.257456E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.754 | TFLOPs: 42.89 | 31: iteration 32740/ 52452 | consumed samples: 16762880 | consumed tokens: 34330378240 | elapsed time per iteration (s): 1.73 | learning rate: 7.677E-05 | global batch size: 512 | lm loss: 1.260697E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.808 | TFLOPs: 44.40 | 31: iteration 32750/ 52452 | consumed samples: 16768000 | consumed tokens: 34340864000 | elapsed time per iteration (s): 1.77 | learning rate: 7.672E-05 | global batch size: 512 | lm loss: 1.253772E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.809 | TFLOPs: 43.35 | 31: iteration 32760/ 52452 | consumed samples: 16773120 | consumed tokens: 34351349760 | elapsed time per iteration (s): 1.78 | learning rate: 7.666E-05 | global batch size: 512 | lm loss: 1.246790E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.808 | TFLOPs: 43.20 | 31: iteration 32770/ 52452 | consumed samples: 16778240 | consumed tokens: 34361835520 | elapsed time per iteration (s): 1.76 | learning rate: 7.661E-05 | global batch size: 512 | lm loss: 1.254036E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.718 | TFLOPs: 43.64 | 31: iteration 32780/ 52452 | consumed samples: 16783360 | consumed tokens: 34372321280 | elapsed time per iteration (s): 1.83 | learning rate: 7.656E-05 | global batch size: 512 | lm loss: 1.244251E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.118 | TFLOPs: 42.04 | 31: iteration 32790/ 52452 | consumed samples: 16788480 | consumed tokens: 34382807040 | elapsed time per iteration (s): 1.75 | learning rate: 7.651E-05 | global batch size: 512 | lm loss: 1.248762E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.960 | TFLOPs: 43.82 | 31: iteration 32800/ 52452 | consumed samples: 16793600 | consumed tokens: 34393292800 | elapsed time per iteration (s): 1.84 | learning rate: 7.646E-05 | global batch size: 512 | lm loss: 1.251194E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.159 | TFLOPs: 41.75 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32800 | lm loss value: 4.111373E+00 | lm loss PPL: 6.103048E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32800 | lm loss value: 4.029698E+00 | lm loss PPL: 5.624392E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32810/ 52452 | consumed samples: 16798720 | consumed tokens: 34403778560 | elapsed time per iteration (s): 9.19 | learning rate: 7.641E-05 | global batch size: 512 | lm loss: 1.258309E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.727 | TFLOPs: 8.36 | 31: iteration 32820/ 52452 | consumed samples: 16803840 | consumed tokens: 34414264320 | elapsed time per iteration (s): 1.81 | learning rate: 7.636E-05 | global batch size: 512 | lm loss: 1.247029E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.625 | TFLOPs: 42.42 | 31: iteration 32830/ 52452 | consumed samples: 16808960 | consumed tokens: 34424750080 | elapsed time per iteration (s): 1.84 | learning rate: 7.631E-05 | global batch size: 512 | lm loss: 1.264644E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.720 | TFLOPs: 41.68 | 31: iteration 32840/ 52452 | consumed samples: 16814080 | consumed tokens: 34435235840 | elapsed time per iteration (s): 1.79 | learning rate: 7.626E-05 | global batch size: 512 | lm loss: 1.234883E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.832 | TFLOPs: 42.90 | 31: iteration 32850/ 52452 | consumed samples: 16819200 | consumed tokens: 34445721600 | elapsed time per iteration (s): 1.79 | learning rate: 7.621E-05 | global batch size: 512 | lm loss: 1.247095E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.239 | TFLOPs: 42.96 | 31: iteration 32860/ 52452 | consumed samples: 16824320 | consumed tokens: 34456207360 | elapsed time per iteration (s): 1.78 | learning rate: 7.616E-05 | global batch size: 512 | lm loss: 1.244936E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.025 | TFLOPs: 43.08 | 31: iteration 32870/ 52452 | consumed samples: 16829440 | consumed tokens: 34466693120 | elapsed time per iteration (s): 1.85 | learning rate: 7.611E-05 | global batch size: 512 | lm loss: 1.265560E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.481 | TFLOPs: 41.50 | 31: iteration 32880/ 52452 | consumed samples: 16834560 | consumed tokens: 34477178880 | elapsed time per iteration (s): 1.77 | learning rate: 7.606E-05 | global batch size: 512 | lm loss: 1.256861E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.581 | TFLOPs: 43.46 | 31: iteration 32890/ 52452 | consumed samples: 16839680 | consumed tokens: 34487664640 | elapsed time per iteration (s): 1.81 | learning rate: 7.601E-05 | global batch size: 512 | lm loss: 1.249554E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.407 | TFLOPs: 42.54 | 31: iteration 32900/ 52452 | consumed samples: 16844800 | consumed tokens: 34498150400 | elapsed time per iteration (s): 1.75 | learning rate: 7.596E-05 | global batch size: 512 | lm loss: 1.250126E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.883 | TFLOPs: 43.81 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 32900 | lm loss value: 4.181015E+00 | lm loss PPL: 6.543223E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 32900 | lm loss value: 4.073375E+00 | lm loss PPL: 5.875491E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 32910/ 52452 | consumed samples: 16849920 | consumed tokens: 34508636160 | elapsed time per iteration (s): 9.14 | learning rate: 7.591E-05 | global batch size: 512 | lm loss: 1.246155E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.035 | TFLOPs: 8.41 | 31: iteration 32920/ 52452 | consumed samples: 16855040 | consumed tokens: 34519121920 | elapsed time per iteration (s): 1.84 | learning rate: 7.586E-05 | global batch size: 512 | lm loss: 1.244627E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.011 | TFLOPs: 41.88 | 31: iteration 32930/ 52452 | consumed samples: 16860160 | consumed tokens: 34529607680 | elapsed time per iteration (s): 1.82 | learning rate: 7.581E-05 | global batch size: 512 | lm loss: 1.249764E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.642 | TFLOPs: 42.12 | 31: iteration 32940/ 52452 | consumed samples: 16865280 | consumed tokens: 34540093440 | elapsed time per iteration (s): 1.80 | learning rate: 7.576E-05 | global batch size: 512 | lm loss: 1.252587E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.983 | TFLOPs: 42.77 | 31: iteration 32950/ 52452 | consumed samples: 16870400 | consumed tokens: 34550579200 | elapsed time per iteration (s): 1.74 | learning rate: 7.571E-05 | global batch size: 512 | lm loss: 1.250440E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.959 | TFLOPs: 44.27 | 31: iteration 32960/ 52452 | consumed samples: 16875520 | consumed tokens: 34561064960 | elapsed time per iteration (s): 1.75 | learning rate: 7.566E-05 | global batch size: 512 | lm loss: 1.243627E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.871 | TFLOPs: 43.81 | 31: iteration 32970/ 52452 | consumed samples: 16880640 | consumed tokens: 34571550720 | elapsed time per iteration (s): 1.82 | learning rate: 7.561E-05 | global batch size: 512 | lm loss: 1.245523E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.618 | TFLOPs: 42.27 | 31: iteration 32980/ 52452 | consumed samples: 16885760 | consumed tokens: 34582036480 | elapsed time per iteration (s): 1.75 | learning rate: 7.555E-05 | global batch size: 512 | lm loss: 1.248732E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.987 | TFLOPs: 43.83 | 31: iteration 32990/ 52452 | consumed samples: 16890880 | consumed tokens: 34592522240 | elapsed time per iteration (s): 1.73 | learning rate: 7.550E-05 | global batch size: 512 | lm loss: 1.255860E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.531 | TFLOPs: 44.51 | 31: iteration 33000/ 52452 | consumed samples: 16896000 | consumed tokens: 34603008000 | elapsed time per iteration (s): 1.74 | learning rate: 7.545E-05 | global batch size: 512 | lm loss: 1.245120E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.414 | TFLOPs: 44.04 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33000 | lm loss value: 4.219450E+00 | lm loss PPL: 6.799608E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33000 | lm loss value: 4.089076E+00 | lm loss PPL: 5.968472E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33010/ 52452 | consumed samples: 16901120 | consumed tokens: 34613493760 | elapsed time per iteration (s): 9.22 | learning rate: 7.540E-05 | global batch size: 512 | lm loss: 1.252343E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.549 | TFLOPs: 8.34 | 31: iteration 33020/ 52452 | consumed samples: 16906240 | consumed tokens: 34623979520 | elapsed time per iteration (s): 1.73 | learning rate: 7.535E-05 | global batch size: 512 | lm loss: 1.248818E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.378 | TFLOPs: 44.33 | 31: iteration 33030/ 52452 | consumed samples: 16911360 | consumed tokens: 34634465280 | elapsed time per iteration (s): 1.74 | learning rate: 7.530E-05 | global batch size: 512 | lm loss: 1.252925E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.975 | TFLOPs: 44.27 | 31: iteration 33040/ 52452 | consumed samples: 16916480 | consumed tokens: 34644951040 | elapsed time per iteration (s): 1.77 | learning rate: 7.525E-05 | global batch size: 512 | lm loss: 1.247520E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.459 | TFLOPs: 43.30 | 31: iteration 33050/ 52452 | consumed samples: 16921600 | consumed tokens: 34655436800 | elapsed time per iteration (s): 1.76 | learning rate: 7.520E-05 | global batch size: 512 | lm loss: 1.238324E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.337 | TFLOPs: 43.73 | 31: iteration 33060/ 52452 | consumed samples: 16926720 | consumed tokens: 34665922560 | elapsed time per iteration (s): 1.78 | learning rate: 7.515E-05 | global batch size: 512 | lm loss: 1.251428E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.291 | TFLOPs: 43.12 | 31: iteration 33070/ 52452 | consumed samples: 16931840 | consumed tokens: 34676408320 | elapsed time per iteration (s): 1.85 | learning rate: 7.510E-05 | global batch size: 512 | lm loss: 1.254972E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.139 | TFLOPs: 41.45 | 31: iteration 33080/ 52452 | consumed samples: 16936960 | consumed tokens: 34686894080 | elapsed time per iteration (s): 1.76 | learning rate: 7.505E-05 | global batch size: 512 | lm loss: 1.260099E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.644 | TFLOPs: 43.77 | 31: iteration 33090/ 52452 | consumed samples: 16942080 | consumed tokens: 34697379840 | elapsed time per iteration (s): 1.77 | learning rate: 7.500E-05 | global batch size: 512 | lm loss: 1.244674E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.735 | TFLOPs: 43.49 | 31: iteration 33100/ 52452 | consumed samples: 16947200 | consumed tokens: 34707865600 | elapsed time per iteration (s): 1.82 | learning rate: 7.495E-05 | global batch size: 512 | lm loss: 1.263106E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.669 | TFLOPs: 42.28 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33100 | lm loss value: 4.101320E+00 | lm loss PPL: 6.042001E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33100 | lm loss value: 4.019818E+00 | lm loss PPL: 5.569099E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33110/ 52452 | consumed samples: 16952320 | consumed tokens: 34718351360 | elapsed time per iteration (s): 9.13 | learning rate: 7.490E-05 | global batch size: 512 | lm loss: 1.230465E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.051 | TFLOPs: 8.41 | 31: iteration 33120/ 52452 | consumed samples: 16957440 | consumed tokens: 34728837120 | elapsed time per iteration (s): 1.76 | learning rate: 7.485E-05 | global batch size: 512 | lm loss: 1.259981E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.582 | TFLOPs: 43.61 | 31: iteration 33130/ 52452 | consumed samples: 16962560 | consumed tokens: 34739322880 | elapsed time per iteration (s): 1.75 | learning rate: 7.480E-05 | global batch size: 512 | lm loss: 1.239038E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.745 | TFLOPs: 43.94 | 31: iteration 33140/ 52452 | consumed samples: 16967680 | consumed tokens: 34749808640 | elapsed time per iteration (s): 1.75 | learning rate: 7.475E-05 | global batch size: 512 | lm loss: 1.245461E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.500 | TFLOPs: 43.90 | 31: iteration 33150/ 52452 | consumed samples: 16972800 | consumed tokens: 34760294400 | elapsed time per iteration (s): 1.74 | learning rate: 7.470E-05 | global batch size: 512 | lm loss: 1.253123E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.850 | TFLOPs: 44.11 | 31: iteration 33160/ 52452 | consumed samples: 16977920 | consumed tokens: 34770780160 | elapsed time per iteration (s): 1.78 | learning rate: 7.465E-05 | global batch size: 512 | lm loss: 1.251013E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.943 | TFLOPs: 43.22 | 31: iteration 33170/ 52452 | consumed samples: 16983040 | consumed tokens: 34781265920 | elapsed time per iteration (s): 1.75 | learning rate: 7.460E-05 | global batch size: 512 | lm loss: 1.255985E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.533 | TFLOPs: 43.91 | 31: iteration 33180/ 52452 | consumed samples: 16988160 | consumed tokens: 34791751680 | elapsed time per iteration (s): 1.80 | learning rate: 7.455E-05 | global batch size: 512 | lm loss: 1.236888E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.954 | TFLOPs: 42.77 | 31: iteration 33190/ 52452 | consumed samples: 16993280 | consumed tokens: 34802237440 | elapsed time per iteration (s): 1.72 | learning rate: 7.450E-05 | global batch size: 512 | lm loss: 1.250155E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.309 | TFLOPs: 44.62 | 31: iteration 33200/ 52452 | consumed samples: 16998400 | consumed tokens: 34812723200 | elapsed time per iteration (s): 1.73 | learning rate: 7.445E-05 | global batch size: 512 | lm loss: 1.255984E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.446 | TFLOPs: 44.49 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33200 | lm loss value: 4.078807E+00 | lm loss PPL: 5.907497E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33200 | lm loss value: 4.003272E+00 | lm loss PPL: 5.477709E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33210/ 52452 | consumed samples: 17003520 | consumed tokens: 34823208960 | elapsed time per iteration (s): 9.24 | learning rate: 7.440E-05 | global batch size: 512 | lm loss: 1.265619E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.426 | TFLOPs: 8.32 | 31: iteration 33220/ 52452 | consumed samples: 17008640 | consumed tokens: 34833694720 | elapsed time per iteration (s): 1.80 | learning rate: 7.435E-05 | global batch size: 512 | lm loss: 1.249749E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.341 | TFLOPs: 42.68 | 31: iteration 33230/ 52452 | consumed samples: 17013760 | consumed tokens: 34844180480 | elapsed time per iteration (s): 1.77 | learning rate: 7.430E-05 | global batch size: 512 | lm loss: 1.253763E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.489 | TFLOPs: 43.45 | 31: iteration 33240/ 52452 | consumed samples: 17018880 | consumed tokens: 34854666240 | elapsed time per iteration (s): 1.81 | learning rate: 7.425E-05 | global batch size: 512 | lm loss: 1.256285E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.989 | TFLOPs: 42.48 | 31: iteration 33250/ 52452 | consumed samples: 17024000 | consumed tokens: 34865152000 | elapsed time per iteration (s): 1.91 | learning rate: 7.420E-05 | global batch size: 512 | lm loss: 1.247821E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.384 | TFLOPs: 40.13 | 31: iteration 33260/ 52452 | consumed samples: 17029120 | consumed tokens: 34875637760 | elapsed time per iteration (s): 1.77 | learning rate: 7.415E-05 | global batch size: 512 | lm loss: 1.250749E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.882 | TFLOPs: 43.36 | 31: iteration 33270/ 52452 | consumed samples: 17034240 | consumed tokens: 34886123520 | elapsed time per iteration (s): 1.84 | learning rate: 7.410E-05 | global batch size: 512 | lm loss: 1.239091E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.033 | TFLOPs: 41.73 | 31: iteration 33280/ 52452 | consumed samples: 17039360 | consumed tokens: 34896609280 | elapsed time per iteration (s): 1.78 | learning rate: 7.405E-05 | global batch size: 512 | lm loss: 1.245613E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.908 | TFLOPs: 43.06 | 31: iteration 33290/ 52452 | consumed samples: 17044480 | consumed tokens: 34907095040 | elapsed time per iteration (s): 1.76 | learning rate: 7.400E-05 | global batch size: 512 | lm loss: 1.242745E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.631 | TFLOPs: 43.62 | 31: iteration 33300/ 52452 | consumed samples: 17049600 | consumed tokens: 34917580800 | elapsed time per iteration (s): 1.82 | learning rate: 7.395E-05 | global batch size: 512 | lm loss: 1.241130E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.649 | TFLOPs: 42.12 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33300 | lm loss value: 4.192530E+00 | lm loss PPL: 6.619005E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33300 | lm loss value: 4.066058E+00 | lm loss PPL: 5.832657E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33310/ 52452 | consumed samples: 17054720 | consumed tokens: 34928066560 | elapsed time per iteration (s): 9.22 | learning rate: 7.390E-05 | global batch size: 512 | lm loss: 1.249417E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.554 | TFLOPs: 8.34 | 31: iteration 33320/ 52452 | consumed samples: 17059840 | consumed tokens: 34938552320 | elapsed time per iteration (s): 1.77 | learning rate: 7.385E-05 | global batch size: 512 | lm loss: 1.241518E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.593 | TFLOPs: 43.47 | 31: iteration 33330/ 52452 | consumed samples: 17064960 | consumed tokens: 34949038080 | elapsed time per iteration (s): 1.77 | learning rate: 7.380E-05 | global batch size: 512 | lm loss: 1.239512E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.986 | TFLOPs: 43.38 | 31: iteration 33340/ 52452 | consumed samples: 17070080 | consumed tokens: 34959523840 | elapsed time per iteration (s): 1.76 | learning rate: 7.375E-05 | global batch size: 512 | lm loss: 1.252515E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.337 | TFLOPs: 43.58 | 31: iteration 33350/ 52452 | consumed samples: 17075200 | consumed tokens: 34970009600 | elapsed time per iteration (s): 1.75 | learning rate: 7.370E-05 | global batch size: 512 | lm loss: 1.253654E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.780 | TFLOPs: 43.79 | 31: iteration 33360/ 52452 | consumed samples: 17080320 | consumed tokens: 34980495360 | elapsed time per iteration (s): 1.78 | learning rate: 7.365E-05 | global batch size: 512 | lm loss: 1.237495E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.132 | TFLOPs: 43.10 | 31: iteration 33370/ 52452 | consumed samples: 17085440 | consumed tokens: 34990981120 | elapsed time per iteration (s): 1.79 | learning rate: 7.360E-05 | global batch size: 512 | lm loss: 1.248228E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.514 | TFLOPs: 42.85 | 31: iteration 33380/ 52452 | consumed samples: 17090560 | consumed tokens: 35001466880 | elapsed time per iteration (s): 1.73 | learning rate: 7.355E-05 | global batch size: 512 | lm loss: 1.250913E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.908 | TFLOPs: 44.41 | 31: iteration 33390/ 52452 | consumed samples: 17095680 | consumed tokens: 35011952640 | elapsed time per iteration (s): 1.78 | learning rate: 7.350E-05 | global batch size: 512 | lm loss: 1.240522E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.683 | TFLOPs: 43.18 | 31: iteration 33400/ 52452 | consumed samples: 17100800 | consumed tokens: 35022438400 | elapsed time per iteration (s): 1.77 | learning rate: 7.345E-05 | global batch size: 512 | lm loss: 1.237560E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.361 | TFLOPs: 43.43 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33400 | lm loss value: 4.169903E+00 | lm loss PPL: 6.470916E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33400 | lm loss value: 4.058118E+00 | lm loss PPL: 5.786530E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33410/ 52452 | consumed samples: 17105920 | consumed tokens: 35032924160 | elapsed time per iteration (s): 9.18 | learning rate: 7.340E-05 | global batch size: 512 | lm loss: 1.246146E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.765 | TFLOPs: 8.37 | 31: iteration 33420/ 52452 | consumed samples: 17111040 | consumed tokens: 35043409920 | elapsed time per iteration (s): 1.89 | learning rate: 7.335E-05 | global batch size: 512 | lm loss: 1.240309E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.726 | TFLOPs: 40.63 | 31: iteration 33430/ 52452 | consumed samples: 17116160 | consumed tokens: 35053895680 | elapsed time per iteration (s): 1.91 | learning rate: 7.330E-05 | global batch size: 512 | lm loss: 1.250455E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.231 | TFLOPs: 40.26 | 31: iteration 33440/ 52452 | consumed samples: 17121280 | consumed tokens: 35064381440 | elapsed time per iteration (s): 1.81 | learning rate: 7.325E-05 | global batch size: 512 | lm loss: 1.238764E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.847 | TFLOPs: 42.45 | 31: iteration 33450/ 52452 | consumed samples: 17126400 | consumed tokens: 35074867200 | elapsed time per iteration (s): 1.75 | learning rate: 7.320E-05 | global batch size: 512 | lm loss: 1.250954E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.166 | TFLOPs: 43.85 | 31: iteration 33460/ 52452 | consumed samples: 17131520 | consumed tokens: 35085352960 | elapsed time per iteration (s): 1.82 | learning rate: 7.315E-05 | global batch size: 512 | lm loss: 1.252742E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.738 | TFLOPs: 42.29 | 31: iteration 33470/ 52452 | consumed samples: 17136640 | consumed tokens: 35095838720 | elapsed time per iteration (s): 1.81 | learning rate: 7.311E-05 | global batch size: 512 | lm loss: 1.259864E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.980 | TFLOPs: 42.47 | 31: iteration 33480/ 52452 | consumed samples: 17141760 | consumed tokens: 35106324480 | elapsed time per iteration (s): 1.78 | learning rate: 7.306E-05 | global batch size: 512 | lm loss: 1.248515E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.983 | TFLOPs: 43.22 | 31: iteration 33490/ 52452 | consumed samples: 17146880 | consumed tokens: 35116810240 | elapsed time per iteration (s): 1.76 | learning rate: 7.301E-05 | global batch size: 512 | lm loss: 1.249290E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.623 | TFLOPs: 43.77 | 31: iteration 33500/ 52452 | consumed samples: 17152000 | consumed tokens: 35127296000 | elapsed time per iteration (s): 1.79 | learning rate: 7.296E-05 | global batch size: 512 | lm loss: 1.246124E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.269 | TFLOPs: 42.82 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33500 | lm loss value: 4.144269E+00 | lm loss PPL: 6.307153E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33500 | lm loss value: 4.051769E+00 | lm loss PPL: 5.749907E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33510/ 52452 | consumed samples: 17157120 | consumed tokens: 35137781760 | elapsed time per iteration (s): 9.19 | learning rate: 7.291E-05 | global batch size: 512 | lm loss: 1.252052E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.694 | TFLOPs: 8.36 | 31: iteration 33520/ 52452 | consumed samples: 17162240 | consumed tokens: 35148267520 | elapsed time per iteration (s): 1.84 | learning rate: 7.286E-05 | global batch size: 512 | lm loss: 1.232622E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.046 | TFLOPs: 41.73 | 31: iteration 33530/ 52452 | consumed samples: 17167360 | consumed tokens: 35158753280 | elapsed time per iteration (s): 1.90 | learning rate: 7.281E-05 | global batch size: 512 | lm loss: 1.248893E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.696 | TFLOPs: 40.48 | 31: iteration 33540/ 52452 | consumed samples: 17172480 | consumed tokens: 35169239040 | elapsed time per iteration (s): 1.86 | learning rate: 7.276E-05 | global batch size: 512 | lm loss: 1.242517E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.003 | TFLOPs: 41.43 | 31: iteration 33550/ 52452 | consumed samples: 17177600 | consumed tokens: 35179724800 | elapsed time per iteration (s): 1.92 | learning rate: 7.271E-05 | global batch size: 512 | lm loss: 1.226882E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.457 | TFLOPs: 39.99 | 31: iteration 33560/ 52452 | consumed samples: 17182720 | consumed tokens: 35190210560 | elapsed time per iteration (s): 1.88 | learning rate: 7.266E-05 | global batch size: 512 | lm loss: 1.236378E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.406 | TFLOPs: 40.89 | 31: iteration 33570/ 52452 | consumed samples: 17187840 | consumed tokens: 35200696320 | elapsed time per iteration (s): 1.79 | learning rate: 7.261E-05 | global batch size: 512 | lm loss: 1.244297E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.393 | TFLOPs: 42.84 | 31: iteration 33580/ 52452 | consumed samples: 17192960 | consumed tokens: 35211182080 | elapsed time per iteration (s): 1.79 | learning rate: 7.256E-05 | global batch size: 512 | lm loss: 1.232150E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.468 | TFLOPs: 43.00 | 31: iteration 33590/ 52452 | consumed samples: 17198080 | consumed tokens: 35221667840 | elapsed time per iteration (s): 1.84 | learning rate: 7.251E-05 | global batch size: 512 | lm loss: 1.252694E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.524 | TFLOPs: 41.65 | 31: iteration 33600/ 52452 | consumed samples: 17203200 | consumed tokens: 35232153600 | elapsed time per iteration (s): 1.86 | learning rate: 7.246E-05 | global batch size: 512 | lm loss: 1.228372E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.875 | TFLOPs: 41.41 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33600 | lm loss value: 4.200422E+00 | lm loss PPL: 6.671450E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33600 | lm loss value: 4.088557E+00 | lm loss PPL: 5.965374E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33610/ 52452 | consumed samples: 17208320 | consumed tokens: 35242639360 | elapsed time per iteration (s): 9.23 | learning rate: 7.241E-05 | global batch size: 512 | lm loss: 1.242230E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.464 | TFLOPs: 8.32 | 31: iteration 33620/ 52452 | consumed samples: 17213440 | consumed tokens: 35253125120 | elapsed time per iteration (s): 1.81 | learning rate: 7.236E-05 | global batch size: 512 | lm loss: 1.242850E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.643 | TFLOPs: 42.42 | 31: iteration 33630/ 52452 | consumed samples: 17218560 | consumed tokens: 35263610880 | elapsed time per iteration (s): 1.81 | learning rate: 7.231E-05 | global batch size: 512 | lm loss: 1.235166E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.377 | TFLOPs: 42.53 | 31: iteration 33640/ 52452 | consumed samples: 17223680 | consumed tokens: 35274096640 | elapsed time per iteration (s): 1.78 | learning rate: 7.226E-05 | global batch size: 512 | lm loss: 1.240042E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.060 | TFLOPs: 43.09 | 31: iteration 33650/ 52452 | consumed samples: 17228800 | consumed tokens: 35284582400 | elapsed time per iteration (s): 1.74 | learning rate: 7.221E-05 | global batch size: 512 | lm loss: 1.239080E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.412 | TFLOPs: 44.04 | 31: iteration 33660/ 52452 | consumed samples: 17233920 | consumed tokens: 35295068160 | elapsed time per iteration (s): 1.80 | learning rate: 7.216E-05 | global batch size: 512 | lm loss: 1.226221E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.847 | TFLOPs: 42.60 | 31: iteration 33670/ 52452 | consumed samples: 17239040 | consumed tokens: 35305553920 | elapsed time per iteration (s): 1.82 | learning rate: 7.211E-05 | global batch size: 512 | lm loss: 1.254515E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.697 | TFLOPs: 42.28 | 31: iteration 33680/ 52452 | consumed samples: 17244160 | consumed tokens: 35316039680 | elapsed time per iteration (s): 1.74 | learning rate: 7.207E-05 | global batch size: 512 | lm loss: 1.247441E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.297 | TFLOPs: 44.17 | 31: iteration 33690/ 52452 | consumed samples: 17249280 | consumed tokens: 35326525440 | elapsed time per iteration (s): 1.79 | learning rate: 7.202E-05 | global batch size: 512 | lm loss: 1.248223E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.753 | TFLOPs: 42.89 | 31: iteration 33700/ 52452 | consumed samples: 17254400 | consumed tokens: 35337011200 | elapsed time per iteration (s): 1.75 | learning rate: 7.197E-05 | global batch size: 512 | lm loss: 1.234536E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.510 | TFLOPs: 43.90 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33700 | lm loss value: 4.164656E+00 | lm loss PPL: 6.437051E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33700 | lm loss value: 4.069583E+00 | lm loss PPL: 5.853255E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33710/ 52452 | consumed samples: 17259520 | consumed tokens: 35347496960 | elapsed time per iteration (s): 9.26 | learning rate: 7.192E-05 | global batch size: 512 | lm loss: 1.231560E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.306 | TFLOPs: 8.30 | 31: iteration 33720/ 52452 | consumed samples: 17264640 | consumed tokens: 35357982720 | elapsed time per iteration (s): 1.77 | learning rate: 7.187E-05 | global batch size: 512 | lm loss: 1.243672E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.001 | TFLOPs: 43.53 | 31: iteration 33730/ 52452 | consumed samples: 17269760 | consumed tokens: 35368468480 | elapsed time per iteration (s): 1.75 | learning rate: 7.182E-05 | global batch size: 512 | lm loss: 1.254538E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.839 | TFLOPs: 43.80 | 31: iteration 33740/ 52452 | consumed samples: 17274880 | consumed tokens: 35378954240 | elapsed time per iteration (s): 1.78 | learning rate: 7.177E-05 | global batch size: 512 | lm loss: 1.248153E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.170 | TFLOPs: 43.25 | 31: iteration 33750/ 52452 | consumed samples: 17280000 | consumed tokens: 35389440000 | elapsed time per iteration (s): 1.86 | learning rate: 7.172E-05 | global batch size: 512 | lm loss: 1.245128E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.187 | TFLOPs: 41.30 | 31: iteration 33760/ 52452 | consumed samples: 17285120 | consumed tokens: 35399925760 | elapsed time per iteration (s): 1.76 | learning rate: 7.167E-05 | global batch size: 512 | lm loss: 1.236444E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.469 | TFLOPs: 43.60 | 31: iteration 33770/ 52452 | consumed samples: 17290240 | consumed tokens: 35410411520 | elapsed time per iteration (s): 1.74 | learning rate: 7.162E-05 | global batch size: 512 | lm loss: 1.230012E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.395 | TFLOPs: 44.19 | 31: iteration 33780/ 52452 | consumed samples: 17295360 | consumed tokens: 35420897280 | elapsed time per iteration (s): 1.76 | learning rate: 7.157E-05 | global batch size: 512 | lm loss: 1.231734E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.779 | TFLOPs: 43.64 | 31: iteration 33790/ 52452 | consumed samples: 17300480 | consumed tokens: 35431383040 | elapsed time per iteration (s): 1.76 | learning rate: 7.152E-05 | global batch size: 512 | lm loss: 1.240158E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.584 | TFLOPs: 43.62 | 31: iteration 33800/ 52452 | consumed samples: 17305600 | consumed tokens: 35441868800 | elapsed time per iteration (s): 1.77 | learning rate: 7.147E-05 | global batch size: 512 | lm loss: 1.221348E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.944 | TFLOPs: 43.37 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33800 | lm loss value: 4.155145E+00 | lm loss PPL: 6.376122E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33800 | lm loss value: 4.056421E+00 | lm loss PPL: 5.776721E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33810/ 52452 | consumed samples: 17310720 | consumed tokens: 35452354560 | elapsed time per iteration (s): 9.19 | learning rate: 7.142E-05 | global batch size: 512 | lm loss: 1.225037E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.708 | TFLOPs: 8.36 | 31: iteration 33820/ 52452 | consumed samples: 17315840 | consumed tokens: 35462840320 | elapsed time per iteration (s): 1.73 | learning rate: 7.138E-05 | global batch size: 512 | lm loss: 1.247774E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.936 | TFLOPs: 44.42 | 31: iteration 33830/ 52452 | consumed samples: 17320960 | consumed tokens: 35473326080 | elapsed time per iteration (s): 1.79 | learning rate: 7.133E-05 | global batch size: 512 | lm loss: 1.236747E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.567 | TFLOPs: 43.01 | 31: iteration 33840/ 52452 | consumed samples: 17326080 | consumed tokens: 35483811840 | elapsed time per iteration (s): 1.77 | learning rate: 7.128E-05 | global batch size: 512 | lm loss: 1.256045E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.390 | TFLOPs: 43.44 | 31: iteration 33850/ 52452 | consumed samples: 17331200 | consumed tokens: 35494297600 | elapsed time per iteration (s): 1.78 | learning rate: 7.123E-05 | global batch size: 512 | lm loss: 1.246473E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.222 | TFLOPs: 43.26 | 31: iteration 33860/ 52452 | consumed samples: 17336320 | consumed tokens: 35504783360 | elapsed time per iteration (s): 1.80 | learning rate: 7.118E-05 | global batch size: 512 | lm loss: 1.233342E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.265 | TFLOPs: 42.67 | 31: iteration 33870/ 52452 | consumed samples: 17341440 | consumed tokens: 35515269120 | elapsed time per iteration (s): 1.76 | learning rate: 7.113E-05 | global batch size: 512 | lm loss: 1.245882E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.497 | TFLOPs: 43.75 | 31: iteration 33880/ 52452 | consumed samples: 17346560 | consumed tokens: 35525754880 | elapsed time per iteration (s): 1.75 | learning rate: 7.108E-05 | global batch size: 512 | lm loss: 1.242368E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.263 | TFLOPs: 43.87 | 31: iteration 33890/ 52452 | consumed samples: 17351680 | consumed tokens: 35536240640 | elapsed time per iteration (s): 1.79 | learning rate: 7.103E-05 | global batch size: 512 | lm loss: 1.240568E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.662 | TFLOPs: 42.88 | 31: iteration 33900/ 52452 | consumed samples: 17356800 | consumed tokens: 35546726400 | elapsed time per iteration (s): 1.79 | learning rate: 7.098E-05 | global batch size: 512 | lm loss: 1.238217E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.876 | TFLOPs: 42.91 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 33900 | lm loss value: 4.118505E+00 | lm loss PPL: 6.146725E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 33900 | lm loss value: 4.027742E+00 | lm loss PPL: 5.613401E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 33910/ 52452 | consumed samples: 17361920 | consumed tokens: 35557212160 | elapsed time per iteration (s): 9.21 | learning rate: 7.093E-05 | global batch size: 512 | lm loss: 1.227813E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.605 | TFLOPs: 8.35 | 31: iteration 33920/ 52452 | consumed samples: 17367040 | consumed tokens: 35567697920 | elapsed time per iteration (s): 1.73 | learning rate: 7.088E-05 | global batch size: 512 | lm loss: 1.237406E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.301 | TFLOPs: 44.32 | 31: iteration 33930/ 52452 | consumed samples: 17372160 | consumed tokens: 35578183680 | elapsed time per iteration (s): 1.79 | learning rate: 7.084E-05 | global batch size: 512 | lm loss: 1.244923E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.816 | TFLOPs: 43.05 | 31: iteration 33940/ 52452 | consumed samples: 17377280 | consumed tokens: 35588669440 | elapsed time per iteration (s): 1.78 | learning rate: 7.079E-05 | global batch size: 512 | lm loss: 1.247563E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.105 | TFLOPs: 43.24 | 31: iteration 33950/ 52452 | consumed samples: 17382400 | consumed tokens: 35599155200 | elapsed time per iteration (s): 1.75 | learning rate: 7.074E-05 | global batch size: 512 | lm loss: 1.236757E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.908 | TFLOPs: 43.81 | 31: iteration 33960/ 52452 | consumed samples: 17387520 | consumed tokens: 35609640960 | elapsed time per iteration (s): 1.83 | learning rate: 7.069E-05 | global batch size: 512 | lm loss: 1.233610E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.724 | TFLOPs: 41.98 | 31: iteration 33970/ 52452 | consumed samples: 17392640 | consumed tokens: 35620126720 | elapsed time per iteration (s): 1.77 | learning rate: 7.064E-05 | global batch size: 512 | lm loss: 1.243602E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.753 | TFLOPs: 43.49 | 31: iteration 33980/ 52452 | consumed samples: 17397760 | consumed tokens: 35630612480 | elapsed time per iteration (s): 1.73 | learning rate: 7.059E-05 | global batch size: 512 | lm loss: 1.241968E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.511 | TFLOPs: 44.35 | 31: iteration 33990/ 52452 | consumed samples: 17402880 | consumed tokens: 35641098240 | elapsed time per iteration (s): 1.74 | learning rate: 7.054E-05 | global batch size: 512 | lm loss: 1.239670E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.923 | TFLOPs: 44.12 | 0: [2022-12-26 08:55:52,307] [INFO] [logging.py:68:log_dist] [Rank 0] step=34000, skipped=0, lr=[7.049254357622476e-05, 7.049254357622476e-05, 7.049254357622476e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 34000/ 52452 | consumed samples: 17408000 | consumed tokens: 35651584000 | elapsed time per iteration (s): 1.75 | learning rate: 7.049E-05 | global batch size: 512 | lm loss: 1.243021E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.075 | TFLOPs: 43.84 | 0: steps: 34000 loss: 1.2401 iter time (s): 1.789 samples/sec: 286.186 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34000 | lm loss value: 4.148736E+00 | lm loss PPL: 6.335390E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34000 | lm loss value: 4.057475E+00 | lm loss PPL: 5.782812E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34010/ 52452 | consumed samples: 17413120 | consumed tokens: 35662069760 | elapsed time per iteration (s): 9.10 | learning rate: 7.044E-05 | global batch size: 512 | lm loss: 1.243306E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.249 | TFLOPs: 8.44 | 31: iteration 34020/ 52452 | consumed samples: 17418240 | consumed tokens: 35672555520 | elapsed time per iteration (s): 1.77 | learning rate: 7.039E-05 | global batch size: 512 | lm loss: 1.231641E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.556 | TFLOPs: 43.31 | 31: iteration 34030/ 52452 | consumed samples: 17423360 | consumed tokens: 35683041280 | elapsed time per iteration (s): 1.75 | learning rate: 7.035E-05 | global batch size: 512 | lm loss: 1.248195E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.375 | TFLOPs: 44.03 | 31: iteration 34040/ 52452 | consumed samples: 17428480 | consumed tokens: 35693527040 | elapsed time per iteration (s): 1.78 | learning rate: 7.030E-05 | global batch size: 512 | lm loss: 1.236076E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.134 | TFLOPs: 43.10 | 31: iteration 34050/ 52452 | consumed samples: 17433600 | consumed tokens: 35704012800 | elapsed time per iteration (s): 1.75 | learning rate: 7.025E-05 | global batch size: 512 | lm loss: 1.240721E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.902 | TFLOPs: 43.96 | 31: iteration 34060/ 52452 | consumed samples: 17438720 | consumed tokens: 35714498560 | elapsed time per iteration (s): 1.83 | learning rate: 7.020E-05 | global batch size: 512 | lm loss: 1.232089E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.280 | TFLOPs: 42.07 | 31: iteration 34070/ 52452 | consumed samples: 17443840 | consumed tokens: 35724984320 | elapsed time per iteration (s): 1.75 | learning rate: 7.015E-05 | global batch size: 512 | lm loss: 1.242645E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.205 | TFLOPs: 43.86 | 31: iteration 34080/ 52452 | consumed samples: 17448960 | consumed tokens: 35735470080 | elapsed time per iteration (s): 1.76 | learning rate: 7.010E-05 | global batch size: 512 | lm loss: 1.239816E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.258 | TFLOPs: 43.57 | 31: iteration 34090/ 52452 | consumed samples: 17454080 | consumed tokens: 35745955840 | elapsed time per iteration (s): 1.82 | learning rate: 7.005E-05 | global batch size: 512 | lm loss: 1.249353E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.953 | TFLOPs: 42.17 | 31: iteration 34100/ 52452 | consumed samples: 17459200 | consumed tokens: 35756441600 | elapsed time per iteration (s): 1.75 | learning rate: 7.000E-05 | global batch size: 512 | lm loss: 1.239534E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.362 | TFLOPs: 43.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34100 | lm loss value: 4.119075E+00 | lm loss PPL: 6.150232E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34100 | lm loss value: 4.039594E+00 | lm loss PPL: 5.680329E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34110/ 52452 | consumed samples: 17464320 | consumed tokens: 35766927360 | elapsed time per iteration (s): 9.13 | learning rate: 6.996E-05 | global batch size: 512 | lm loss: 1.246773E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.067 | TFLOPs: 8.42 | 31: iteration 34120/ 52452 | consumed samples: 17469440 | consumed tokens: 35777413120 | elapsed time per iteration (s): 1.78 | learning rate: 6.991E-05 | global batch size: 512 | lm loss: 1.233917E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.228 | TFLOPs: 43.11 | 31: iteration 34130/ 52452 | consumed samples: 17474560 | consumed tokens: 35787898880 | elapsed time per iteration (s): 1.83 | learning rate: 6.986E-05 | global batch size: 512 | lm loss: 1.230922E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.881 | TFLOPs: 42.01 | 31: iteration 34140/ 52452 | consumed samples: 17479680 | consumed tokens: 35798384640 | elapsed time per iteration (s): 1.83 | learning rate: 6.981E-05 | global batch size: 512 | lm loss: 1.240514E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.133 | TFLOPs: 42.05 | 31: iteration 34150/ 52452 | consumed samples: 17484800 | consumed tokens: 35808870400 | elapsed time per iteration (s): 1.81 | learning rate: 6.976E-05 | global batch size: 512 | lm loss: 1.250114E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.812 | TFLOPs: 42.45 | 31: iteration 34160/ 52452 | consumed samples: 17489920 | consumed tokens: 35819356160 | elapsed time per iteration (s): 1.78 | learning rate: 6.971E-05 | global batch size: 512 | lm loss: 1.227862E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.006 | TFLOPs: 43.23 | 31: iteration 34170/ 52452 | consumed samples: 17495040 | consumed tokens: 35829841920 | elapsed time per iteration (s): 1.82 | learning rate: 6.966E-05 | global batch size: 512 | lm loss: 1.224872E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.250 | TFLOPs: 42.21 | 31: iteration 34180/ 52452 | consumed samples: 17500160 | consumed tokens: 35840327680 | elapsed time per iteration (s): 1.79 | learning rate: 6.961E-05 | global batch size: 512 | lm loss: 1.227922E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.973 | TFLOPs: 42.92 | 31: iteration 34190/ 52452 | consumed samples: 17505280 | consumed tokens: 35850813440 | elapsed time per iteration (s): 1.80 | learning rate: 6.957E-05 | global batch size: 512 | lm loss: 1.232860E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.686 | TFLOPs: 42.58 | 31: iteration 34200/ 52452 | consumed samples: 17510400 | consumed tokens: 35861299200 | elapsed time per iteration (s): 1.77 | learning rate: 6.952E-05 | global batch size: 512 | lm loss: 1.239371E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.071 | TFLOPs: 43.39 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34200 | lm loss value: 4.178185E+00 | lm loss PPL: 6.524735E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34200 | lm loss value: 4.076197E+00 | lm loss PPL: 5.892095E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34210/ 52452 | consumed samples: 17515520 | consumed tokens: 35871784960 | elapsed time per iteration (s): 9.37 | learning rate: 6.947E-05 | global batch size: 512 | lm loss: 1.231081E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.637 | TFLOPs: 8.20 | 31: iteration 34220/ 52452 | consumed samples: 17520640 | consumed tokens: 35882270720 | elapsed time per iteration (s): 1.82 | learning rate: 6.942E-05 | global batch size: 512 | lm loss: 1.243000E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.999 | TFLOPs: 42.18 | 31: iteration 34230/ 52452 | consumed samples: 17525760 | consumed tokens: 35892756480 | elapsed time per iteration (s): 1.79 | learning rate: 6.937E-05 | global batch size: 512 | lm loss: 1.244844E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.344 | TFLOPs: 42.98 | 31: iteration 34240/ 52452 | consumed samples: 17530880 | consumed tokens: 35903242240 | elapsed time per iteration (s): 1.81 | learning rate: 6.932E-05 | global batch size: 512 | lm loss: 1.234117E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.252 | TFLOPs: 42.36 | 31: iteration 34250/ 52452 | consumed samples: 17536000 | consumed tokens: 35913728000 | elapsed time per iteration (s): 1.82 | learning rate: 6.927E-05 | global batch size: 512 | lm loss: 1.235616E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.614 | TFLOPs: 42.12 | 31: iteration 34260/ 52452 | consumed samples: 17541120 | consumed tokens: 35924213760 | elapsed time per iteration (s): 1.82 | learning rate: 6.923E-05 | global batch size: 512 | lm loss: 1.243688E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.324 | TFLOPs: 42.23 | 31: iteration 34270/ 52452 | consumed samples: 17546240 | consumed tokens: 35934699520 | elapsed time per iteration (s): 1.86 | learning rate: 6.918E-05 | global batch size: 512 | lm loss: 1.250465E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.747 | TFLOPs: 41.39 | 31: iteration 34280/ 52452 | consumed samples: 17551360 | consumed tokens: 35945185280 | elapsed time per iteration (s): 1.81 | learning rate: 6.913E-05 | global batch size: 512 | lm loss: 1.244618E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.585 | TFLOPs: 42.56 | 31: iteration 34290/ 52452 | consumed samples: 17556480 | consumed tokens: 35955671040 | elapsed time per iteration (s): 1.79 | learning rate: 6.908E-05 | global batch size: 512 | lm loss: 1.242404E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.120 | TFLOPs: 42.94 | 31: iteration 34300/ 52452 | consumed samples: 17561600 | consumed tokens: 35966156800 | elapsed time per iteration (s): 1.83 | learning rate: 6.903E-05 | global batch size: 512 | lm loss: 1.242808E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.307 | TFLOPs: 42.07 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34300 | lm loss value: 4.032204E+00 | lm loss PPL: 5.638506E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34300 | lm loss value: 3.965368E+00 | lm loss PPL: 5.273969E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34310/ 52452 | consumed samples: 17566720 | consumed tokens: 35976642560 | elapsed time per iteration (s): 9.24 | learning rate: 6.898E-05 | global batch size: 512 | lm loss: 1.251914E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.398 | TFLOPs: 8.31 | 31: iteration 34320/ 52452 | consumed samples: 17571840 | consumed tokens: 35987128320 | elapsed time per iteration (s): 1.75 | learning rate: 6.893E-05 | global batch size: 512 | lm loss: 1.244303E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.177 | TFLOPs: 43.85 | 31: iteration 34330/ 52452 | consumed samples: 17576960 | consumed tokens: 35997614080 | elapsed time per iteration (s): 1.92 | learning rate: 6.889E-05 | global batch size: 512 | lm loss: 1.233328E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.416 | TFLOPs: 39.99 | 31: iteration 34340/ 52452 | consumed samples: 17582080 | consumed tokens: 36008099840 | elapsed time per iteration (s): 1.74 | learning rate: 6.884E-05 | global batch size: 512 | lm loss: 1.243908E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.595 | TFLOPs: 44.22 | 31: iteration 34350/ 52452 | consumed samples: 17587200 | consumed tokens: 36018585600 | elapsed time per iteration (s): 1.76 | learning rate: 6.879E-05 | global batch size: 512 | lm loss: 1.231565E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.563 | TFLOPs: 43.61 | 31: iteration 34360/ 52452 | consumed samples: 17592320 | consumed tokens: 36029071360 | elapsed time per iteration (s): 1.76 | learning rate: 6.874E-05 | global batch size: 512 | lm loss: 1.231734E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.030 | TFLOPs: 43.68 | 31: iteration 34370/ 52452 | consumed samples: 17597440 | consumed tokens: 36039557120 | elapsed time per iteration (s): 1.73 | learning rate: 6.869E-05 | global batch size: 512 | lm loss: 1.231597E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.893 | TFLOPs: 44.41 | 31: iteration 34380/ 52452 | consumed samples: 17602560 | consumed tokens: 36050042880 | elapsed time per iteration (s): 1.73 | learning rate: 6.864E-05 | global batch size: 512 | lm loss: 1.233209E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.991 | TFLOPs: 44.43 | 31: iteration 34390/ 52452 | consumed samples: 17607680 | consumed tokens: 36060528640 | elapsed time per iteration (s): 1.76 | learning rate: 6.860E-05 | global batch size: 512 | lm loss: 1.239221E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.507 | TFLOPs: 43.75 | 31: iteration 34400/ 52452 | consumed samples: 17612800 | consumed tokens: 36071014400 | elapsed time per iteration (s): 1.85 | learning rate: 6.855E-05 | global batch size: 512 | lm loss: 1.230086E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.331 | TFLOPs: 41.63 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34400 | lm loss value: 4.219881E+00 | lm loss PPL: 6.802536E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34400 | lm loss value: 4.110678E+00 | lm loss PPL: 6.098807E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34410/ 52452 | consumed samples: 17617920 | consumed tokens: 36081500160 | elapsed time per iteration (s): 9.19 | learning rate: 6.850E-05 | global batch size: 512 | lm loss: 1.239297E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.720 | TFLOPs: 8.36 | 31: iteration 34420/ 52452 | consumed samples: 17623040 | consumed tokens: 36091985920 | elapsed time per iteration (s): 1.87 | learning rate: 6.845E-05 | global batch size: 512 | lm loss: 1.239696E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.459 | TFLOPs: 41.19 | 31: iteration 34430/ 52452 | consumed samples: 17628160 | consumed tokens: 36102471680 | elapsed time per iteration (s): 1.76 | learning rate: 6.840E-05 | global batch size: 512 | lm loss: 1.235374E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.714 | TFLOPs: 43.63 | 31: iteration 34440/ 52452 | consumed samples: 17633280 | consumed tokens: 36112957440 | elapsed time per iteration (s): 1.75 | learning rate: 6.835E-05 | global batch size: 512 | lm loss: 1.241617E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.077 | TFLOPs: 43.99 | 31: iteration 34450/ 52452 | consumed samples: 17638400 | consumed tokens: 36123443200 | elapsed time per iteration (s): 1.81 | learning rate: 6.831E-05 | global batch size: 512 | lm loss: 1.241238E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.615 | TFLOPs: 42.42 | 31: iteration 34460/ 52452 | consumed samples: 17643520 | consumed tokens: 36133928960 | elapsed time per iteration (s): 1.88 | learning rate: 6.826E-05 | global batch size: 512 | lm loss: 1.232550E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.748 | TFLOPs: 40.79 | 31: iteration 34470/ 52452 | consumed samples: 17648640 | consumed tokens: 36144414720 | elapsed time per iteration (s): 1.74 | learning rate: 6.821E-05 | global batch size: 512 | lm loss: 1.237070E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.457 | TFLOPs: 44.05 | 31: iteration 34480/ 52452 | consumed samples: 17653760 | consumed tokens: 36154900480 | elapsed time per iteration (s): 1.85 | learning rate: 6.816E-05 | global batch size: 512 | lm loss: 1.229425E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.438 | TFLOPs: 41.49 | 31: iteration 34490/ 52452 | consumed samples: 17658880 | consumed tokens: 36165386240 | elapsed time per iteration (s): 1.80 | learning rate: 6.811E-05 | global batch size: 512 | lm loss: 1.228190E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.850 | TFLOPs: 42.60 | 31: iteration 34500/ 52452 | consumed samples: 17664000 | consumed tokens: 36175872000 | elapsed time per iteration (s): 1.81 | learning rate: 6.806E-05 | global batch size: 512 | lm loss: 1.230946E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.738 | TFLOPs: 42.44 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34500 | lm loss value: 4.199604E+00 | lm loss PPL: 6.665993E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34500 | lm loss value: 4.080986E+00 | lm loss PPL: 5.920379E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34510/ 52452 | consumed samples: 17669120 | consumed tokens: 36186357760 | elapsed time per iteration (s): 9.18 | learning rate: 6.802E-05 | global batch size: 512 | lm loss: 1.237482E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.789 | TFLOPs: 8.37 | 31: iteration 34520/ 52452 | consumed samples: 17674240 | consumed tokens: 36196843520 | elapsed time per iteration (s): 1.86 | learning rate: 6.797E-05 | global batch size: 512 | lm loss: 1.230593E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.668 | TFLOPs: 41.23 | 31: iteration 34530/ 52452 | consumed samples: 17679360 | consumed tokens: 36207329280 | elapsed time per iteration (s): 1.78 | learning rate: 6.792E-05 | global batch size: 512 | lm loss: 1.229677E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.554 | TFLOPs: 43.16 | 31: iteration 34540/ 52452 | consumed samples: 17684480 | consumed tokens: 36217815040 | elapsed time per iteration (s): 1.85 | learning rate: 6.787E-05 | global batch size: 512 | lm loss: 1.237290E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.279 | TFLOPs: 41.62 | 31: iteration 34550/ 52452 | consumed samples: 17689600 | consumed tokens: 36228300800 | elapsed time per iteration (s): 1.82 | learning rate: 6.782E-05 | global batch size: 512 | lm loss: 1.237344E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.817 | TFLOPs: 42.15 | 31: iteration 34560/ 52452 | consumed samples: 17694720 | consumed tokens: 36238786560 | elapsed time per iteration (s): 1.75 | learning rate: 6.778E-05 | global batch size: 512 | lm loss: 1.232019E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.216 | TFLOPs: 44.01 | 31: iteration 34570/ 52452 | consumed samples: 17699840 | consumed tokens: 36249272320 | elapsed time per iteration (s): 1.77 | learning rate: 6.773E-05 | global batch size: 512 | lm loss: 1.240711E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.833 | TFLOPs: 43.35 | 31: iteration 34580/ 52452 | consumed samples: 17704960 | consumed tokens: 36259758080 | elapsed time per iteration (s): 1.77 | learning rate: 6.768E-05 | global batch size: 512 | lm loss: 1.240304E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.941 | TFLOPs: 43.37 | 31: iteration 34590/ 52452 | consumed samples: 17710080 | consumed tokens: 36270243840 | elapsed time per iteration (s): 2.01 | learning rate: 6.763E-05 | global batch size: 512 | lm loss: 1.223745E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 254.507 | TFLOPs: 38.20 | 31: iteration 34600/ 52452 | consumed samples: 17715200 | consumed tokens: 36280729600 | elapsed time per iteration (s): 1.76 | learning rate: 6.758E-05 | global batch size: 512 | lm loss: 1.237510E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.568 | TFLOPs: 43.61 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34600 | lm loss value: 4.118065E+00 | lm loss PPL: 6.144023E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34600 | lm loss value: 4.021238E+00 | lm loss PPL: 5.577012E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34610/ 52452 | consumed samples: 17720320 | consumed tokens: 36291215360 | elapsed time per iteration (s): 9.12 | learning rate: 6.754E-05 | global batch size: 512 | lm loss: 1.227028E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.142 | TFLOPs: 8.43 | 31: iteration 34620/ 52452 | consumed samples: 17725440 | consumed tokens: 36301701120 | elapsed time per iteration (s): 1.81 | learning rate: 6.749E-05 | global batch size: 512 | lm loss: 1.236974E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.788 | TFLOPs: 42.44 | 31: iteration 34630/ 52452 | consumed samples: 17730560 | consumed tokens: 36312186880 | elapsed time per iteration (s): 1.78 | learning rate: 6.744E-05 | global batch size: 512 | lm loss: 1.227318E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.835 | TFLOPs: 43.05 | 31: iteration 34640/ 52452 | consumed samples: 17735680 | consumed tokens: 36322672640 | elapsed time per iteration (s): 1.82 | learning rate: 6.739E-05 | global batch size: 512 | lm loss: 1.239159E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.780 | TFLOPs: 42.14 | 31: iteration 34650/ 52452 | consumed samples: 17740800 | consumed tokens: 36333158400 | elapsed time per iteration (s): 1.76 | learning rate: 6.734E-05 | global batch size: 512 | lm loss: 1.236712E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.469 | TFLOPs: 43.75 | 31: iteration 34660/ 52452 | consumed samples: 17745920 | consumed tokens: 36343644160 | elapsed time per iteration (s): 1.84 | learning rate: 6.730E-05 | global batch size: 512 | lm loss: 1.238540E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.768 | TFLOPs: 41.84 | 31: iteration 34670/ 52452 | consumed samples: 17751040 | consumed tokens: 36354129920 | elapsed time per iteration (s): 1.81 | learning rate: 6.725E-05 | global batch size: 512 | lm loss: 1.219889E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.827 | TFLOPs: 42.45 | 31: iteration 34680/ 52452 | consumed samples: 17756160 | consumed tokens: 36364615680 | elapsed time per iteration (s): 1.76 | learning rate: 6.720E-05 | global batch size: 512 | lm loss: 1.246066E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.701 | TFLOPs: 43.78 | 31: iteration 34690/ 52452 | consumed samples: 17761280 | consumed tokens: 36375101440 | elapsed time per iteration (s): 1.77 | learning rate: 6.715E-05 | global batch size: 512 | lm loss: 1.243820E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.249 | TFLOPs: 43.41 | 31: iteration 34700/ 52452 | consumed samples: 17766400 | consumed tokens: 36385587200 | elapsed time per iteration (s): 1.76 | learning rate: 6.710E-05 | global batch size: 512 | lm loss: 1.234179E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.266 | TFLOPs: 43.72 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34700 | lm loss value: 4.178973E+00 | lm loss PPL: 6.529877E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34700 | lm loss value: 4.065014E+00 | lm loss PPL: 5.826575E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34710/ 52452 | consumed samples: 17771520 | consumed tokens: 36396072960 | elapsed time per iteration (s): 9.37 | learning rate: 6.706E-05 | global batch size: 512 | lm loss: 1.237873E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.618 | TFLOPs: 8.20 | 31: iteration 34720/ 52452 | consumed samples: 17776640 | consumed tokens: 36406558720 | elapsed time per iteration (s): 1.79 | learning rate: 6.701E-05 | global batch size: 512 | lm loss: 1.233078E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.262 | TFLOPs: 42.97 | 31: iteration 34730/ 52452 | consumed samples: 17781760 | consumed tokens: 36417044480 | elapsed time per iteration (s): 1.74 | learning rate: 6.696E-05 | global batch size: 512 | lm loss: 1.230127E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.024 | TFLOPs: 44.28 | 31: iteration 34740/ 52452 | consumed samples: 17786880 | consumed tokens: 36427530240 | elapsed time per iteration (s): 1.72 | learning rate: 6.691E-05 | global batch size: 512 | lm loss: 1.220883E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.289 | TFLOPs: 44.62 | 31: iteration 34750/ 52452 | consumed samples: 17792000 | consumed tokens: 36438016000 | elapsed time per iteration (s): 1.78 | learning rate: 6.687E-05 | global batch size: 512 | lm loss: 1.233974E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.690 | TFLOPs: 43.18 | 31: iteration 34760/ 52452 | consumed samples: 17797120 | consumed tokens: 36448501760 | elapsed time per iteration (s): 1.75 | learning rate: 6.682E-05 | global batch size: 512 | lm loss: 1.225858E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.769 | TFLOPs: 43.94 | 31: iteration 34770/ 52452 | consumed samples: 17802240 | consumed tokens: 36458987520 | elapsed time per iteration (s): 1.75 | learning rate: 6.677E-05 | global batch size: 512 | lm loss: 1.236026E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.350 | TFLOPs: 43.88 | 31: iteration 34780/ 52452 | consumed samples: 17807360 | consumed tokens: 36469473280 | elapsed time per iteration (s): 1.81 | learning rate: 6.672E-05 | global batch size: 512 | lm loss: 1.237379E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.825 | TFLOPs: 42.45 | 31: iteration 34790/ 52452 | consumed samples: 17812480 | consumed tokens: 36479959040 | elapsed time per iteration (s): 1.80 | learning rate: 6.667E-05 | global batch size: 512 | lm loss: 1.243962E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.036 | TFLOPs: 42.78 | 31: iteration 34800/ 52452 | consumed samples: 17817600 | consumed tokens: 36490444800 | elapsed time per iteration (s): 1.74 | learning rate: 6.663E-05 | global batch size: 512 | lm loss: 1.242023E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.018 | TFLOPs: 44.28 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34800 | lm loss value: 4.248795E+00 | lm loss PPL: 7.002099E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34800 | lm loss value: 4.119670E+00 | lm loss PPL: 6.153893E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34810/ 52452 | consumed samples: 17822720 | consumed tokens: 36500930560 | elapsed time per iteration (s): 9.11 | learning rate: 6.658E-05 | global batch size: 512 | lm loss: 1.232720E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.227 | TFLOPs: 8.44 | 31: iteration 34820/ 52452 | consumed samples: 17827840 | consumed tokens: 36511416320 | elapsed time per iteration (s): 1.80 | learning rate: 6.653E-05 | global batch size: 512 | lm loss: 1.229445E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.174 | TFLOPs: 42.65 | 31: iteration 34830/ 52452 | consumed samples: 17832960 | consumed tokens: 36521902080 | elapsed time per iteration (s): 1.72 | learning rate: 6.648E-05 | global batch size: 512 | lm loss: 1.240652E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.449 | TFLOPs: 44.65 | 31: iteration 34840/ 52452 | consumed samples: 17838080 | consumed tokens: 36532387840 | elapsed time per iteration (s): 1.87 | learning rate: 6.644E-05 | global batch size: 512 | lm loss: 1.224725E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.944 | TFLOPs: 41.12 | 31: iteration 34850/ 52452 | consumed samples: 17843200 | consumed tokens: 36542873600 | elapsed time per iteration (s): 1.73 | learning rate: 6.639E-05 | global batch size: 512 | lm loss: 1.239620E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.318 | TFLOPs: 44.33 | 31: iteration 34860/ 52452 | consumed samples: 17848320 | consumed tokens: 36553359360 | elapsed time per iteration (s): 1.80 | learning rate: 6.634E-05 | global batch size: 512 | lm loss: 1.233902E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.143 | TFLOPs: 42.65 | 31: iteration 34870/ 52452 | consumed samples: 17853440 | consumed tokens: 36563845120 | elapsed time per iteration (s): 1.77 | learning rate: 6.629E-05 | global batch size: 512 | lm loss: 1.229302E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.957 | TFLOPs: 43.52 | 31: iteration 34880/ 52452 | consumed samples: 17858560 | consumed tokens: 36574330880 | elapsed time per iteration (s): 1.83 | learning rate: 6.625E-05 | global batch size: 512 | lm loss: 1.217432E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.873 | TFLOPs: 42.01 | 31: iteration 34890/ 52452 | consumed samples: 17863680 | consumed tokens: 36584816640 | elapsed time per iteration (s): 1.76 | learning rate: 6.620E-05 | global batch size: 512 | lm loss: 1.230663E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.549 | TFLOPs: 43.76 | 31: iteration 34900/ 52452 | consumed samples: 17868800 | consumed tokens: 36595302400 | elapsed time per iteration (s): 1.78 | learning rate: 6.615E-05 | global batch size: 512 | lm loss: 1.229735E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.899 | TFLOPs: 43.06 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 34900 | lm loss value: 4.116540E+00 | lm loss PPL: 6.134661E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 34900 | lm loss value: 4.031250E+00 | lm loss PPL: 5.633125E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 34910/ 52452 | consumed samples: 17873920 | consumed tokens: 36605788160 | elapsed time per iteration (s): 9.17 | learning rate: 6.610E-05 | global batch size: 512 | lm loss: 1.250153E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.808 | TFLOPs: 8.38 | 31: iteration 34920/ 52452 | consumed samples: 17879040 | consumed tokens: 36616273920 | elapsed time per iteration (s): 1.76 | learning rate: 6.606E-05 | global batch size: 512 | lm loss: 1.240467E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.436 | TFLOPs: 43.59 | 31: iteration 34930/ 52452 | consumed samples: 17884160 | consumed tokens: 36626759680 | elapsed time per iteration (s): 1.75 | learning rate: 6.601E-05 | global batch size: 512 | lm loss: 1.235580E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.585 | TFLOPs: 43.92 | 31: iteration 34940/ 52452 | consumed samples: 17889280 | consumed tokens: 36637245440 | elapsed time per iteration (s): 1.80 | learning rate: 6.596E-05 | global batch size: 512 | lm loss: 1.243759E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.197 | TFLOPs: 42.66 | 31: iteration 34950/ 52452 | consumed samples: 17894400 | consumed tokens: 36647731200 | elapsed time per iteration (s): 1.80 | learning rate: 6.591E-05 | global batch size: 512 | lm loss: 1.238244E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.178 | TFLOPs: 42.65 | 31: iteration 34960/ 52452 | consumed samples: 17899520 | consumed tokens: 36658216960 | elapsed time per iteration (s): 1.84 | learning rate: 6.587E-05 | global batch size: 512 | lm loss: 1.238174E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.076 | TFLOPs: 41.74 | 31: iteration 34970/ 52452 | consumed samples: 17904640 | consumed tokens: 36668702720 | elapsed time per iteration (s): 1.78 | learning rate: 6.582E-05 | global batch size: 512 | lm loss: 1.220780E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.387 | TFLOPs: 43.29 | 31: iteration 34980/ 52452 | consumed samples: 17909760 | consumed tokens: 36679188480 | elapsed time per iteration (s): 1.86 | learning rate: 6.577E-05 | global batch size: 512 | lm loss: 1.233853E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.549 | TFLOPs: 41.21 | 31: iteration 34990/ 52452 | consumed samples: 17914880 | consumed tokens: 36689674240 | elapsed time per iteration (s): 1.85 | learning rate: 6.572E-05 | global batch size: 512 | lm loss: 1.217141E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.472 | TFLOPs: 41.65 | 31: iteration 35000/ 52452 | consumed samples: 17920000 | consumed tokens: 36700160000 | elapsed time per iteration (s): 1.79 | learning rate: 6.568E-05 | global batch size: 512 | lm loss: 1.250419E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.443 | TFLOPs: 42.99 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35000 | lm loss value: 4.131312E+00 | lm loss PPL: 6.225958E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35000 | lm loss value: 4.012644E+00 | lm loss PPL: 5.529286E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35010/ 52452 | consumed samples: 17925120 | consumed tokens: 36710645760 | elapsed time per iteration (s): 9.18 | learning rate: 6.563E-05 | global batch size: 512 | lm loss: 1.229274E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.755 | TFLOPs: 8.37 | 31: iteration 35020/ 52452 | consumed samples: 17930240 | consumed tokens: 36721131520 | elapsed time per iteration (s): 1.81 | learning rate: 6.558E-05 | global batch size: 512 | lm loss: 1.229621E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.103 | TFLOPs: 42.49 | 31: iteration 35030/ 52452 | consumed samples: 17935360 | consumed tokens: 36731617280 | elapsed time per iteration (s): 1.79 | learning rate: 6.553E-05 | global batch size: 512 | lm loss: 1.231529E+00 | grad norm: 0.193 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.707 | TFLOPs: 42.88 | 31: iteration 35040/ 52452 | consumed samples: 17940480 | consumed tokens: 36742103040 | elapsed time per iteration (s): 1.77 | learning rate: 6.549E-05 | global batch size: 512 | lm loss: 1.229525E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.045 | TFLOPs: 43.38 | 31: iteration 35050/ 52452 | consumed samples: 17945600 | consumed tokens: 36752588800 | elapsed time per iteration (s): 1.79 | learning rate: 6.544E-05 | global batch size: 512 | lm loss: 1.230693E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.600 | TFLOPs: 42.87 | 31: iteration 35060/ 52452 | consumed samples: 17950720 | consumed tokens: 36763074560 | elapsed time per iteration (s): 1.83 | learning rate: 6.539E-05 | global batch size: 512 | lm loss: 1.234673E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.681 | TFLOPs: 41.98 | 31: iteration 35070/ 52452 | consumed samples: 17955840 | consumed tokens: 36773560320 | elapsed time per iteration (s): 1.78 | learning rate: 6.534E-05 | global batch size: 512 | lm loss: 1.225178E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.243 | TFLOPs: 43.26 | 31: iteration 35080/ 52452 | consumed samples: 17960960 | consumed tokens: 36784046080 | elapsed time per iteration (s): 1.80 | learning rate: 6.530E-05 | global batch size: 512 | lm loss: 1.228017E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.056 | TFLOPs: 42.79 | 31: iteration 35090/ 52452 | consumed samples: 17966080 | consumed tokens: 36794531840 | elapsed time per iteration (s): 1.75 | learning rate: 6.525E-05 | global batch size: 512 | lm loss: 1.225050E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.174 | TFLOPs: 44.00 | 31: iteration 35100/ 52452 | consumed samples: 17971200 | consumed tokens: 36805017600 | elapsed time per iteration (s): 1.82 | learning rate: 6.520E-05 | global batch size: 512 | lm loss: 1.237003E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.706 | TFLOPs: 42.13 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35100 | lm loss value: 4.318161E+00 | lm loss PPL: 7.505045E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35100 | lm loss value: 4.170444E+00 | lm loss PPL: 6.474422E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35110/ 52452 | consumed samples: 17976320 | consumed tokens: 36815503360 | elapsed time per iteration (s): 9.20 | learning rate: 6.516E-05 | global batch size: 512 | lm loss: 1.234822E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.635 | TFLOPs: 8.35 | 31: iteration 35120/ 52452 | consumed samples: 17981440 | consumed tokens: 36825989120 | elapsed time per iteration (s): 1.78 | learning rate: 6.511E-05 | global batch size: 512 | lm loss: 1.229063E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.439 | TFLOPs: 43.14 | 31: iteration 35130/ 52452 | consumed samples: 17986560 | consumed tokens: 36836474880 | elapsed time per iteration (s): 1.93 | learning rate: 6.506E-05 | global batch size: 512 | lm loss: 1.231329E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.961 | TFLOPs: 39.92 | 31: iteration 35140/ 52452 | consumed samples: 17991680 | consumed tokens: 36846960640 | elapsed time per iteration (s): 1.81 | learning rate: 6.501E-05 | global batch size: 512 | lm loss: 1.230633E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.258 | TFLOPs: 42.52 | 31: iteration 35150/ 52452 | consumed samples: 17996800 | consumed tokens: 36857446400 | elapsed time per iteration (s): 1.86 | learning rate: 6.497E-05 | global batch size: 512 | lm loss: 1.229881E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.910 | TFLOPs: 41.26 | 31: iteration 35160/ 52452 | consumed samples: 18001920 | consumed tokens: 36867932160 | elapsed time per iteration (s): 1.77 | learning rate: 6.492E-05 | global batch size: 512 | lm loss: 1.232392E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.537 | TFLOPs: 43.31 | 31: iteration 35170/ 52452 | consumed samples: 18007040 | consumed tokens: 36878417920 | elapsed time per iteration (s): 1.79 | learning rate: 6.487E-05 | global batch size: 512 | lm loss: 1.233078E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.145 | TFLOPs: 42.95 | 31: iteration 35180/ 52452 | consumed samples: 18012160 | consumed tokens: 36888903680 | elapsed time per iteration (s): 1.83 | learning rate: 6.483E-05 | global batch size: 512 | lm loss: 1.233060E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.483 | TFLOPs: 42.10 | 31: iteration 35190/ 52452 | consumed samples: 18017280 | consumed tokens: 36899389440 | elapsed time per iteration (s): 1.81 | learning rate: 6.478E-05 | global batch size: 512 | lm loss: 1.224869E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.216 | TFLOPs: 42.36 | 31: iteration 35200/ 52452 | consumed samples: 18022400 | consumed tokens: 36909875200 | elapsed time per iteration (s): 1.85 | learning rate: 6.473E-05 | global batch size: 512 | lm loss: 1.239842E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.544 | TFLOPs: 41.51 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35200 | lm loss value: 4.050616E+00 | lm loss PPL: 5.743284E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35200 | lm loss value: 3.976968E+00 | lm loss PPL: 5.335503E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35210/ 52452 | consumed samples: 18027520 | consumed tokens: 36920360960 | elapsed time per iteration (s): 9.28 | learning rate: 6.468E-05 | global batch size: 512 | lm loss: 1.230695E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.150 | TFLOPs: 8.28 | 31: iteration 35220/ 52452 | consumed samples: 18032640 | consumed tokens: 36930846720 | elapsed time per iteration (s): 1.85 | learning rate: 6.464E-05 | global batch size: 512 | lm loss: 1.231060E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.878 | TFLOPs: 41.56 | 31: iteration 35230/ 52452 | consumed samples: 18037760 | consumed tokens: 36941332480 | elapsed time per iteration (s): 1.77 | learning rate: 6.459E-05 | global batch size: 512 | lm loss: 1.222861E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.040 | TFLOPs: 43.38 | 31: iteration 35240/ 52452 | consumed samples: 18042880 | consumed tokens: 36951818240 | elapsed time per iteration (s): 1.78 | learning rate: 6.454E-05 | global batch size: 512 | lm loss: 1.233556E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.542 | TFLOPs: 43.16 | 31: iteration 35250/ 52452 | consumed samples: 18048000 | consumed tokens: 36962304000 | elapsed time per iteration (s): 1.87 | learning rate: 6.450E-05 | global batch size: 512 | lm loss: 1.233284E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.381 | TFLOPs: 41.18 | 31: iteration 35260/ 52452 | consumed samples: 18053120 | consumed tokens: 36972789760 | elapsed time per iteration (s): 1.78 | learning rate: 6.445E-05 | global batch size: 512 | lm loss: 1.237272E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.241 | TFLOPs: 43.11 | 31: iteration 35270/ 52452 | consumed samples: 18058240 | consumed tokens: 36983275520 | elapsed time per iteration (s): 1.82 | learning rate: 6.440E-05 | global batch size: 512 | lm loss: 1.236618E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.942 | TFLOPs: 42.17 | 31: iteration 35280/ 52452 | consumed samples: 18063360 | consumed tokens: 36993761280 | elapsed time per iteration (s): 1.78 | learning rate: 6.436E-05 | global batch size: 512 | lm loss: 1.230360E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.586 | TFLOPs: 43.17 | 31: iteration 35290/ 52452 | consumed samples: 18068480 | consumed tokens: 37004247040 | elapsed time per iteration (s): 1.77 | learning rate: 6.431E-05 | global batch size: 512 | lm loss: 1.234312E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.855 | TFLOPs: 43.51 | 31: iteration 35300/ 52452 | consumed samples: 18073600 | consumed tokens: 37014732800 | elapsed time per iteration (s): 1.80 | learning rate: 6.426E-05 | global batch size: 512 | lm loss: 1.211993E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.415 | TFLOPs: 42.69 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35300 | lm loss value: 4.112022E+00 | lm loss PPL: 6.107010E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35300 | lm loss value: 4.016132E+00 | lm loss PPL: 5.548606E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35310/ 52452 | consumed samples: 18078720 | consumed tokens: 37025218560 | elapsed time per iteration (s): 9.26 | learning rate: 6.421E-05 | global batch size: 512 | lm loss: 1.240540E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.280 | TFLOPs: 8.30 | 31: iteration 35320/ 52452 | consumed samples: 18083840 | consumed tokens: 37035704320 | elapsed time per iteration (s): 1.78 | learning rate: 6.417E-05 | global batch size: 512 | lm loss: 1.227370E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.199 | TFLOPs: 43.26 | 31: iteration 35330/ 52452 | consumed samples: 18088960 | consumed tokens: 37046190080 | elapsed time per iteration (s): 1.76 | learning rate: 6.412E-05 | global batch size: 512 | lm loss: 1.220236E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.082 | TFLOPs: 43.69 | 31: iteration 35340/ 52452 | consumed samples: 18094080 | consumed tokens: 37056675840 | elapsed time per iteration (s): 1.77 | learning rate: 6.407E-05 | global batch size: 512 | lm loss: 1.235585E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.531 | TFLOPs: 43.31 | 31: iteration 35350/ 52452 | consumed samples: 18099200 | consumed tokens: 37067161600 | elapsed time per iteration (s): 1.79 | learning rate: 6.403E-05 | global batch size: 512 | lm loss: 1.218687E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.146 | TFLOPs: 42.95 | 31: iteration 35360/ 52452 | consumed samples: 18104320 | consumed tokens: 37077647360 | elapsed time per iteration (s): 1.79 | learning rate: 6.398E-05 | global batch size: 512 | lm loss: 1.235299E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.758 | TFLOPs: 43.04 | 31: iteration 35370/ 52452 | consumed samples: 18109440 | consumed tokens: 37088133120 | elapsed time per iteration (s): 1.87 | learning rate: 6.393E-05 | global batch size: 512 | lm loss: 1.232219E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.174 | TFLOPs: 41.00 | 31: iteration 35380/ 52452 | consumed samples: 18114560 | consumed tokens: 37098618880 | elapsed time per iteration (s): 1.77 | learning rate: 6.389E-05 | global batch size: 512 | lm loss: 1.241424E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.628 | TFLOPs: 43.47 | 31: iteration 35390/ 52452 | consumed samples: 18119680 | consumed tokens: 37109104640 | elapsed time per iteration (s): 1.78 | learning rate: 6.384E-05 | global batch size: 512 | lm loss: 1.211903E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.037 | TFLOPs: 43.08 | 31: iteration 35400/ 52452 | consumed samples: 18124800 | consumed tokens: 37119590400 | elapsed time per iteration (s): 1.80 | learning rate: 6.379E-05 | global batch size: 512 | lm loss: 1.230683E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.300 | TFLOPs: 42.67 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35400 | lm loss value: 4.125548E+00 | lm loss PPL: 6.190172E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35400 | lm loss value: 4.026765E+00 | lm loss PPL: 5.607922E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35410/ 52452 | consumed samples: 18129920 | consumed tokens: 37130076160 | elapsed time per iteration (s): 9.19 | learning rate: 6.375E-05 | global batch size: 512 | lm loss: 1.222462E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.693 | TFLOPs: 8.36 | 31: iteration 35420/ 52452 | consumed samples: 18135040 | consumed tokens: 37140561920 | elapsed time per iteration (s): 1.75 | learning rate: 6.370E-05 | global batch size: 512 | lm loss: 1.235634E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.482 | TFLOPs: 43.90 | 31: iteration 35430/ 52452 | consumed samples: 18140160 | consumed tokens: 37151047680 | elapsed time per iteration (s): 1.76 | learning rate: 6.365E-05 | global batch size: 512 | lm loss: 1.219128E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.249 | TFLOPs: 43.71 | 31: iteration 35440/ 52452 | consumed samples: 18145280 | consumed tokens: 37161533440 | elapsed time per iteration (s): 1.75 | learning rate: 6.361E-05 | global batch size: 512 | lm loss: 1.240397E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.953 | TFLOPs: 43.82 | 31: iteration 35450/ 52452 | consumed samples: 18150400 | consumed tokens: 37172019200 | elapsed time per iteration (s): 1.78 | learning rate: 6.356E-05 | global batch size: 512 | lm loss: 1.235629E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.864 | TFLOPs: 43.06 | 31: iteration 35460/ 52452 | consumed samples: 18155520 | consumed tokens: 37182504960 | elapsed time per iteration (s): 1.76 | learning rate: 6.351E-05 | global batch size: 512 | lm loss: 1.223377E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.100 | TFLOPs: 43.69 | 31: iteration 35470/ 52452 | consumed samples: 18160640 | consumed tokens: 37192990720 | elapsed time per iteration (s): 1.78 | learning rate: 6.347E-05 | global batch size: 512 | lm loss: 1.221912E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.869 | TFLOPs: 43.06 | 31: iteration 35480/ 52452 | consumed samples: 18165760 | consumed tokens: 37203476480 | elapsed time per iteration (s): 1.77 | learning rate: 6.342E-05 | global batch size: 512 | lm loss: 1.231956E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.750 | TFLOPs: 43.49 | 31: iteration 35490/ 52452 | consumed samples: 18170880 | consumed tokens: 37213962240 | elapsed time per iteration (s): 1.74 | learning rate: 6.337E-05 | global batch size: 512 | lm loss: 1.223123E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.908 | TFLOPs: 44.11 | 31: iteration 35500/ 52452 | consumed samples: 18176000 | consumed tokens: 37224448000 | elapsed time per iteration (s): 1.78 | learning rate: 6.333E-05 | global batch size: 512 | lm loss: 1.229928E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.410 | TFLOPs: 43.29 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35500 | lm loss value: 4.172642E+00 | lm loss PPL: 6.488667E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35500 | lm loss value: 4.060638E+00 | lm loss PPL: 5.801134E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35510/ 52452 | consumed samples: 18181120 | consumed tokens: 37234933760 | elapsed time per iteration (s): 9.13 | learning rate: 6.328E-05 | global batch size: 512 | lm loss: 1.218754E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.066 | TFLOPs: 8.42 | 31: iteration 35520/ 52452 | consumed samples: 18186240 | consumed tokens: 37245419520 | elapsed time per iteration (s): 1.79 | learning rate: 6.323E-05 | global batch size: 512 | lm loss: 1.246314E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.140 | TFLOPs: 42.95 | 31: iteration 35530/ 52452 | consumed samples: 18191360 | consumed tokens: 37255905280 | elapsed time per iteration (s): 1.78 | learning rate: 6.319E-05 | global batch size: 512 | lm loss: 1.238003E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.518 | TFLOPs: 43.15 | 31: iteration 35540/ 52452 | consumed samples: 18196480 | consumed tokens: 37266391040 | elapsed time per iteration (s): 1.80 | learning rate: 6.314E-05 | global batch size: 512 | lm loss: 1.219938E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.984 | TFLOPs: 42.62 | 31: iteration 35550/ 52452 | consumed samples: 18201600 | consumed tokens: 37276876800 | elapsed time per iteration (s): 1.76 | learning rate: 6.309E-05 | global batch size: 512 | lm loss: 1.234193E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.489 | TFLOPs: 43.60 | 31: iteration 35560/ 52452 | consumed samples: 18206720 | consumed tokens: 37287362560 | elapsed time per iteration (s): 1.79 | learning rate: 6.305E-05 | global batch size: 512 | lm loss: 1.231999E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.858 | TFLOPs: 42.91 | 31: iteration 35570/ 52452 | consumed samples: 18211840 | consumed tokens: 37297848320 | elapsed time per iteration (s): 1.82 | learning rate: 6.300E-05 | global batch size: 512 | lm loss: 1.229612E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.794 | TFLOPs: 42.30 | 31: iteration 35580/ 52452 | consumed samples: 18216960 | consumed tokens: 37308334080 | elapsed time per iteration (s): 1.83 | learning rate: 6.295E-05 | global batch size: 512 | lm loss: 1.221515E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.312 | TFLOPs: 42.07 | 31: iteration 35590/ 52452 | consumed samples: 18222080 | consumed tokens: 37318819840 | elapsed time per iteration (s): 1.73 | learning rate: 6.291E-05 | global batch size: 512 | lm loss: 1.223235E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.798 | TFLOPs: 44.40 | 31: iteration 35600/ 52452 | consumed samples: 18227200 | consumed tokens: 37329305600 | elapsed time per iteration (s): 1.77 | learning rate: 6.286E-05 | global batch size: 512 | lm loss: 1.217836E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.886 | TFLOPs: 43.36 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35600 | lm loss value: 4.150978E+00 | lm loss PPL: 6.349604E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35600 | lm loss value: 4.046114E+00 | lm loss PPL: 5.717484E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35610/ 52452 | consumed samples: 18232320 | consumed tokens: 37339791360 | elapsed time per iteration (s): 9.26 | learning rate: 6.282E-05 | global batch size: 512 | lm loss: 1.217496E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.318 | TFLOPs: 8.30 | 31: iteration 35620/ 52452 | consumed samples: 18237440 | consumed tokens: 37350277120 | elapsed time per iteration (s): 1.75 | learning rate: 6.277E-05 | global batch size: 512 | lm loss: 1.221896E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.270 | TFLOPs: 44.02 | 31: iteration 35630/ 52452 | consumed samples: 18242560 | consumed tokens: 37360762880 | elapsed time per iteration (s): 1.73 | learning rate: 6.272E-05 | global batch size: 512 | lm loss: 1.233546E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.000 | TFLOPs: 44.43 | 31: iteration 35640/ 52452 | consumed samples: 18247680 | consumed tokens: 37371248640 | elapsed time per iteration (s): 1.80 | learning rate: 6.268E-05 | global batch size: 512 | lm loss: 1.236186E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.874 | TFLOPs: 42.76 | 31: iteration 35650/ 52452 | consumed samples: 18252800 | consumed tokens: 37381734400 | elapsed time per iteration (s): 1.74 | learning rate: 6.263E-05 | global batch size: 512 | lm loss: 1.223139E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.826 | TFLOPs: 44.10 | 31: iteration 35660/ 52452 | consumed samples: 18257920 | consumed tokens: 37392220160 | elapsed time per iteration (s): 1.99 | learning rate: 6.258E-05 | global batch size: 512 | lm loss: 1.224713E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 257.288 | TFLOPs: 38.62 | 31: iteration 35670/ 52452 | consumed samples: 18263040 | consumed tokens: 37402705920 | elapsed time per iteration (s): 1.76 | learning rate: 6.254E-05 | global batch size: 512 | lm loss: 1.227343E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.558 | TFLOPs: 43.76 | 31: iteration 35680/ 52452 | consumed samples: 18268160 | consumed tokens: 37413191680 | elapsed time per iteration (s): 1.88 | learning rate: 6.249E-05 | global batch size: 512 | lm loss: 1.229334E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.754 | TFLOPs: 40.94 | 31: iteration 35690/ 52452 | consumed samples: 18273280 | consumed tokens: 37423677440 | elapsed time per iteration (s): 1.82 | learning rate: 6.245E-05 | global batch size: 512 | lm loss: 1.222097E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.096 | TFLOPs: 42.19 | 31: iteration 35700/ 52452 | consumed samples: 18278400 | consumed tokens: 37434163200 | elapsed time per iteration (s): 1.76 | learning rate: 6.240E-05 | global batch size: 512 | lm loss: 1.220123E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.603 | TFLOPs: 43.62 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35700 | lm loss value: 4.144220E+00 | lm loss PPL: 6.306843E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35700 | lm loss value: 4.030753E+00 | lm loss PPL: 5.630327E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35710/ 52452 | consumed samples: 18283520 | consumed tokens: 37444648960 | elapsed time per iteration (s): 9.19 | learning rate: 6.235E-05 | global batch size: 512 | lm loss: 1.230861E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.742 | TFLOPs: 8.37 | 31: iteration 35720/ 52452 | consumed samples: 18288640 | consumed tokens: 37455134720 | elapsed time per iteration (s): 1.76 | learning rate: 6.231E-05 | global batch size: 512 | lm loss: 1.217554E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.627 | TFLOPs: 43.62 | 31: iteration 35730/ 52452 | consumed samples: 18293760 | consumed tokens: 37465620480 | elapsed time per iteration (s): 1.85 | learning rate: 6.226E-05 | global batch size: 512 | lm loss: 1.220586E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.535 | TFLOPs: 41.51 | 31: iteration 35740/ 52452 | consumed samples: 18298880 | consumed tokens: 37476106240 | elapsed time per iteration (s): 1.78 | learning rate: 6.221E-05 | global batch size: 512 | lm loss: 1.230441E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.008 | TFLOPs: 43.23 | 31: iteration 35750/ 52452 | consumed samples: 18304000 | consumed tokens: 37486592000 | elapsed time per iteration (s): 1.74 | learning rate: 6.217E-05 | global batch size: 512 | lm loss: 1.236286E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.951 | TFLOPs: 44.12 | 31: iteration 35760/ 52452 | consumed samples: 18309120 | consumed tokens: 37497077760 | elapsed time per iteration (s): 1.75 | learning rate: 6.212E-05 | global batch size: 512 | lm loss: 1.235082E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.220 | TFLOPs: 43.86 | 31: iteration 35770/ 52452 | consumed samples: 18314240 | consumed tokens: 37507563520 | elapsed time per iteration (s): 1.76 | learning rate: 6.208E-05 | global batch size: 512 | lm loss: 1.236372E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.710 | TFLOPs: 43.78 | 31: iteration 35780/ 52452 | consumed samples: 18319360 | consumed tokens: 37518049280 | elapsed time per iteration (s): 1.73 | learning rate: 6.203E-05 | global batch size: 512 | lm loss: 1.235809E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.133 | TFLOPs: 44.45 | 31: iteration 35790/ 52452 | consumed samples: 18324480 | consumed tokens: 37528535040 | elapsed time per iteration (s): 1.79 | learning rate: 6.198E-05 | global batch size: 512 | lm loss: 1.215136E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.540 | TFLOPs: 43.01 | 31: iteration 35800/ 52452 | consumed samples: 18329600 | consumed tokens: 37539020800 | elapsed time per iteration (s): 1.77 | learning rate: 6.194E-05 | global batch size: 512 | lm loss: 1.217999E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.057 | TFLOPs: 43.54 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35800 | lm loss value: 4.181689E+00 | lm loss PPL: 6.547637E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35800 | lm loss value: 4.061635E+00 | lm loss PPL: 5.806915E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35810/ 52452 | consumed samples: 18334720 | consumed tokens: 37549506560 | elapsed time per iteration (s): 9.17 | learning rate: 6.189E-05 | global batch size: 512 | lm loss: 1.224284E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.809 | TFLOPs: 8.38 | 31: iteration 35820/ 52452 | consumed samples: 18339840 | consumed tokens: 37559992320 | elapsed time per iteration (s): 1.86 | learning rate: 6.185E-05 | global batch size: 512 | lm loss: 1.244607E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.734 | TFLOPs: 41.39 | 31: iteration 35830/ 52452 | consumed samples: 18344960 | consumed tokens: 37570478080 | elapsed time per iteration (s): 1.79 | learning rate: 6.180E-05 | global batch size: 512 | lm loss: 1.216473E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.727 | TFLOPs: 43.04 | 31: iteration 35840/ 52452 | consumed samples: 18350080 | consumed tokens: 37580963840 | elapsed time per iteration (s): 1.76 | learning rate: 6.175E-05 | global batch size: 512 | lm loss: 1.226258E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.569 | TFLOPs: 43.61 | 31: iteration 35850/ 52452 | consumed samples: 18355200 | consumed tokens: 37591449600 | elapsed time per iteration (s): 1.78 | learning rate: 6.171E-05 | global batch size: 512 | lm loss: 1.233773E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.358 | TFLOPs: 43.13 | 31: iteration 35860/ 52452 | consumed samples: 18360320 | consumed tokens: 37601935360 | elapsed time per iteration (s): 1.79 | learning rate: 6.166E-05 | global batch size: 512 | lm loss: 1.232764E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.138 | TFLOPs: 42.95 | 31: iteration 35870/ 52452 | consumed samples: 18365440 | consumed tokens: 37612421120 | elapsed time per iteration (s): 1.80 | learning rate: 6.162E-05 | global batch size: 512 | lm loss: 1.228367E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.970 | TFLOPs: 42.62 | 31: iteration 35880/ 52452 | consumed samples: 18370560 | consumed tokens: 37622906880 | elapsed time per iteration (s): 1.78 | learning rate: 6.157E-05 | global batch size: 512 | lm loss: 1.220955E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.381 | TFLOPs: 43.28 | 31: iteration 35890/ 52452 | consumed samples: 18375680 | consumed tokens: 37633392640 | elapsed time per iteration (s): 1.79 | learning rate: 6.152E-05 | global batch size: 512 | lm loss: 1.227563E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.716 | TFLOPs: 42.88 | 31: iteration 35900/ 52452 | consumed samples: 18380800 | consumed tokens: 37643878400 | elapsed time per iteration (s): 1.77 | learning rate: 6.148E-05 | global batch size: 512 | lm loss: 1.224942E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.802 | TFLOPs: 43.35 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 35900 | lm loss value: 4.144133E+00 | lm loss PPL: 6.306290E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 35900 | lm loss value: 4.019869E+00 | lm loss PPL: 5.569383E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 35910/ 52452 | consumed samples: 18385920 | consumed tokens: 37654364160 | elapsed time per iteration (s): 9.15 | learning rate: 6.143E-05 | global batch size: 512 | lm loss: 1.233158E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.948 | TFLOPs: 8.40 | 31: iteration 35920/ 52452 | consumed samples: 18391040 | consumed tokens: 37664849920 | elapsed time per iteration (s): 1.73 | learning rate: 6.139E-05 | global batch size: 512 | lm loss: 1.225917E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.311 | TFLOPs: 44.32 | 31: iteration 35930/ 52452 | consumed samples: 18396160 | consumed tokens: 37675335680 | elapsed time per iteration (s): 1.76 | learning rate: 6.134E-05 | global batch size: 512 | lm loss: 1.218250E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.145 | TFLOPs: 43.55 | 31: iteration 35940/ 52452 | consumed samples: 18401280 | consumed tokens: 37685821440 | elapsed time per iteration (s): 1.85 | learning rate: 6.130E-05 | global batch size: 512 | lm loss: 1.223154E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.587 | TFLOPs: 41.51 | 31: iteration 35950/ 52452 | consumed samples: 18406400 | consumed tokens: 37696307200 | elapsed time per iteration (s): 1.75 | learning rate: 6.125E-05 | global batch size: 512 | lm loss: 1.222344E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.343 | TFLOPs: 43.88 | 31: iteration 35960/ 52452 | consumed samples: 18411520 | consumed tokens: 37706792960 | elapsed time per iteration (s): 1.77 | learning rate: 6.120E-05 | global batch size: 512 | lm loss: 1.231491E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.168 | TFLOPs: 43.40 | 31: iteration 35970/ 52452 | consumed samples: 18416640 | consumed tokens: 37717278720 | elapsed time per iteration (s): 1.74 | learning rate: 6.116E-05 | global batch size: 512 | lm loss: 1.221606E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.667 | TFLOPs: 44.23 | 31: iteration 35980/ 52452 | consumed samples: 18421760 | consumed tokens: 37727764480 | elapsed time per iteration (s): 1.78 | learning rate: 6.111E-05 | global batch size: 512 | lm loss: 1.214577E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.113 | TFLOPs: 43.09 | 31: iteration 35990/ 52452 | consumed samples: 18426880 | consumed tokens: 37738250240 | elapsed time per iteration (s): 1.75 | learning rate: 6.107E-05 | global batch size: 512 | lm loss: 1.225807E+00 | grad norm: 0.146 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.696 | TFLOPs: 43.93 | 0: [2022-12-26 10:20:16,820] [INFO] [logging.py:68:log_dist] [Rank 0] step=36000, skipped=0, lr=[6.102066254221278e-05, 6.102066254221278e-05, 6.102066254221278e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 36000/ 52452 | consumed samples: 18432000 | consumed tokens: 37748736000 | elapsed time per iteration (s): 1.74 | learning rate: 6.102E-05 | global batch size: 512 | lm loss: 1.202713E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.946 | TFLOPs: 44.27 | 0: steps: 36000 loss: 1.1988 iter time (s): 1.793 samples/sec: 285.588 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36000 | lm loss value: 4.328148E+00 | lm loss PPL: 7.580376E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36000 | lm loss value: 4.157289E+00 | lm loss PPL: 6.389806E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36010/ 52452 | consumed samples: 18437120 | consumed tokens: 37759221760 | elapsed time per iteration (s): 9.12 | learning rate: 6.097E-05 | global batch size: 512 | lm loss: 1.218647E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.144 | TFLOPs: 8.43 | 31: iteration 36020/ 52452 | consumed samples: 18442240 | consumed tokens: 37769707520 | elapsed time per iteration (s): 1.74 | learning rate: 6.093E-05 | global batch size: 512 | lm loss: 1.223565E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.508 | TFLOPs: 44.20 | 31: iteration 36030/ 52452 | consumed samples: 18447360 | consumed tokens: 37780193280 | elapsed time per iteration (s): 1.82 | learning rate: 6.088E-05 | global batch size: 512 | lm loss: 1.226078E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.966 | TFLOPs: 42.32 | 31: iteration 36040/ 52452 | consumed samples: 18452480 | consumed tokens: 37790679040 | elapsed time per iteration (s): 1.79 | learning rate: 6.084E-05 | global batch size: 512 | lm loss: 1.231117E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.059 | TFLOPs: 42.94 | 31: iteration 36050/ 52452 | consumed samples: 18457600 | consumed tokens: 37801164800 | elapsed time per iteration (s): 1.76 | learning rate: 6.079E-05 | global batch size: 512 | lm loss: 1.220517E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.834 | TFLOPs: 43.65 | 31: iteration 36060/ 52452 | consumed samples: 18462720 | consumed tokens: 37811650560 | elapsed time per iteration (s): 1.86 | learning rate: 6.075E-05 | global batch size: 512 | lm loss: 1.236194E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.953 | TFLOPs: 41.27 | 31: iteration 36070/ 52452 | consumed samples: 18467840 | consumed tokens: 37822136320 | elapsed time per iteration (s): 1.77 | learning rate: 6.070E-05 | global batch size: 512 | lm loss: 1.229003E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.688 | TFLOPs: 43.33 | 31: iteration 36080/ 52452 | consumed samples: 18472960 | consumed tokens: 37832622080 | elapsed time per iteration (s): 1.82 | learning rate: 6.066E-05 | global batch size: 512 | lm loss: 1.226607E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.791 | TFLOPs: 42.30 | 31: iteration 36090/ 52452 | consumed samples: 18478080 | consumed tokens: 37843107840 | elapsed time per iteration (s): 1.75 | learning rate: 6.061E-05 | global batch size: 512 | lm loss: 1.216225E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.587 | TFLOPs: 43.92 | 31: iteration 36100/ 52452 | consumed samples: 18483200 | consumed tokens: 37853593600 | elapsed time per iteration (s): 1.75 | learning rate: 6.056E-05 | global batch size: 512 | lm loss: 1.228038E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.243 | TFLOPs: 43.86 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36100 | lm loss value: 4.294633E+00 | lm loss PPL: 7.330530E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36100 | lm loss value: 4.115003E+00 | lm loss PPL: 6.125238E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36110/ 52452 | consumed samples: 18488320 | consumed tokens: 37864079360 | elapsed time per iteration (s): 9.16 | learning rate: 6.052E-05 | global batch size: 512 | lm loss: 1.239099E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.924 | TFLOPs: 8.39 | 31: iteration 36120/ 52452 | consumed samples: 18493440 | consumed tokens: 37874565120 | elapsed time per iteration (s): 1.79 | learning rate: 6.047E-05 | global batch size: 512 | lm loss: 1.214294E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.581 | TFLOPs: 42.86 | 31: iteration 36130/ 52452 | consumed samples: 18498560 | consumed tokens: 37885050880 | elapsed time per iteration (s): 1.78 | learning rate: 6.043E-05 | global batch size: 512 | lm loss: 1.224764E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.572 | TFLOPs: 43.16 | 31: iteration 36140/ 52452 | consumed samples: 18503680 | consumed tokens: 37895536640 | elapsed time per iteration (s): 1.79 | learning rate: 6.038E-05 | global batch size: 512 | lm loss: 1.208276E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.158 | TFLOPs: 42.95 | 31: iteration 36150/ 52452 | consumed samples: 18508800 | consumed tokens: 37906022400 | elapsed time per iteration (s): 1.85 | learning rate: 6.034E-05 | global batch size: 512 | lm loss: 1.221345E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.694 | TFLOPs: 41.53 | 31: iteration 36160/ 52452 | consumed samples: 18513920 | consumed tokens: 37916508160 | elapsed time per iteration (s): 1.76 | learning rate: 6.029E-05 | global batch size: 512 | lm loss: 1.212489E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.464 | TFLOPs: 43.60 | 31: iteration 36170/ 52452 | consumed samples: 18519040 | consumed tokens: 37926993920 | elapsed time per iteration (s): 1.80 | learning rate: 6.025E-05 | global batch size: 512 | lm loss: 1.214705E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.280 | TFLOPs: 42.67 | 31: iteration 36180/ 52452 | consumed samples: 18524160 | consumed tokens: 37937479680 | elapsed time per iteration (s): 1.74 | learning rate: 6.020E-05 | global batch size: 512 | lm loss: 1.233371E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.267 | TFLOPs: 44.17 | 31: iteration 36190/ 52452 | consumed samples: 18529280 | consumed tokens: 37947965440 | elapsed time per iteration (s): 1.76 | learning rate: 6.016E-05 | global batch size: 512 | lm loss: 1.223255E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.260 | TFLOPs: 43.57 | 31: iteration 36200/ 52452 | consumed samples: 18534400 | consumed tokens: 37958451200 | elapsed time per iteration (s): 1.78 | learning rate: 6.011E-05 | global batch size: 512 | lm loss: 1.218676E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.419 | TFLOPs: 43.29 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36200 | lm loss value: 4.279440E+00 | lm loss PPL: 7.220003E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36200 | lm loss value: 4.124812E+00 | lm loss PPL: 6.185619E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36210/ 52452 | consumed samples: 18539520 | consumed tokens: 37968936960 | elapsed time per iteration (s): 9.25 | learning rate: 6.007E-05 | global batch size: 512 | lm loss: 1.237430E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.344 | TFLOPs: 8.31 | 31: iteration 36220/ 52452 | consumed samples: 18544640 | consumed tokens: 37979422720 | elapsed time per iteration (s): 1.76 | learning rate: 6.002E-05 | global batch size: 512 | lm loss: 1.208674E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.353 | TFLOPs: 43.58 | 31: iteration 36230/ 52452 | consumed samples: 18549760 | consumed tokens: 37989908480 | elapsed time per iteration (s): 1.74 | learning rate: 5.997E-05 | global batch size: 512 | lm loss: 1.221340E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.488 | TFLOPs: 44.05 | 31: iteration 36240/ 52452 | consumed samples: 18554880 | consumed tokens: 38000394240 | elapsed time per iteration (s): 1.78 | learning rate: 5.993E-05 | global batch size: 512 | lm loss: 1.223285E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.314 | TFLOPs: 43.27 | 31: iteration 36250/ 52452 | consumed samples: 18560000 | consumed tokens: 38010880000 | elapsed time per iteration (s): 1.81 | learning rate: 5.988E-05 | global batch size: 512 | lm loss: 1.225828E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.171 | TFLOPs: 42.50 | 31: iteration 36260/ 52452 | consumed samples: 18565120 | consumed tokens: 38021365760 | elapsed time per iteration (s): 1.72 | learning rate: 5.984E-05 | global batch size: 512 | lm loss: 1.214452E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.168 | TFLOPs: 44.60 | 31: iteration 36270/ 52452 | consumed samples: 18570240 | consumed tokens: 38031851520 | elapsed time per iteration (s): 1.79 | learning rate: 5.979E-05 | global batch size: 512 | lm loss: 1.234163E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.575 | TFLOPs: 43.01 | 31: iteration 36280/ 52452 | consumed samples: 18575360 | consumed tokens: 38042337280 | elapsed time per iteration (s): 1.82 | learning rate: 5.975E-05 | global batch size: 512 | lm loss: 1.221533E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.134 | TFLOPs: 42.20 | 31: iteration 36290/ 52452 | consumed samples: 18580480 | consumed tokens: 38052823040 | elapsed time per iteration (s): 1.73 | learning rate: 5.970E-05 | global batch size: 512 | lm loss: 1.222901E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.210 | TFLOPs: 44.31 | 31: iteration 36300/ 52452 | consumed samples: 18585600 | consumed tokens: 38063308800 | elapsed time per iteration (s): 1.80 | learning rate: 5.966E-05 | global batch size: 512 | lm loss: 1.223542E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.049 | TFLOPs: 42.63 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36300 | lm loss value: 4.147600E+00 | lm loss PPL: 6.328192E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36300 | lm loss value: 4.031649E+00 | lm loss PPL: 5.635374E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36310/ 52452 | consumed samples: 18590720 | consumed tokens: 38073794560 | elapsed time per iteration (s): 9.16 | learning rate: 5.961E-05 | global batch size: 512 | lm loss: 1.214032E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.914 | TFLOPs: 8.39 | 31: iteration 36320/ 52452 | consumed samples: 18595840 | consumed tokens: 38084280320 | elapsed time per iteration (s): 1.75 | learning rate: 5.957E-05 | global batch size: 512 | lm loss: 1.219042E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.369 | TFLOPs: 44.03 | 31: iteration 36330/ 52452 | consumed samples: 18600960 | consumed tokens: 38094766080 | elapsed time per iteration (s): 1.80 | learning rate: 5.952E-05 | global batch size: 512 | lm loss: 1.225764E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.664 | TFLOPs: 42.58 | 31: iteration 36340/ 52452 | consumed samples: 18606080 | consumed tokens: 38105251840 | elapsed time per iteration (s): 1.78 | learning rate: 5.948E-05 | global batch size: 512 | lm loss: 1.225787E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.742 | TFLOPs: 43.19 | 31: iteration 36350/ 52452 | consumed samples: 18611200 | consumed tokens: 38115737600 | elapsed time per iteration (s): 1.84 | learning rate: 5.943E-05 | global batch size: 512 | lm loss: 1.231378E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.496 | TFLOPs: 41.80 | 31: iteration 36360/ 52452 | consumed samples: 18616320 | consumed tokens: 38126223360 | elapsed time per iteration (s): 1.79 | learning rate: 5.939E-05 | global batch size: 512 | lm loss: 1.212157E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.673 | TFLOPs: 42.88 | 31: iteration 36370/ 52452 | consumed samples: 18621440 | consumed tokens: 38136709120 | elapsed time per iteration (s): 1.78 | learning rate: 5.934E-05 | global batch size: 512 | lm loss: 1.212806E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.139 | TFLOPs: 43.25 | 31: iteration 36380/ 52452 | consumed samples: 18626560 | consumed tokens: 38147194880 | elapsed time per iteration (s): 1.79 | learning rate: 5.930E-05 | global batch size: 512 | lm loss: 1.229162E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.368 | TFLOPs: 42.98 | 31: iteration 36390/ 52452 | consumed samples: 18631680 | consumed tokens: 38157680640 | elapsed time per iteration (s): 1.75 | learning rate: 5.925E-05 | global batch size: 512 | lm loss: 1.202919E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.019 | TFLOPs: 43.83 | 31: iteration 36400/ 52452 | consumed samples: 18636800 | consumed tokens: 38168166400 | elapsed time per iteration (s): 1.74 | learning rate: 5.921E-05 | global batch size: 512 | lm loss: 1.217261E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.532 | TFLOPs: 44.06 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36400 | lm loss value: 4.214484E+00 | lm loss PPL: 6.765923E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36400 | lm loss value: 4.067972E+00 | lm loss PPL: 5.843834E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36410/ 52452 | consumed samples: 18641920 | consumed tokens: 38178652160 | elapsed time per iteration (s): 9.19 | learning rate: 5.916E-05 | global batch size: 512 | lm loss: 1.212280E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.682 | TFLOPs: 8.36 | 31: iteration 36420/ 52452 | consumed samples: 18647040 | consumed tokens: 38189137920 | elapsed time per iteration (s): 1.78 | learning rate: 5.912E-05 | global batch size: 512 | lm loss: 1.213750E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.231 | TFLOPs: 43.26 | 31: iteration 36430/ 52452 | consumed samples: 18652160 | consumed tokens: 38199623680 | elapsed time per iteration (s): 1.73 | learning rate: 5.907E-05 | global batch size: 512 | lm loss: 1.222011E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.459 | TFLOPs: 44.35 | 31: iteration 36440/ 52452 | consumed samples: 18657280 | consumed tokens: 38210109440 | elapsed time per iteration (s): 1.75 | learning rate: 5.903E-05 | global batch size: 512 | lm loss: 1.212805E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.273 | TFLOPs: 43.87 | 31: iteration 36450/ 52452 | consumed samples: 18662400 | consumed tokens: 38220595200 | elapsed time per iteration (s): 1.74 | learning rate: 5.898E-05 | global batch size: 512 | lm loss: 1.222980E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.273 | TFLOPs: 44.17 | 31: iteration 36460/ 52452 | consumed samples: 18667520 | consumed tokens: 38231080960 | elapsed time per iteration (s): 1.76 | learning rate: 5.894E-05 | global batch size: 512 | lm loss: 1.225263E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.447 | TFLOPs: 43.74 | 31: iteration 36470/ 52452 | consumed samples: 18672640 | consumed tokens: 38241566720 | elapsed time per iteration (s): 1.73 | learning rate: 5.889E-05 | global batch size: 512 | lm loss: 1.205683E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.050 | TFLOPs: 44.44 | 31: iteration 36480/ 52452 | consumed samples: 18677760 | consumed tokens: 38252052480 | elapsed time per iteration (s): 1.72 | learning rate: 5.885E-05 | global batch size: 512 | lm loss: 1.216438E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.820 | TFLOPs: 44.55 | 31: iteration 36490/ 52452 | consumed samples: 18682880 | consumed tokens: 38262538240 | elapsed time per iteration (s): 1.73 | learning rate: 5.880E-05 | global batch size: 512 | lm loss: 1.221886E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.507 | TFLOPs: 44.35 | 31: iteration 36500/ 52452 | consumed samples: 18688000 | consumed tokens: 38273024000 | elapsed time per iteration (s): 1.74 | learning rate: 5.876E-05 | global batch size: 512 | lm loss: 1.221353E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.924 | TFLOPs: 44.27 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36500 | lm loss value: 4.226024E+00 | lm loss PPL: 6.844457E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36500 | lm loss value: 4.085815E+00 | lm loss PPL: 5.949040E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36510/ 52452 | consumed samples: 18693120 | consumed tokens: 38283509760 | elapsed time per iteration (s): 9.12 | learning rate: 5.871E-05 | global batch size: 512 | lm loss: 1.212785E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.166 | TFLOPs: 8.43 | 31: iteration 36520/ 52452 | consumed samples: 18698240 | consumed tokens: 38293995520 | elapsed time per iteration (s): 1.74 | learning rate: 5.867E-05 | global batch size: 512 | lm loss: 1.215289E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.987 | TFLOPs: 44.28 | 31: iteration 36530/ 52452 | consumed samples: 18703360 | consumed tokens: 38304481280 | elapsed time per iteration (s): 1.73 | learning rate: 5.863E-05 | global batch size: 512 | lm loss: 1.215315E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.290 | TFLOPs: 44.32 | 31: iteration 36540/ 52452 | consumed samples: 18708480 | consumed tokens: 38314967040 | elapsed time per iteration (s): 1.76 | learning rate: 5.858E-05 | global batch size: 512 | lm loss: 1.218061E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.684 | TFLOPs: 43.63 | 31: iteration 36550/ 52452 | consumed samples: 18713600 | consumed tokens: 38325452800 | elapsed time per iteration (s): 1.74 | learning rate: 5.854E-05 | global batch size: 512 | lm loss: 1.215215E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.526 | TFLOPs: 44.21 | 31: iteration 36560/ 52452 | consumed samples: 18718720 | consumed tokens: 38335938560 | elapsed time per iteration (s): 1.74 | learning rate: 5.849E-05 | global batch size: 512 | lm loss: 1.211917E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.694 | TFLOPs: 44.23 | 31: iteration 36570/ 52452 | consumed samples: 18723840 | consumed tokens: 38346424320 | elapsed time per iteration (s): 1.73 | learning rate: 5.845E-05 | global batch size: 512 | lm loss: 1.207456E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.594 | TFLOPs: 44.52 | 31: iteration 36580/ 52452 | consumed samples: 18728960 | consumed tokens: 38356910080 | elapsed time per iteration (s): 1.85 | learning rate: 5.840E-05 | global batch size: 512 | lm loss: 1.218111E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.051 | TFLOPs: 41.43 | 31: iteration 36590/ 52452 | consumed samples: 18734080 | consumed tokens: 38367395840 | elapsed time per iteration (s): 1.78 | learning rate: 5.836E-05 | global batch size: 512 | lm loss: 1.213646E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.167 | TFLOPs: 43.10 | 31: iteration 36600/ 52452 | consumed samples: 18739200 | consumed tokens: 38377881600 | elapsed time per iteration (s): 1.74 | learning rate: 5.831E-05 | global batch size: 512 | lm loss: 1.212820E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.827 | TFLOPs: 44.10 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36600 | lm loss value: 4.222711E+00 | lm loss PPL: 6.821815E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36600 | lm loss value: 4.079088E+00 | lm loss PPL: 5.909157E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36610/ 52452 | consumed samples: 18744320 | consumed tokens: 38388367360 | elapsed time per iteration (s): 9.19 | learning rate: 5.827E-05 | global batch size: 512 | lm loss: 1.207777E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.720 | TFLOPs: 8.36 | 31: iteration 36620/ 52452 | consumed samples: 18749440 | consumed tokens: 38398853120 | elapsed time per iteration (s): 1.77 | learning rate: 5.822E-05 | global batch size: 512 | lm loss: 1.216686E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.826 | TFLOPs: 43.50 | 31: iteration 36630/ 52452 | consumed samples: 18754560 | consumed tokens: 38409338880 | elapsed time per iteration (s): 1.76 | learning rate: 5.818E-05 | global batch size: 512 | lm loss: 1.210557E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.101 | TFLOPs: 43.54 | 31: iteration 36640/ 52452 | consumed samples: 18759680 | consumed tokens: 38419824640 | elapsed time per iteration (s): 1.78 | learning rate: 5.813E-05 | global batch size: 512 | lm loss: 1.207473E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.044 | TFLOPs: 43.23 | 31: iteration 36650/ 52452 | consumed samples: 18764800 | consumed tokens: 38430310400 | elapsed time per iteration (s): 1.73 | learning rate: 5.809E-05 | global batch size: 512 | lm loss: 1.205344E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.888 | TFLOPs: 44.41 | 31: iteration 36660/ 52452 | consumed samples: 18769920 | consumed tokens: 38440796160 | elapsed time per iteration (s): 1.87 | learning rate: 5.805E-05 | global batch size: 512 | lm loss: 1.211982E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.344 | TFLOPs: 41.18 | 31: iteration 36670/ 52452 | consumed samples: 18775040 | consumed tokens: 38451281920 | elapsed time per iteration (s): 1.76 | learning rate: 5.800E-05 | global batch size: 512 | lm loss: 1.220033E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.416 | TFLOPs: 43.74 | 31: iteration 36680/ 52452 | consumed samples: 18780160 | consumed tokens: 38461767680 | elapsed time per iteration (s): 1.75 | learning rate: 5.796E-05 | global batch size: 512 | lm loss: 1.224883E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.235 | TFLOPs: 44.01 | 31: iteration 36690/ 52452 | consumed samples: 18785280 | consumed tokens: 38472253440 | elapsed time per iteration (s): 1.80 | learning rate: 5.791E-05 | global batch size: 512 | lm loss: 1.221867E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.926 | TFLOPs: 42.62 | 31: iteration 36700/ 52452 | consumed samples: 18790400 | consumed tokens: 38482739200 | elapsed time per iteration (s): 1.73 | learning rate: 5.787E-05 | global batch size: 512 | lm loss: 1.210509E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.459 | TFLOPs: 44.35 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36700 | lm loss value: 4.151419E+00 | lm loss PPL: 6.352409E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36700 | lm loss value: 4.039859E+00 | lm loss PPL: 5.681835E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36710/ 52452 | consumed samples: 18795520 | consumed tokens: 38493224960 | elapsed time per iteration (s): 9.24 | learning rate: 5.782E-05 | global batch size: 512 | lm loss: 1.218840E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.382 | TFLOPs: 8.31 | 31: iteration 36720/ 52452 | consumed samples: 18800640 | consumed tokens: 38503710720 | elapsed time per iteration (s): 1.73 | learning rate: 5.778E-05 | global batch size: 512 | lm loss: 1.216842E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.757 | TFLOPs: 44.54 | 31: iteration 36730/ 52452 | consumed samples: 18805760 | consumed tokens: 38514196480 | elapsed time per iteration (s): 1.77 | learning rate: 5.773E-05 | global batch size: 512 | lm loss: 1.216744E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.958 | TFLOPs: 43.52 | 31: iteration 36740/ 52452 | consumed samples: 18810880 | consumed tokens: 38524682240 | elapsed time per iteration (s): 1.76 | learning rate: 5.769E-05 | global batch size: 512 | lm loss: 1.212798E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.645 | TFLOPs: 43.62 | 31: iteration 36750/ 52452 | consumed samples: 18816000 | consumed tokens: 38535168000 | elapsed time per iteration (s): 1.79 | learning rate: 5.765E-05 | global batch size: 512 | lm loss: 1.224161E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.481 | TFLOPs: 42.85 | 31: iteration 36760/ 52452 | consumed samples: 18821120 | consumed tokens: 38545653760 | elapsed time per iteration (s): 1.85 | learning rate: 5.760E-05 | global batch size: 512 | lm loss: 1.205454E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.698 | TFLOPs: 41.53 | 31: iteration 36770/ 52452 | consumed samples: 18826240 | consumed tokens: 38556139520 | elapsed time per iteration (s): 1.79 | learning rate: 5.756E-05 | global batch size: 512 | lm loss: 1.217088E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.099 | TFLOPs: 42.94 | 31: iteration 36780/ 52452 | consumed samples: 18831360 | consumed tokens: 38566625280 | elapsed time per iteration (s): 1.81 | learning rate: 5.751E-05 | global batch size: 512 | lm loss: 1.222699E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.126 | TFLOPs: 42.50 | 31: iteration 36790/ 52452 | consumed samples: 18836480 | consumed tokens: 38577111040 | elapsed time per iteration (s): 1.77 | learning rate: 5.747E-05 | global batch size: 512 | lm loss: 1.215382E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.081 | TFLOPs: 43.39 | 31: iteration 36800/ 52452 | consumed samples: 18841600 | consumed tokens: 38587596800 | elapsed time per iteration (s): 1.80 | learning rate: 5.743E-05 | global batch size: 512 | lm loss: 1.203140E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.753 | TFLOPs: 42.74 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36800 | lm loss value: 4.344815E+00 | lm loss PPL: 7.707780E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36800 | lm loss value: 4.180152E+00 | lm loss PPL: 6.537582E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36810/ 52452 | consumed samples: 18846720 | consumed tokens: 38598082560 | elapsed time per iteration (s): 9.21 | learning rate: 5.738E-05 | global batch size: 512 | lm loss: 1.206384E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.617 | TFLOPs: 8.35 | 31: iteration 36820/ 52452 | consumed samples: 18851840 | consumed tokens: 38608568320 | elapsed time per iteration (s): 1.88 | learning rate: 5.734E-05 | global batch size: 512 | lm loss: 1.211530E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.055 | TFLOPs: 40.83 | 31: iteration 36830/ 52452 | consumed samples: 18856960 | consumed tokens: 38619054080 | elapsed time per iteration (s): 1.85 | learning rate: 5.729E-05 | global batch size: 512 | lm loss: 1.215349E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.873 | TFLOPs: 41.56 | 31: iteration 36840/ 52452 | consumed samples: 18862080 | consumed tokens: 38629539840 | elapsed time per iteration (s): 1.79 | learning rate: 5.725E-05 | global batch size: 512 | lm loss: 1.207293E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.669 | TFLOPs: 43.03 | 31: iteration 36850/ 52452 | consumed samples: 18867200 | consumed tokens: 38640025600 | elapsed time per iteration (s): 1.74 | learning rate: 5.720E-05 | global batch size: 512 | lm loss: 1.210271E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.121 | TFLOPs: 44.15 | 31: iteration 36860/ 52452 | consumed samples: 18872320 | consumed tokens: 38650511360 | elapsed time per iteration (s): 1.80 | learning rate: 5.716E-05 | global batch size: 512 | lm loss: 1.229239E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.740 | TFLOPs: 42.74 | 31: iteration 36870/ 52452 | consumed samples: 18877440 | consumed tokens: 38660997120 | elapsed time per iteration (s): 1.79 | learning rate: 5.712E-05 | global batch size: 512 | lm loss: 1.208156E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.053 | TFLOPs: 42.93 | 31: iteration 36880/ 52452 | consumed samples: 18882560 | consumed tokens: 38671482880 | elapsed time per iteration (s): 1.74 | learning rate: 5.707E-05 | global batch size: 512 | lm loss: 1.218222E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.816 | TFLOPs: 44.25 | 31: iteration 36890/ 52452 | consumed samples: 18887680 | consumed tokens: 38681968640 | elapsed time per iteration (s): 1.75 | learning rate: 5.703E-05 | global batch size: 512 | lm loss: 1.202276E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.077 | TFLOPs: 43.84 | 31: iteration 36900/ 52452 | consumed samples: 18892800 | consumed tokens: 38692454400 | elapsed time per iteration (s): 1.78 | learning rate: 5.698E-05 | global batch size: 512 | lm loss: 1.204904E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.639 | TFLOPs: 43.17 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 36900 | lm loss value: 4.051376E+00 | lm loss PPL: 5.747651E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 36900 | lm loss value: 3.969667E+00 | lm loss PPL: 5.296688E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 36910/ 52452 | consumed samples: 18897920 | consumed tokens: 38702940160 | elapsed time per iteration (s): 9.21 | learning rate: 5.694E-05 | global batch size: 512 | lm loss: 1.214301E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.565 | TFLOPs: 8.34 | 31: iteration 36920/ 52452 | consumed samples: 18903040 | consumed tokens: 38713425920 | elapsed time per iteration (s): 1.74 | learning rate: 5.690E-05 | global batch size: 512 | lm loss: 1.213241E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.582 | TFLOPs: 44.22 | 31: iteration 36930/ 52452 | consumed samples: 18908160 | consumed tokens: 38723911680 | elapsed time per iteration (s): 1.74 | learning rate: 5.685E-05 | global batch size: 512 | lm loss: 1.208456E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.748 | TFLOPs: 44.09 | 31: iteration 36940/ 52452 | consumed samples: 18913280 | consumed tokens: 38734397440 | elapsed time per iteration (s): 1.76 | learning rate: 5.681E-05 | global batch size: 512 | lm loss: 1.221879E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.300 | TFLOPs: 43.57 | 31: iteration 36950/ 52452 | consumed samples: 18918400 | consumed tokens: 38744883200 | elapsed time per iteration (s): 1.83 | learning rate: 5.676E-05 | global batch size: 512 | lm loss: 1.214636E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.347 | TFLOPs: 42.08 | 31: iteration 36960/ 52452 | consumed samples: 18923520 | consumed tokens: 38755368960 | elapsed time per iteration (s): 1.78 | learning rate: 5.672E-05 | global batch size: 512 | lm loss: 1.211161E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.277 | TFLOPs: 43.12 | 31: iteration 36970/ 52452 | consumed samples: 18928640 | consumed tokens: 38765854720 | elapsed time per iteration (s): 1.77 | learning rate: 5.668E-05 | global batch size: 512 | lm loss: 1.217731E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.360 | TFLOPs: 43.43 | 31: iteration 36980/ 52452 | consumed samples: 18933760 | consumed tokens: 38776340480 | elapsed time per iteration (s): 1.79 | learning rate: 5.663E-05 | global batch size: 512 | lm loss: 1.222513E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.210 | TFLOPs: 42.96 | 31: iteration 36990/ 52452 | consumed samples: 18938880 | consumed tokens: 38786826240 | elapsed time per iteration (s): 1.76 | learning rate: 5.659E-05 | global batch size: 512 | lm loss: 1.214336E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.996 | TFLOPs: 43.68 | 31: iteration 37000/ 52452 | consumed samples: 18944000 | consumed tokens: 38797312000 | elapsed time per iteration (s): 1.75 | learning rate: 5.655E-05 | global batch size: 512 | lm loss: 1.222952E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.435 | TFLOPs: 43.89 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37000 | lm loss value: 4.139305E+00 | lm loss PPL: 6.275920E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37000 | lm loss value: 4.026823E+00 | lm loss PPL: 5.608243E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37010/ 52452 | consumed samples: 18949120 | consumed tokens: 38807797760 | elapsed time per iteration (s): 9.17 | learning rate: 5.650E-05 | global batch size: 512 | lm loss: 1.209325E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.849 | TFLOPs: 8.38 | 31: iteration 37020/ 52452 | consumed samples: 18954240 | consumed tokens: 38818283520 | elapsed time per iteration (s): 1.80 | learning rate: 5.646E-05 | global batch size: 512 | lm loss: 1.207290E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.844 | TFLOPs: 42.75 | 31: iteration 37030/ 52452 | consumed samples: 18959360 | consumed tokens: 38828769280 | elapsed time per iteration (s): 1.76 | learning rate: 5.641E-05 | global batch size: 512 | lm loss: 1.213938E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.469 | TFLOPs: 43.60 | 31: iteration 37040/ 52452 | consumed samples: 18964480 | consumed tokens: 38839255040 | elapsed time per iteration (s): 1.79 | learning rate: 5.637E-05 | global batch size: 512 | lm loss: 1.216129E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.284 | TFLOPs: 42.82 | 31: iteration 37050/ 52452 | consumed samples: 18969600 | consumed tokens: 38849740800 | elapsed time per iteration (s): 1.75 | learning rate: 5.633E-05 | global batch size: 512 | lm loss: 1.223566E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.138 | TFLOPs: 43.85 | 31: iteration 37060/ 52452 | consumed samples: 18974720 | consumed tokens: 38860226560 | elapsed time per iteration (s): 1.75 | learning rate: 5.628E-05 | global batch size: 512 | lm loss: 1.219242E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.293 | TFLOPs: 43.87 | 31: iteration 37070/ 52452 | consumed samples: 18979840 | consumed tokens: 38870712320 | elapsed time per iteration (s): 1.79 | learning rate: 5.624E-05 | global batch size: 512 | lm loss: 1.201055E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.796 | TFLOPs: 43.05 | 31: iteration 37080/ 52452 | consumed samples: 18984960 | consumed tokens: 38881198080 | elapsed time per iteration (s): 1.82 | learning rate: 5.620E-05 | global batch size: 512 | lm loss: 1.209762E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.024 | TFLOPs: 42.18 | 31: iteration 37090/ 52452 | consumed samples: 18990080 | consumed tokens: 38891683840 | elapsed time per iteration (s): 1.77 | learning rate: 5.615E-05 | global batch size: 512 | lm loss: 1.213028E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.920 | TFLOPs: 43.52 | 31: iteration 37100/ 52452 | consumed samples: 18995200 | consumed tokens: 38902169600 | elapsed time per iteration (s): 1.75 | learning rate: 5.611E-05 | global batch size: 512 | lm loss: 1.200199E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.376 | TFLOPs: 43.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37100 | lm loss value: 4.277017E+00 | lm loss PPL: 7.202524E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37100 | lm loss value: 4.127115E+00 | lm loss PPL: 6.199878E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37110/ 52452 | consumed samples: 19000320 | consumed tokens: 38912655360 | elapsed time per iteration (s): 9.21 | learning rate: 5.606E-05 | global batch size: 512 | lm loss: 1.226302E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.583 | TFLOPs: 8.34 | 31: iteration 37120/ 52452 | consumed samples: 19005440 | consumed tokens: 38923141120 | elapsed time per iteration (s): 1.77 | learning rate: 5.602E-05 | global batch size: 512 | lm loss: 1.217125E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.147 | TFLOPs: 43.40 | 31: iteration 37130/ 52452 | consumed samples: 19010560 | consumed tokens: 38933626880 | elapsed time per iteration (s): 1.74 | learning rate: 5.598E-05 | global batch size: 512 | lm loss: 1.212774E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.468 | TFLOPs: 44.05 | 31: iteration 37140/ 52452 | consumed samples: 19015680 | consumed tokens: 38944112640 | elapsed time per iteration (s): 1.85 | learning rate: 5.593E-05 | global batch size: 512 | lm loss: 1.211077E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.186 | TFLOPs: 41.60 | 31: iteration 37150/ 52452 | consumed samples: 19020800 | consumed tokens: 38954598400 | elapsed time per iteration (s): 1.81 | learning rate: 5.589E-05 | global batch size: 512 | lm loss: 1.202882E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.025 | TFLOPs: 42.48 | 31: iteration 37160/ 52452 | consumed samples: 19025920 | consumed tokens: 38965084160 | elapsed time per iteration (s): 1.81 | learning rate: 5.585E-05 | global batch size: 512 | lm loss: 1.215500E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.242 | TFLOPs: 42.51 | 31: iteration 37170/ 52452 | consumed samples: 19031040 | consumed tokens: 38975569920 | elapsed time per iteration (s): 1.73 | learning rate: 5.580E-05 | global batch size: 512 | lm loss: 1.212140E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.394 | TFLOPs: 44.49 | 31: iteration 37180/ 52452 | consumed samples: 19036160 | consumed tokens: 38986055680 | elapsed time per iteration (s): 1.81 | learning rate: 5.576E-05 | global batch size: 512 | lm loss: 1.207754E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.231 | TFLOPs: 42.51 | 31: iteration 37190/ 52452 | consumed samples: 19041280 | consumed tokens: 38996541440 | elapsed time per iteration (s): 1.75 | learning rate: 5.572E-05 | global batch size: 512 | lm loss: 1.215398E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.042 | TFLOPs: 43.98 | 31: iteration 37200/ 52452 | consumed samples: 19046400 | consumed tokens: 39007027200 | elapsed time per iteration (s): 1.82 | learning rate: 5.567E-05 | global batch size: 512 | lm loss: 1.212110E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.618 | TFLOPs: 42.12 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37200 | lm loss value: 4.131926E+00 | lm loss PPL: 6.229780E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37200 | lm loss value: 4.022179E+00 | lm loss PPL: 5.582259E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37210/ 52452 | consumed samples: 19051520 | consumed tokens: 39017512960 | elapsed time per iteration (s): 9.25 | learning rate: 5.563E-05 | global batch size: 512 | lm loss: 1.211879E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.334 | TFLOPs: 8.31 | 31: iteration 37220/ 52452 | consumed samples: 19056640 | consumed tokens: 39027998720 | elapsed time per iteration (s): 1.77 | learning rate: 5.559E-05 | global batch size: 512 | lm loss: 1.209575E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.269 | TFLOPs: 43.42 | 31: iteration 37230/ 52452 | consumed samples: 19061760 | consumed tokens: 39038484480 | elapsed time per iteration (s): 1.82 | learning rate: 5.554E-05 | global batch size: 512 | lm loss: 1.220424E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.705 | TFLOPs: 42.13 | 31: iteration 37240/ 52452 | consumed samples: 19066880 | consumed tokens: 39048970240 | elapsed time per iteration (s): 1.99 | learning rate: 5.550E-05 | global batch size: 512 | lm loss: 1.211656E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 257.887 | TFLOPs: 38.71 | 31: iteration 37250/ 52452 | consumed samples: 19072000 | consumed tokens: 39059456000 | elapsed time per iteration (s): 1.72 | learning rate: 5.546E-05 | global batch size: 512 | lm loss: 1.217117E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.340 | TFLOPs: 44.63 | 31: iteration 37260/ 52452 | consumed samples: 19077120 | consumed tokens: 39069941760 | elapsed time per iteration (s): 1.81 | learning rate: 5.541E-05 | global batch size: 512 | lm loss: 1.223224E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.177 | TFLOPs: 42.35 | 31: iteration 37270/ 52452 | consumed samples: 19082240 | consumed tokens: 39080427520 | elapsed time per iteration (s): 1.77 | learning rate: 5.537E-05 | global batch size: 512 | lm loss: 1.220232E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.247 | TFLOPs: 43.41 | 31: iteration 37280/ 52452 | consumed samples: 19087360 | consumed tokens: 39090913280 | elapsed time per iteration (s): 1.83 | learning rate: 5.533E-05 | global batch size: 512 | lm loss: 1.223097E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.096 | TFLOPs: 42.04 | 31: iteration 37290/ 52452 | consumed samples: 19092480 | consumed tokens: 39101399040 | elapsed time per iteration (s): 1.97 | learning rate: 5.528E-05 | global batch size: 512 | lm loss: 1.216404E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 259.278 | TFLOPs: 38.92 | 31: iteration 37300/ 52452 | consumed samples: 19097600 | consumed tokens: 39111884800 | elapsed time per iteration (s): 1.78 | learning rate: 5.524E-05 | global batch size: 512 | lm loss: 1.219979E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.365 | TFLOPs: 43.13 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37300 | lm loss value: 4.148770E+00 | lm loss PPL: 6.335605E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37300 | lm loss value: 4.034903E+00 | lm loss PPL: 5.653741E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37310/ 52452 | consumed samples: 19102720 | consumed tokens: 39122370560 | elapsed time per iteration (s): 9.15 | learning rate: 5.520E-05 | global batch size: 512 | lm loss: 1.207291E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.926 | TFLOPs: 8.39 | 31: iteration 37320/ 52452 | consumed samples: 19107840 | consumed tokens: 39132856320 | elapsed time per iteration (s): 1.74 | learning rate: 5.515E-05 | global batch size: 512 | lm loss: 1.216084E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.628 | TFLOPs: 44.07 | 31: iteration 37330/ 52452 | consumed samples: 19112960 | consumed tokens: 39143342080 | elapsed time per iteration (s): 1.83 | learning rate: 5.511E-05 | global batch size: 512 | lm loss: 1.219170E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.164 | TFLOPs: 42.05 | 31: iteration 37340/ 52452 | consumed samples: 19118080 | consumed tokens: 39153827840 | elapsed time per iteration (s): 1.74 | learning rate: 5.507E-05 | global batch size: 512 | lm loss: 1.216638E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.095 | TFLOPs: 44.29 | 31: iteration 37350/ 52452 | consumed samples: 19123200 | consumed tokens: 39164313600 | elapsed time per iteration (s): 1.77 | learning rate: 5.502E-05 | global batch size: 512 | lm loss: 1.216203E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.000 | TFLOPs: 43.38 | 31: iteration 37360/ 52452 | consumed samples: 19128320 | consumed tokens: 39174799360 | elapsed time per iteration (s): 1.78 | learning rate: 5.498E-05 | global batch size: 512 | lm loss: 1.219441E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.903 | TFLOPs: 43.21 | 31: iteration 37370/ 52452 | consumed samples: 19133440 | consumed tokens: 39185285120 | elapsed time per iteration (s): 1.75 | learning rate: 5.494E-05 | global batch size: 512 | lm loss: 1.210631E+00 | grad norm: 0.182 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.271 | TFLOPs: 43.87 | 31: iteration 37380/ 52452 | consumed samples: 19138560 | consumed tokens: 39195770880 | elapsed time per iteration (s): 1.74 | learning rate: 5.489E-05 | global batch size: 512 | lm loss: 1.219090E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.242 | TFLOPs: 44.16 | 31: iteration 37390/ 52452 | consumed samples: 19143680 | consumed tokens: 39206256640 | elapsed time per iteration (s): 1.77 | learning rate: 5.485E-05 | global batch size: 512 | lm loss: 1.200389E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.921 | TFLOPs: 43.52 | 31: iteration 37400/ 52452 | consumed samples: 19148800 | consumed tokens: 39216742400 | elapsed time per iteration (s): 1.77 | learning rate: 5.481E-05 | global batch size: 512 | lm loss: 1.211646E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.381 | TFLOPs: 43.43 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37400 | lm loss value: 4.178822E+00 | lm loss PPL: 6.528890E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37400 | lm loss value: 4.068494E+00 | lm loss PPL: 5.846883E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37410/ 52452 | consumed samples: 19153920 | consumed tokens: 39227228160 | elapsed time per iteration (s): 9.22 | learning rate: 5.477E-05 | global batch size: 512 | lm loss: 1.208823E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.517 | TFLOPs: 8.33 | 31: iteration 37420/ 52452 | consumed samples: 19159040 | consumed tokens: 39237713920 | elapsed time per iteration (s): 1.81 | learning rate: 5.472E-05 | global batch size: 512 | lm loss: 1.217632E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.174 | TFLOPs: 42.50 | 31: iteration 37430/ 52452 | consumed samples: 19164160 | consumed tokens: 39248199680 | elapsed time per iteration (s): 1.84 | learning rate: 5.468E-05 | global batch size: 512 | lm loss: 1.217260E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.983 | TFLOPs: 41.87 | 31: iteration 37440/ 52452 | consumed samples: 19169280 | consumed tokens: 39258685440 | elapsed time per iteration (s): 1.77 | learning rate: 5.464E-05 | global batch size: 512 | lm loss: 1.212579E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.575 | TFLOPs: 43.46 | 31: iteration 37450/ 52452 | consumed samples: 19174400 | consumed tokens: 39269171200 | elapsed time per iteration (s): 1.77 | learning rate: 5.459E-05 | global batch size: 512 | lm loss: 1.209690E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.868 | TFLOPs: 43.51 | 31: iteration 37460/ 52452 | consumed samples: 19179520 | consumed tokens: 39279656960 | elapsed time per iteration (s): 1.78 | learning rate: 5.455E-05 | global batch size: 512 | lm loss: 1.203312E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.000 | TFLOPs: 43.08 | 31: iteration 37470/ 52452 | consumed samples: 19184640 | consumed tokens: 39290142720 | elapsed time per iteration (s): 1.78 | learning rate: 5.451E-05 | global batch size: 512 | lm loss: 1.222391E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.179 | TFLOPs: 43.10 | 31: iteration 37480/ 52452 | consumed samples: 19189760 | consumed tokens: 39300628480 | elapsed time per iteration (s): 1.75 | learning rate: 5.447E-05 | global batch size: 512 | lm loss: 1.217101E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.293 | TFLOPs: 43.87 | 31: iteration 37490/ 52452 | consumed samples: 19194880 | consumed tokens: 39311114240 | elapsed time per iteration (s): 1.75 | learning rate: 5.442E-05 | global batch size: 512 | lm loss: 1.216795E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.765 | TFLOPs: 43.94 | 31: iteration 37500/ 52452 | consumed samples: 19200000 | consumed tokens: 39321600000 | elapsed time per iteration (s): 1.76 | learning rate: 5.438E-05 | global batch size: 512 | lm loss: 1.221502E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.347 | TFLOPs: 43.73 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37500 | lm loss value: 4.218983E+00 | lm loss PPL: 6.796431E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37500 | lm loss value: 4.069555E+00 | lm loss PPL: 5.853090E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37510/ 52452 | consumed samples: 19205120 | consumed tokens: 39332085760 | elapsed time per iteration (s): 9.19 | learning rate: 5.434E-05 | global batch size: 512 | lm loss: 1.215309E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.693 | TFLOPs: 8.36 | 31: iteration 37520/ 52452 | consumed samples: 19210240 | consumed tokens: 39342571520 | elapsed time per iteration (s): 1.80 | learning rate: 5.429E-05 | global batch size: 512 | lm loss: 1.208870E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.920 | TFLOPs: 42.61 | 31: iteration 37530/ 52452 | consumed samples: 19215360 | consumed tokens: 39353057280 | elapsed time per iteration (s): 1.73 | learning rate: 5.425E-05 | global batch size: 512 | lm loss: 1.207527E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.089 | TFLOPs: 44.44 | 31: iteration 37540/ 52452 | consumed samples: 19220480 | consumed tokens: 39363543040 | elapsed time per iteration (s): 1.75 | learning rate: 5.421E-05 | global batch size: 512 | lm loss: 1.218452E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.170 | TFLOPs: 43.85 | 31: iteration 37550/ 52452 | consumed samples: 19225600 | consumed tokens: 39374028800 | elapsed time per iteration (s): 1.72 | learning rate: 5.417E-05 | global batch size: 512 | lm loss: 1.212181E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.531 | TFLOPs: 44.66 | 31: iteration 37560/ 52452 | consumed samples: 19230720 | consumed tokens: 39384514560 | elapsed time per iteration (s): 1.76 | learning rate: 5.412E-05 | global batch size: 512 | lm loss: 1.210096E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.138 | TFLOPs: 43.70 | 31: iteration 37570/ 52452 | consumed samples: 19235840 | consumed tokens: 39395000320 | elapsed time per iteration (s): 1.77 | learning rate: 5.408E-05 | global batch size: 512 | lm loss: 1.218474E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.151 | TFLOPs: 43.40 | 31: iteration 37580/ 52452 | consumed samples: 19240960 | consumed tokens: 39405486080 | elapsed time per iteration (s): 1.80 | learning rate: 5.404E-05 | global batch size: 512 | lm loss: 1.216476E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.273 | TFLOPs: 42.67 | 31: iteration 37590/ 52452 | consumed samples: 19246080 | consumed tokens: 39415971840 | elapsed time per iteration (s): 1.77 | learning rate: 5.400E-05 | global batch size: 512 | lm loss: 1.227634E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.043 | TFLOPs: 43.38 | 31: iteration 37600/ 52452 | consumed samples: 19251200 | consumed tokens: 39426457600 | elapsed time per iteration (s): 1.76 | learning rate: 5.395E-05 | global batch size: 512 | lm loss: 1.216358E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.017 | TFLOPs: 43.68 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37600 | lm loss value: 4.151406E+00 | lm loss PPL: 6.352327E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37600 | lm loss value: 4.029494E+00 | lm loss PPL: 5.623244E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37610/ 52452 | consumed samples: 19256320 | consumed tokens: 39436943360 | elapsed time per iteration (s): 9.15 | learning rate: 5.391E-05 | global batch size: 512 | lm loss: 1.217511E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.934 | TFLOPs: 8.40 | 31: iteration 37620/ 52452 | consumed samples: 19261440 | consumed tokens: 39447429120 | elapsed time per iteration (s): 2.00 | learning rate: 5.387E-05 | global batch size: 512 | lm loss: 1.203864E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 255.433 | TFLOPs: 38.34 | 31: iteration 37630/ 52452 | consumed samples: 19266560 | consumed tokens: 39457914880 | elapsed time per iteration (s): 1.82 | learning rate: 5.382E-05 | global batch size: 512 | lm loss: 1.207754E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.915 | TFLOPs: 42.31 | 31: iteration 37640/ 52452 | consumed samples: 19271680 | consumed tokens: 39468400640 | elapsed time per iteration (s): 1.79 | learning rate: 5.378E-05 | global batch size: 512 | lm loss: 1.207112E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.944 | TFLOPs: 42.92 | 31: iteration 37650/ 52452 | consumed samples: 19276800 | consumed tokens: 39478886400 | elapsed time per iteration (s): 1.79 | learning rate: 5.374E-05 | global batch size: 512 | lm loss: 1.205067E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.832 | TFLOPs: 43.05 | 31: iteration 37660/ 52452 | consumed samples: 19281920 | consumed tokens: 39489372160 | elapsed time per iteration (s): 1.76 | learning rate: 5.370E-05 | global batch size: 512 | lm loss: 1.205818E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.226 | TFLOPs: 43.71 | 31: iteration 37670/ 52452 | consumed samples: 19287040 | consumed tokens: 39499857920 | elapsed time per iteration (s): 1.73 | learning rate: 5.365E-05 | global batch size: 512 | lm loss: 1.212827E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.782 | TFLOPs: 44.55 | 31: iteration 37680/ 52452 | consumed samples: 19292160 | consumed tokens: 39510343680 | elapsed time per iteration (s): 1.74 | learning rate: 5.361E-05 | global batch size: 512 | lm loss: 1.219681E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.890 | TFLOPs: 44.11 | 31: iteration 37690/ 52452 | consumed samples: 19297280 | consumed tokens: 39520829440 | elapsed time per iteration (s): 1.77 | learning rate: 5.357E-05 | global batch size: 512 | lm loss: 1.209710E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.157 | TFLOPs: 43.40 | 31: iteration 37700/ 52452 | consumed samples: 19302400 | consumed tokens: 39531315200 | elapsed time per iteration (s): 1.76 | learning rate: 5.353E-05 | global batch size: 512 | lm loss: 1.209256E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.980 | TFLOPs: 43.67 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37700 | lm loss value: 4.201333E+00 | lm loss PPL: 6.677525E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37700 | lm loss value: 4.088689E+00 | lm loss PPL: 5.966162E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37710/ 52452 | consumed samples: 19307520 | consumed tokens: 39541800960 | elapsed time per iteration (s): 9.19 | learning rate: 5.349E-05 | global batch size: 512 | lm loss: 1.208271E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.742 | TFLOPs: 8.37 | 31: iteration 37720/ 52452 | consumed samples: 19312640 | consumed tokens: 39552286720 | elapsed time per iteration (s): 1.73 | learning rate: 5.344E-05 | global batch size: 512 | lm loss: 1.213539E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.435 | TFLOPs: 44.34 | 31: iteration 37730/ 52452 | consumed samples: 19317760 | consumed tokens: 39562772480 | elapsed time per iteration (s): 1.77 | learning rate: 5.340E-05 | global batch size: 512 | lm loss: 1.209646E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.111 | TFLOPs: 43.39 | 31: iteration 37740/ 52452 | consumed samples: 19322880 | consumed tokens: 39573258240 | elapsed time per iteration (s): 1.75 | learning rate: 5.336E-05 | global batch size: 512 | lm loss: 1.208598E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.752 | TFLOPs: 43.94 | 31: iteration 37750/ 52452 | consumed samples: 19328000 | consumed tokens: 39583744000 | elapsed time per iteration (s): 1.72 | learning rate: 5.332E-05 | global batch size: 512 | lm loss: 1.204278E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.527 | TFLOPs: 44.66 | 31: iteration 37760/ 52452 | consumed samples: 19333120 | consumed tokens: 39594229760 | elapsed time per iteration (s): 1.74 | learning rate: 5.327E-05 | global batch size: 512 | lm loss: 1.202342E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.663 | TFLOPs: 44.23 | 31: iteration 37770/ 52452 | consumed samples: 19338240 | consumed tokens: 39604715520 | elapsed time per iteration (s): 1.76 | learning rate: 5.323E-05 | global batch size: 512 | lm loss: 1.209922E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.261 | TFLOPs: 43.72 | 31: iteration 37780/ 52452 | consumed samples: 19343360 | consumed tokens: 39615201280 | elapsed time per iteration (s): 1.72 | learning rate: 5.319E-05 | global batch size: 512 | lm loss: 1.210150E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.575 | TFLOPs: 44.66 | 31: iteration 37790/ 52452 | consumed samples: 19348480 | consumed tokens: 39625687040 | elapsed time per iteration (s): 1.72 | learning rate: 5.315E-05 | global batch size: 512 | lm loss: 1.202940E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.555 | TFLOPs: 44.66 | 31: iteration 37800/ 52452 | consumed samples: 19353600 | consumed tokens: 39636172800 | elapsed time per iteration (s): 1.78 | learning rate: 5.310E-05 | global batch size: 512 | lm loss: 1.210539E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.529 | TFLOPs: 43.16 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37800 | lm loss value: 4.211317E+00 | lm loss PPL: 6.744528E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37800 | lm loss value: 4.082330E+00 | lm loss PPL: 5.928345E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37810/ 52452 | consumed samples: 19358720 | consumed tokens: 39646658560 | elapsed time per iteration (s): 9.22 | learning rate: 5.306E-05 | global batch size: 512 | lm loss: 1.209745E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.524 | TFLOPs: 8.33 | 31: iteration 37820/ 52452 | consumed samples: 19363840 | consumed tokens: 39657144320 | elapsed time per iteration (s): 1.82 | learning rate: 5.302E-05 | global batch size: 512 | lm loss: 1.210984E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.974 | TFLOPs: 42.32 | 31: iteration 37830/ 52452 | consumed samples: 19368960 | consumed tokens: 39667630080 | elapsed time per iteration (s): 1.77 | learning rate: 5.298E-05 | global batch size: 512 | lm loss: 1.210387E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.355 | TFLOPs: 43.43 | 31: iteration 37840/ 52452 | consumed samples: 19374080 | consumed tokens: 39678115840 | elapsed time per iteration (s): 1.83 | learning rate: 5.294E-05 | global batch size: 512 | lm loss: 1.211358E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.631 | TFLOPs: 41.97 | 31: iteration 37850/ 52452 | consumed samples: 19379200 | consumed tokens: 39688601600 | elapsed time per iteration (s): 1.74 | learning rate: 5.289E-05 | global batch size: 512 | lm loss: 1.201542E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.709 | TFLOPs: 44.08 | 31: iteration 37860/ 52452 | consumed samples: 19384320 | consumed tokens: 39699087360 | elapsed time per iteration (s): 1.76 | learning rate: 5.285E-05 | global batch size: 512 | lm loss: 1.207218E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.636 | TFLOPs: 43.77 | 31: iteration 37870/ 52452 | consumed samples: 19389440 | consumed tokens: 39709573120 | elapsed time per iteration (s): 1.77 | learning rate: 5.281E-05 | global batch size: 512 | lm loss: 1.206637E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.524 | TFLOPs: 43.46 | 31: iteration 37880/ 52452 | consumed samples: 19394560 | consumed tokens: 39720058880 | elapsed time per iteration (s): 1.81 | learning rate: 5.277E-05 | global batch size: 512 | lm loss: 1.208482E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.201 | TFLOPs: 42.51 | 31: iteration 37890/ 52452 | consumed samples: 19399680 | consumed tokens: 39730544640 | elapsed time per iteration (s): 1.82 | learning rate: 5.273E-05 | global batch size: 512 | lm loss: 1.208135E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.514 | TFLOPs: 42.25 | 31: iteration 37900/ 52452 | consumed samples: 19404800 | consumed tokens: 39741030400 | elapsed time per iteration (s): 1.77 | learning rate: 5.268E-05 | global batch size: 512 | lm loss: 1.215688E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.966 | TFLOPs: 43.52 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 37900 | lm loss value: 4.120909E+00 | lm loss PPL: 6.161524E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 37900 | lm loss value: 4.019612E+00 | lm loss PPL: 5.567949E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 37910/ 52452 | consumed samples: 19409920 | consumed tokens: 39751516160 | elapsed time per iteration (s): 9.20 | learning rate: 5.264E-05 | global batch size: 512 | lm loss: 1.224612E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.679 | TFLOPs: 8.36 | 31: iteration 37920/ 52452 | consumed samples: 19415040 | consumed tokens: 39762001920 | elapsed time per iteration (s): 1.79 | learning rate: 5.260E-05 | global batch size: 512 | lm loss: 1.209928E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.416 | TFLOPs: 42.99 | 31: iteration 37930/ 52452 | consumed samples: 19420160 | consumed tokens: 39772487680 | elapsed time per iteration (s): 1.80 | learning rate: 5.256E-05 | global batch size: 512 | lm loss: 1.203578E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.487 | TFLOPs: 42.70 | 31: iteration 37940/ 52452 | consumed samples: 19425280 | consumed tokens: 39782973440 | elapsed time per iteration (s): 1.80 | learning rate: 5.252E-05 | global batch size: 512 | lm loss: 1.194410E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.631 | TFLOPs: 42.72 | 31: iteration 37950/ 52452 | consumed samples: 19430400 | consumed tokens: 39793459200 | elapsed time per iteration (s): 1.81 | learning rate: 5.247E-05 | global batch size: 512 | lm loss: 1.202552E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.396 | TFLOPs: 42.54 | 31: iteration 37960/ 52452 | consumed samples: 19435520 | consumed tokens: 39803944960 | elapsed time per iteration (s): 1.80 | learning rate: 5.243E-05 | global batch size: 512 | lm loss: 1.207574E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.068 | TFLOPs: 42.79 | 31: iteration 37970/ 52452 | consumed samples: 19440640 | consumed tokens: 39814430720 | elapsed time per iteration (s): 1.77 | learning rate: 5.239E-05 | global batch size: 512 | lm loss: 1.205925E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.725 | TFLOPs: 43.34 | 31: iteration 37980/ 52452 | consumed samples: 19445760 | consumed tokens: 39824916480 | elapsed time per iteration (s): 1.83 | learning rate: 5.235E-05 | global batch size: 512 | lm loss: 1.211239E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.800 | TFLOPs: 42.00 | 31: iteration 37990/ 52452 | consumed samples: 19450880 | consumed tokens: 39835402240 | elapsed time per iteration (s): 1.76 | learning rate: 5.231E-05 | global batch size: 512 | lm loss: 1.222125E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.241 | TFLOPs: 43.56 | 0: [2022-12-26 11:44:14,233] [INFO] [logging.py:68:log_dist] [Rank 0] step=38000, skipped=0, lr=[5.226500204809012e-05, 5.226500204809012e-05, 5.226500204809012e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 38000/ 52452 | consumed samples: 19456000 | consumed tokens: 39845888000 | elapsed time per iteration (s): 1.73 | learning rate: 5.227E-05 | global batch size: 512 | lm loss: 1.205016E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.769 | TFLOPs: 44.39 | 0: steps: 38000 loss: 1.2161 iter time (s): 1.780 samples/sec: 287.708 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38000 | lm loss value: 4.131752E+00 | lm loss PPL: 6.228693E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38000 | lm loss value: 4.026759E+00 | lm loss PPL: 5.607887E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38010/ 52452 | consumed samples: 19461120 | consumed tokens: 39856373760 | elapsed time per iteration (s): 9.24 | learning rate: 5.222E-05 | global batch size: 512 | lm loss: 1.207956E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.404 | TFLOPs: 8.32 | 31: iteration 38020/ 52452 | consumed samples: 19466240 | consumed tokens: 39866859520 | elapsed time per iteration (s): 1.77 | learning rate: 5.218E-05 | global batch size: 512 | lm loss: 1.209186E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.593 | TFLOPs: 43.32 | 31: iteration 38030/ 52452 | consumed samples: 19471360 | consumed tokens: 39877345280 | elapsed time per iteration (s): 1.75 | learning rate: 5.214E-05 | global batch size: 512 | lm loss: 1.200420E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.758 | TFLOPs: 43.79 | 31: iteration 38040/ 52452 | consumed samples: 19476480 | consumed tokens: 39887831040 | elapsed time per iteration (s): 1.76 | learning rate: 5.210E-05 | global batch size: 512 | lm loss: 1.214950E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.590 | TFLOPs: 43.62 | 31: iteration 38050/ 52452 | consumed samples: 19481600 | consumed tokens: 39898316800 | elapsed time per iteration (s): 1.75 | learning rate: 5.206E-05 | global batch size: 512 | lm loss: 1.201236E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.201 | TFLOPs: 43.86 | 31: iteration 38060/ 52452 | consumed samples: 19486720 | consumed tokens: 39908802560 | elapsed time per iteration (s): 1.75 | learning rate: 5.201E-05 | global batch size: 512 | lm loss: 1.198123E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.229 | TFLOPs: 43.86 | 31: iteration 38070/ 52452 | consumed samples: 19491840 | consumed tokens: 39919288320 | elapsed time per iteration (s): 1.74 | learning rate: 5.197E-05 | global batch size: 512 | lm loss: 1.215281E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.519 | TFLOPs: 44.06 | 31: iteration 38080/ 52452 | consumed samples: 19496960 | consumed tokens: 39929774080 | elapsed time per iteration (s): 1.77 | learning rate: 5.193E-05 | global batch size: 512 | lm loss: 1.208917E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.782 | TFLOPs: 43.34 | 31: iteration 38090/ 52452 | consumed samples: 19502080 | consumed tokens: 39940259840 | elapsed time per iteration (s): 1.79 | learning rate: 5.189E-05 | global batch size: 512 | lm loss: 1.208699E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.565 | TFLOPs: 42.86 | 31: iteration 38100/ 52452 | consumed samples: 19507200 | consumed tokens: 39950745600 | elapsed time per iteration (s): 1.78 | learning rate: 5.185E-05 | global batch size: 512 | lm loss: 1.213961E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.680 | TFLOPs: 43.18 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38100 | lm loss value: 4.170744E+00 | lm loss PPL: 6.476365E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38100 | lm loss value: 4.050838E+00 | lm loss PPL: 5.744555E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38110/ 52452 | consumed samples: 19512320 | consumed tokens: 39961231360 | elapsed time per iteration (s): 9.22 | learning rate: 5.181E-05 | global batch size: 512 | lm loss: 1.205389E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.504 | TFLOPs: 8.33 | 31: iteration 38120/ 52452 | consumed samples: 19517440 | consumed tokens: 39971717120 | elapsed time per iteration (s): 1.76 | learning rate: 5.177E-05 | global batch size: 512 | lm loss: 1.205740E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.004 | TFLOPs: 43.68 | 31: iteration 38130/ 52452 | consumed samples: 19522560 | consumed tokens: 39982202880 | elapsed time per iteration (s): 1.93 | learning rate: 5.172E-05 | global batch size: 512 | lm loss: 1.210507E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.329 | TFLOPs: 39.82 | 31: iteration 38140/ 52452 | consumed samples: 19527680 | consumed tokens: 39992688640 | elapsed time per iteration (s): 1.85 | learning rate: 5.168E-05 | global batch size: 512 | lm loss: 1.200723E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.350 | TFLOPs: 41.63 | 31: iteration 38150/ 52452 | consumed samples: 19532800 | consumed tokens: 40003174400 | elapsed time per iteration (s): 1.72 | learning rate: 5.164E-05 | global batch size: 512 | lm loss: 1.210917E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.985 | TFLOPs: 44.58 | 31: iteration 38160/ 52452 | consumed samples: 19537920 | consumed tokens: 40013660160 | elapsed time per iteration (s): 1.74 | learning rate: 5.160E-05 | global batch size: 512 | lm loss: 1.196594E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.686 | TFLOPs: 44.08 | 31: iteration 38170/ 52452 | consumed samples: 19543040 | consumed tokens: 40024145920 | elapsed time per iteration (s): 1.80 | learning rate: 5.156E-05 | global batch size: 512 | lm loss: 1.223808E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.509 | TFLOPs: 42.70 | 31: iteration 38180/ 52452 | consumed samples: 19548160 | consumed tokens: 40034631680 | elapsed time per iteration (s): 1.80 | learning rate: 5.152E-05 | global batch size: 512 | lm loss: 1.204280E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.405 | TFLOPs: 42.69 | 31: iteration 38190/ 52452 | consumed samples: 19553280 | consumed tokens: 40045117440 | elapsed time per iteration (s): 1.73 | learning rate: 5.148E-05 | global batch size: 512 | lm loss: 1.207922E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.727 | TFLOPs: 44.39 | 31: iteration 38200/ 52452 | consumed samples: 19558400 | consumed tokens: 40055603200 | elapsed time per iteration (s): 1.74 | learning rate: 5.143E-05 | global batch size: 512 | lm loss: 1.212402E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.166 | TFLOPs: 44.15 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38200 | lm loss value: 4.206538E+00 | lm loss PPL: 6.712377E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38200 | lm loss value: 4.074882E+00 | lm loss PPL: 5.884351E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38210/ 52452 | consumed samples: 19563520 | consumed tokens: 40066088960 | elapsed time per iteration (s): 9.16 | learning rate: 5.139E-05 | global batch size: 512 | lm loss: 1.219771E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.872 | TFLOPs: 8.39 | 31: iteration 38220/ 52452 | consumed samples: 19568640 | consumed tokens: 40076574720 | elapsed time per iteration (s): 1.76 | learning rate: 5.135E-05 | global batch size: 512 | lm loss: 1.205477E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.652 | TFLOPs: 43.78 | 31: iteration 38230/ 52452 | consumed samples: 19573760 | consumed tokens: 40087060480 | elapsed time per iteration (s): 1.74 | learning rate: 5.131E-05 | global batch size: 512 | lm loss: 1.209459E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.052 | TFLOPs: 44.29 | 31: iteration 38240/ 52452 | consumed samples: 19578880 | consumed tokens: 40097546240 | elapsed time per iteration (s): 1.78 | learning rate: 5.127E-05 | global batch size: 512 | lm loss: 1.210475E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.104 | TFLOPs: 43.24 | 31: iteration 38250/ 52452 | consumed samples: 19584000 | consumed tokens: 40108032000 | elapsed time per iteration (s): 1.77 | learning rate: 5.123E-05 | global batch size: 512 | lm loss: 1.192726E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.815 | TFLOPs: 43.35 | 31: iteration 38260/ 52452 | consumed samples: 19589120 | consumed tokens: 40118517760 | elapsed time per iteration (s): 1.91 | learning rate: 5.119E-05 | global batch size: 512 | lm loss: 1.201923E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.064 | TFLOPs: 40.23 | 31: iteration 38270/ 52452 | consumed samples: 19594240 | consumed tokens: 40129003520 | elapsed time per iteration (s): 1.84 | learning rate: 5.114E-05 | global batch size: 512 | lm loss: 1.206501E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.416 | TFLOPs: 41.79 | 31: iteration 38280/ 52452 | consumed samples: 19599360 | consumed tokens: 40139489280 | elapsed time per iteration (s): 1.74 | learning rate: 5.110E-05 | global batch size: 512 | lm loss: 1.205578E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.351 | TFLOPs: 44.18 | 31: iteration 38290/ 52452 | consumed samples: 19604480 | consumed tokens: 40149975040 | elapsed time per iteration (s): 1.76 | learning rate: 5.106E-05 | global batch size: 512 | lm loss: 1.197352E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.186 | TFLOPs: 43.56 | 31: iteration 38300/ 52452 | consumed samples: 19609600 | consumed tokens: 40160460800 | elapsed time per iteration (s): 1.77 | learning rate: 5.102E-05 | global batch size: 512 | lm loss: 1.201778E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.024 | TFLOPs: 43.38 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38300 | lm loss value: 4.295536E+00 | lm loss PPL: 7.337150E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38300 | lm loss value: 4.131341E+00 | lm loss PPL: 6.226136E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38310/ 52452 | consumed samples: 19614720 | consumed tokens: 40170946560 | elapsed time per iteration (s): 9.28 | learning rate: 5.098E-05 | global batch size: 512 | lm loss: 1.207405E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.191 | TFLOPs: 8.28 | 31: iteration 38320/ 52452 | consumed samples: 19619840 | consumed tokens: 40181432320 | elapsed time per iteration (s): 1.80 | learning rate: 5.094E-05 | global batch size: 512 | lm loss: 1.199835E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.605 | TFLOPs: 42.72 | 31: iteration 38330/ 52452 | consumed samples: 19624960 | consumed tokens: 40191918080 | elapsed time per iteration (s): 1.98 | learning rate: 5.090E-05 | global batch size: 512 | lm loss: 1.207263E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 258.221 | TFLOPs: 38.76 | 31: iteration 38340/ 52452 | consumed samples: 19630080 | consumed tokens: 40202403840 | elapsed time per iteration (s): 1.91 | learning rate: 5.086E-05 | global batch size: 512 | lm loss: 1.213762E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.384 | TFLOPs: 40.28 | 31: iteration 38350/ 52452 | consumed samples: 19635200 | consumed tokens: 40212889600 | elapsed time per iteration (s): 1.72 | learning rate: 5.082E-05 | global batch size: 512 | lm loss: 1.214142E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.417 | TFLOPs: 44.64 | 31: iteration 38360/ 52452 | consumed samples: 19640320 | consumed tokens: 40223375360 | elapsed time per iteration (s): 1.72 | learning rate: 5.078E-05 | global batch size: 512 | lm loss: 1.208896E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.481 | TFLOPs: 44.65 | 31: iteration 38370/ 52452 | consumed samples: 19645440 | consumed tokens: 40233861120 | elapsed time per iteration (s): 1.72 | learning rate: 5.073E-05 | global batch size: 512 | lm loss: 1.202807E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.937 | TFLOPs: 44.57 | 31: iteration 38380/ 52452 | consumed samples: 19650560 | consumed tokens: 40244346880 | elapsed time per iteration (s): 1.78 | learning rate: 5.069E-05 | global batch size: 512 | lm loss: 1.207816E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.136 | TFLOPs: 43.25 | 31: iteration 38390/ 52452 | consumed samples: 19655680 | consumed tokens: 40254832640 | elapsed time per iteration (s): 1.72 | learning rate: 5.065E-05 | global batch size: 512 | lm loss: 1.214661E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.443 | TFLOPs: 44.64 | 31: iteration 38400/ 52452 | consumed samples: 19660800 | consumed tokens: 40265318400 | elapsed time per iteration (s): 1.72 | learning rate: 5.061E-05 | global batch size: 512 | lm loss: 1.209731E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.518 | TFLOPs: 44.66 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38400 | lm loss value: 4.197764E+00 | lm loss PPL: 6.653741E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38400 | lm loss value: 4.074195E+00 | lm loss PPL: 5.880315E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38410/ 52452 | consumed samples: 19665920 | consumed tokens: 40275804160 | elapsed time per iteration (s): 9.18 | learning rate: 5.057E-05 | global batch size: 512 | lm loss: 1.190978E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.756 | TFLOPs: 8.37 | 31: iteration 38420/ 52452 | consumed samples: 19671040 | consumed tokens: 40286289920 | elapsed time per iteration (s): 1.74 | learning rate: 5.053E-05 | global batch size: 512 | lm loss: 1.211770E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.905 | TFLOPs: 44.11 | 31: iteration 38430/ 52452 | consumed samples: 19676160 | consumed tokens: 40296775680 | elapsed time per iteration (s): 1.73 | learning rate: 5.049E-05 | global batch size: 512 | lm loss: 1.207964E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.819 | TFLOPs: 44.40 | 31: iteration 38440/ 52452 | consumed samples: 19681280 | consumed tokens: 40307261440 | elapsed time per iteration (s): 1.75 | learning rate: 5.045E-05 | global batch size: 512 | lm loss: 1.193035E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.928 | TFLOPs: 43.97 | 31: iteration 38450/ 52452 | consumed samples: 19686400 | consumed tokens: 40317747200 | elapsed time per iteration (s): 1.75 | learning rate: 5.041E-05 | global batch size: 512 | lm loss: 1.205498E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.103 | TFLOPs: 43.99 | 31: iteration 38460/ 52452 | consumed samples: 19691520 | consumed tokens: 40328232960 | elapsed time per iteration (s): 1.74 | learning rate: 5.037E-05 | global batch size: 512 | lm loss: 1.205179E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.118 | TFLOPs: 44.15 | 31: iteration 38470/ 52452 | consumed samples: 19696640 | consumed tokens: 40338718720 | elapsed time per iteration (s): 1.75 | learning rate: 5.033E-05 | global batch size: 512 | lm loss: 1.191874E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.995 | TFLOPs: 43.98 | 31: iteration 38480/ 52452 | consumed samples: 19701760 | consumed tokens: 40349204480 | elapsed time per iteration (s): 1.73 | learning rate: 5.028E-05 | global batch size: 512 | lm loss: 1.210784E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.935 | TFLOPs: 44.42 | 31: iteration 38490/ 52452 | consumed samples: 19706880 | consumed tokens: 40359690240 | elapsed time per iteration (s): 1.74 | learning rate: 5.024E-05 | global batch size: 512 | lm loss: 1.197244E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.718 | TFLOPs: 44.09 | 31: iteration 38500/ 52452 | consumed samples: 19712000 | consumed tokens: 40370176000 | elapsed time per iteration (s): 1.75 | learning rate: 5.020E-05 | global batch size: 512 | lm loss: 1.204132E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.814 | TFLOPs: 43.80 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38500 | lm loss value: 4.250336E+00 | lm loss PPL: 7.012895E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38500 | lm loss value: 4.095313E+00 | lm loss PPL: 6.005811E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38510/ 52452 | consumed samples: 19717120 | consumed tokens: 40380661760 | elapsed time per iteration (s): 9.10 | learning rate: 5.016E-05 | global batch size: 512 | lm loss: 1.207453E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.258 | TFLOPs: 8.44 | 31: iteration 38520/ 52452 | consumed samples: 19722240 | consumed tokens: 40391147520 | elapsed time per iteration (s): 1.74 | learning rate: 5.012E-05 | global batch size: 512 | lm loss: 1.204327E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.795 | TFLOPs: 44.25 | 31: iteration 38530/ 52452 | consumed samples: 19727360 | consumed tokens: 40401633280 | elapsed time per iteration (s): 1.72 | learning rate: 5.008E-05 | global batch size: 512 | lm loss: 1.211353E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.931 | TFLOPs: 44.57 | 31: iteration 38540/ 52452 | consumed samples: 19732480 | consumed tokens: 40412119040 | elapsed time per iteration (s): 1.73 | learning rate: 5.004E-05 | global batch size: 512 | lm loss: 1.210678E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.353 | TFLOPs: 44.48 | 31: iteration 38550/ 52452 | consumed samples: 19737600 | consumed tokens: 40422604800 | elapsed time per iteration (s): 1.72 | learning rate: 5.000E-05 | global batch size: 512 | lm loss: 1.209555E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.429 | TFLOPs: 44.64 | 31: iteration 38560/ 52452 | consumed samples: 19742720 | consumed tokens: 40433090560 | elapsed time per iteration (s): 1.74 | learning rate: 4.996E-05 | global batch size: 512 | lm loss: 1.207925E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.003 | TFLOPs: 44.28 | 31: iteration 38570/ 52452 | consumed samples: 19747840 | consumed tokens: 40443576320 | elapsed time per iteration (s): 1.84 | learning rate: 4.992E-05 | global batch size: 512 | lm loss: 1.206738E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.330 | TFLOPs: 41.78 | 31: iteration 38580/ 52452 | consumed samples: 19752960 | consumed tokens: 40454062080 | elapsed time per iteration (s): 1.93 | learning rate: 4.988E-05 | global batch size: 512 | lm loss: 1.195258E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.548 | TFLOPs: 39.86 | 31: iteration 38590/ 52452 | consumed samples: 19758080 | consumed tokens: 40464547840 | elapsed time per iteration (s): 1.82 | learning rate: 4.984E-05 | global batch size: 512 | lm loss: 1.185243E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.456 | TFLOPs: 42.24 | 31: iteration 38600/ 52452 | consumed samples: 19763200 | consumed tokens: 40475033600 | elapsed time per iteration (s): 1.77 | learning rate: 4.980E-05 | global batch size: 512 | lm loss: 1.198975E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.509 | TFLOPs: 43.30 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38600 | lm loss value: 4.230671E+00 | lm loss PPL: 6.876338E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38600 | lm loss value: 4.092854E+00 | lm loss PPL: 5.991061E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38610/ 52452 | consumed samples: 19768320 | consumed tokens: 40485519360 | elapsed time per iteration (s): 9.16 | learning rate: 4.976E-05 | global batch size: 512 | lm loss: 1.203757E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.889 | TFLOPs: 8.39 | 31: iteration 38620/ 52452 | consumed samples: 19773440 | consumed tokens: 40496005120 | elapsed time per iteration (s): 1.75 | learning rate: 4.972E-05 | global batch size: 512 | lm loss: 1.216003E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.789 | TFLOPs: 43.95 | 31: iteration 38630/ 52452 | consumed samples: 19778560 | consumed tokens: 40506490880 | elapsed time per iteration (s): 1.75 | learning rate: 4.968E-05 | global batch size: 512 | lm loss: 1.205521E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.242 | TFLOPs: 44.01 | 31: iteration 38640/ 52452 | consumed samples: 19783680 | consumed tokens: 40516976640 | elapsed time per iteration (s): 1.74 | learning rate: 4.964E-05 | global batch size: 512 | lm loss: 1.221097E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.071 | TFLOPs: 44.29 | 31: iteration 38650/ 52452 | consumed samples: 19788800 | consumed tokens: 40527462400 | elapsed time per iteration (s): 1.76 | learning rate: 4.960E-05 | global batch size: 512 | lm loss: 1.204056E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.153 | TFLOPs: 43.70 | 31: iteration 38660/ 52452 | consumed samples: 19793920 | consumed tokens: 40537948160 | elapsed time per iteration (s): 1.73 | learning rate: 4.955E-05 | global batch size: 512 | lm loss: 1.201756E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.122 | TFLOPs: 44.45 | 31: iteration 38670/ 52452 | consumed samples: 19799040 | consumed tokens: 40548433920 | elapsed time per iteration (s): 1.73 | learning rate: 4.951E-05 | global batch size: 512 | lm loss: 1.209163E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.522 | TFLOPs: 44.51 | 31: iteration 38680/ 52452 | consumed samples: 19804160 | consumed tokens: 40558919680 | elapsed time per iteration (s): 1.72 | learning rate: 4.947E-05 | global batch size: 512 | lm loss: 1.207778E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.597 | TFLOPs: 44.67 | 31: iteration 38690/ 52452 | consumed samples: 19809280 | consumed tokens: 40569405440 | elapsed time per iteration (s): 1.74 | learning rate: 4.943E-05 | global batch size: 512 | lm loss: 1.194332E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.212 | TFLOPs: 44.16 | 31: iteration 38700/ 52452 | consumed samples: 19814400 | consumed tokens: 40579891200 | elapsed time per iteration (s): 1.75 | learning rate: 4.939E-05 | global batch size: 512 | lm loss: 1.200325E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.057 | TFLOPs: 43.99 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38700 | lm loss value: 4.205564E+00 | lm loss PPL: 6.705841E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38700 | lm loss value: 4.077169E+00 | lm loss PPL: 5.897826E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38710/ 52452 | consumed samples: 19819520 | consumed tokens: 40590376960 | elapsed time per iteration (s): 9.12 | learning rate: 4.935E-05 | global batch size: 512 | lm loss: 1.204791E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.140 | TFLOPs: 8.43 | 31: iteration 38720/ 52452 | consumed samples: 19824640 | consumed tokens: 40600862720 | elapsed time per iteration (s): 1.72 | learning rate: 4.931E-05 | global batch size: 512 | lm loss: 1.191976E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.449 | TFLOPs: 44.65 | 31: iteration 38730/ 52452 | consumed samples: 19829760 | consumed tokens: 40611348480 | elapsed time per iteration (s): 1.78 | learning rate: 4.927E-05 | global batch size: 512 | lm loss: 1.198628E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.904 | TFLOPs: 43.06 | 31: iteration 38740/ 52452 | consumed samples: 19834880 | consumed tokens: 40621834240 | elapsed time per iteration (s): 1.74 | learning rate: 4.923E-05 | global batch size: 512 | lm loss: 1.196887E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.217 | TFLOPs: 44.16 | 31: iteration 38750/ 52452 | consumed samples: 19840000 | consumed tokens: 40632320000 | elapsed time per iteration (s): 1.76 | learning rate: 4.919E-05 | global batch size: 512 | lm loss: 1.208245E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.651 | TFLOPs: 43.78 | 31: iteration 38760/ 52452 | consumed samples: 19845120 | consumed tokens: 40642805760 | elapsed time per iteration (s): 1.76 | learning rate: 4.915E-05 | global batch size: 512 | lm loss: 1.196346E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.849 | TFLOPs: 43.65 | 31: iteration 38770/ 52452 | consumed samples: 19850240 | consumed tokens: 40653291520 | elapsed time per iteration (s): 1.78 | learning rate: 4.911E-05 | global batch size: 512 | lm loss: 1.205757E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.698 | TFLOPs: 43.18 | 31: iteration 38780/ 52452 | consumed samples: 19855360 | consumed tokens: 40663777280 | elapsed time per iteration (s): 1.82 | learning rate: 4.907E-05 | global batch size: 512 | lm loss: 1.204732E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.947 | TFLOPs: 42.32 | 31: iteration 38790/ 52452 | consumed samples: 19860480 | consumed tokens: 40674263040 | elapsed time per iteration (s): 1.76 | learning rate: 4.903E-05 | global batch size: 512 | lm loss: 1.202941E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.165 | TFLOPs: 43.70 | 31: iteration 38800/ 52452 | consumed samples: 19865600 | consumed tokens: 40684748800 | elapsed time per iteration (s): 1.73 | learning rate: 4.899E-05 | global batch size: 512 | lm loss: 1.204143E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.155 | TFLOPs: 44.45 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38800 | lm loss value: 4.191340E+00 | lm loss PPL: 6.611135E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38800 | lm loss value: 4.060740E+00 | lm loss PPL: 5.801726E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38810/ 52452 | consumed samples: 19870720 | consumed tokens: 40695234560 | elapsed time per iteration (s): 9.15 | learning rate: 4.895E-05 | global batch size: 512 | lm loss: 1.203446E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.954 | TFLOPs: 8.40 | 31: iteration 38820/ 52452 | consumed samples: 19875840 | consumed tokens: 40705720320 | elapsed time per iteration (s): 1.80 | learning rate: 4.891E-05 | global batch size: 512 | lm loss: 1.205427E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.810 | TFLOPs: 42.60 | 31: iteration 38830/ 52452 | consumed samples: 19880960 | consumed tokens: 40716206080 | elapsed time per iteration (s): 1.75 | learning rate: 4.887E-05 | global batch size: 512 | lm loss: 1.208399E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.140 | TFLOPs: 44.00 | 31: iteration 38840/ 52452 | consumed samples: 19886080 | consumed tokens: 40726691840 | elapsed time per iteration (s): 1.73 | learning rate: 4.883E-05 | global batch size: 512 | lm loss: 1.212524E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.999 | TFLOPs: 44.43 | 31: iteration 38850/ 52452 | consumed samples: 19891200 | consumed tokens: 40737177600 | elapsed time per iteration (s): 1.74 | learning rate: 4.879E-05 | global batch size: 512 | lm loss: 1.204518E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.984 | TFLOPs: 44.28 | 31: iteration 38860/ 52452 | consumed samples: 19896320 | consumed tokens: 40747663360 | elapsed time per iteration (s): 1.78 | learning rate: 4.875E-05 | global batch size: 512 | lm loss: 1.200949E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.888 | TFLOPs: 43.21 | 31: iteration 38870/ 52452 | consumed samples: 19901440 | consumed tokens: 40758149120 | elapsed time per iteration (s): 1.74 | learning rate: 4.871E-05 | global batch size: 512 | lm loss: 1.200996E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.916 | TFLOPs: 44.12 | 31: iteration 38880/ 52452 | consumed samples: 19906560 | consumed tokens: 40768634880 | elapsed time per iteration (s): 1.73 | learning rate: 4.867E-05 | global batch size: 512 | lm loss: 1.198363E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.854 | TFLOPs: 44.41 | 31: iteration 38890/ 52452 | consumed samples: 19911680 | consumed tokens: 40779120640 | elapsed time per iteration (s): 1.77 | learning rate: 4.863E-05 | global batch size: 512 | lm loss: 1.207323E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.912 | TFLOPs: 43.36 | 31: iteration 38900/ 52452 | consumed samples: 19916800 | consumed tokens: 40789606400 | elapsed time per iteration (s): 1.78 | learning rate: 4.859E-05 | global batch size: 512 | lm loss: 1.191457E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.851 | TFLOPs: 43.20 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 38900 | lm loss value: 4.239337E+00 | lm loss PPL: 6.936188E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 38900 | lm loss value: 4.104941E+00 | lm loss PPL: 6.063919E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 38910/ 52452 | consumed samples: 19921920 | consumed tokens: 40800092160 | elapsed time per iteration (s): 9.13 | learning rate: 4.855E-05 | global batch size: 512 | lm loss: 1.210796E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.057 | TFLOPs: 8.41 | 31: iteration 38920/ 52452 | consumed samples: 19927040 | consumed tokens: 40810577920 | elapsed time per iteration (s): 1.74 | learning rate: 4.851E-05 | global batch size: 512 | lm loss: 1.187973E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.963 | TFLOPs: 44.12 | 31: iteration 38930/ 52452 | consumed samples: 19932160 | consumed tokens: 40821063680 | elapsed time per iteration (s): 1.75 | learning rate: 4.847E-05 | global batch size: 512 | lm loss: 1.202230E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.857 | TFLOPs: 43.81 | 31: iteration 38940/ 52452 | consumed samples: 19937280 | consumed tokens: 40831549440 | elapsed time per iteration (s): 1.77 | learning rate: 4.843E-05 | global batch size: 512 | lm loss: 1.186424E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.974 | TFLOPs: 43.37 | 31: iteration 38950/ 52452 | consumed samples: 19942400 | consumed tokens: 40842035200 | elapsed time per iteration (s): 1.74 | learning rate: 4.839E-05 | global batch size: 512 | lm loss: 1.202866E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.434 | TFLOPs: 44.19 | 31: iteration 38960/ 52452 | consumed samples: 19947520 | consumed tokens: 40852520960 | elapsed time per iteration (s): 1.76 | learning rate: 4.835E-05 | global batch size: 512 | lm loss: 1.206190E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.734 | TFLOPs: 43.64 | 31: iteration 38970/ 52452 | consumed samples: 19952640 | consumed tokens: 40863006720 | elapsed time per iteration (s): 1.76 | learning rate: 4.832E-05 | global batch size: 512 | lm loss: 1.193615E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.936 | TFLOPs: 43.67 | 31: iteration 38980/ 52452 | consumed samples: 19957760 | consumed tokens: 40873492480 | elapsed time per iteration (s): 1.73 | learning rate: 4.828E-05 | global batch size: 512 | lm loss: 1.189838E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.765 | TFLOPs: 44.39 | 31: iteration 38990/ 52452 | consumed samples: 19962880 | consumed tokens: 40883978240 | elapsed time per iteration (s): 1.81 | learning rate: 4.824E-05 | global batch size: 512 | lm loss: 1.196786E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.702 | TFLOPs: 42.43 | 31: iteration 39000/ 52452 | consumed samples: 19968000 | consumed tokens: 40894464000 | elapsed time per iteration (s): 1.81 | learning rate: 4.820E-05 | global batch size: 512 | lm loss: 1.192819E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.343 | TFLOPs: 42.53 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39000 | lm loss value: 4.098819E+00 | lm loss PPL: 6.026905E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39000 | lm loss value: 4.010356E+00 | lm loss PPL: 5.516650E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39010/ 52452 | consumed samples: 19973120 | consumed tokens: 40904949760 | elapsed time per iteration (s): 9.16 | learning rate: 4.816E-05 | global batch size: 512 | lm loss: 1.201168E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.876 | TFLOPs: 8.39 | 31: iteration 39020/ 52452 | consumed samples: 19978240 | consumed tokens: 40915435520 | elapsed time per iteration (s): 1.75 | learning rate: 4.812E-05 | global batch size: 512 | lm loss: 1.200737E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.056 | TFLOPs: 43.84 | 31: iteration 39030/ 52452 | consumed samples: 19983360 | consumed tokens: 40925921280 | elapsed time per iteration (s): 1.78 | learning rate: 4.808E-05 | global batch size: 512 | lm loss: 1.210842E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.381 | TFLOPs: 43.13 | 31: iteration 39040/ 52452 | consumed samples: 19988480 | consumed tokens: 40936407040 | elapsed time per iteration (s): 1.73 | learning rate: 4.804E-05 | global batch size: 512 | lm loss: 1.203156E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.749 | TFLOPs: 44.39 | 31: iteration 39050/ 52452 | consumed samples: 19993600 | consumed tokens: 40946892800 | elapsed time per iteration (s): 1.73 | learning rate: 4.800E-05 | global batch size: 512 | lm loss: 1.203667E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.728 | TFLOPs: 44.39 | 31: iteration 39060/ 52452 | consumed samples: 19998720 | consumed tokens: 40957378560 | elapsed time per iteration (s): 1.79 | learning rate: 4.796E-05 | global batch size: 512 | lm loss: 1.197046E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.325 | TFLOPs: 42.83 | 31: iteration 39070/ 52452 | consumed samples: 20003840 | consumed tokens: 40967864320 | elapsed time per iteration (s): 1.78 | learning rate: 4.792E-05 | global batch size: 512 | lm loss: 1.205726E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.861 | TFLOPs: 43.21 | 31: iteration 39080/ 52452 | consumed samples: 20008960 | consumed tokens: 40978350080 | elapsed time per iteration (s): 1.79 | learning rate: 4.788E-05 | global batch size: 512 | lm loss: 1.199174E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.549 | TFLOPs: 42.86 | 31: iteration 39090/ 52452 | consumed samples: 20014080 | consumed tokens: 40988835840 | elapsed time per iteration (s): 1.76 | learning rate: 4.784E-05 | global batch size: 512 | lm loss: 1.194274E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.977 | TFLOPs: 43.67 | 31: iteration 39100/ 52452 | consumed samples: 20019200 | consumed tokens: 40999321600 | elapsed time per iteration (s): 1.77 | learning rate: 4.780E-05 | global batch size: 512 | lm loss: 1.214173E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.797 | TFLOPs: 43.35 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39100 | lm loss value: 4.176131E+00 | lm loss PPL: 6.511343E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39100 | lm loss value: 4.048439E+00 | lm loss PPL: 5.730793E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39110/ 52452 | consumed samples: 20024320 | consumed tokens: 41009807360 | elapsed time per iteration (s): 9.14 | learning rate: 4.776E-05 | global batch size: 512 | lm loss: 1.197048E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.028 | TFLOPs: 8.41 | 31: iteration 39120/ 52452 | consumed samples: 20029440 | consumed tokens: 41020293120 | elapsed time per iteration (s): 1.76 | learning rate: 4.772E-05 | global batch size: 512 | lm loss: 1.198242E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.251 | TFLOPs: 43.57 | 31: iteration 39130/ 52452 | consumed samples: 20034560 | consumed tokens: 41030778880 | elapsed time per iteration (s): 1.81 | learning rate: 4.768E-05 | global batch size: 512 | lm loss: 1.204129E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.151 | TFLOPs: 42.50 | 31: iteration 39140/ 52452 | consumed samples: 20039680 | consumed tokens: 41041264640 | elapsed time per iteration (s): 1.76 | learning rate: 4.764E-05 | global batch size: 512 | lm loss: 1.207274E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.052 | TFLOPs: 43.69 | 31: iteration 39150/ 52452 | consumed samples: 20044800 | consumed tokens: 41051750400 | elapsed time per iteration (s): 1.77 | learning rate: 4.761E-05 | global batch size: 512 | lm loss: 1.195827E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.580 | TFLOPs: 43.31 | 31: iteration 39160/ 52452 | consumed samples: 20049920 | consumed tokens: 41062236160 | elapsed time per iteration (s): 1.75 | learning rate: 4.757E-05 | global batch size: 512 | lm loss: 1.206375E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.260 | TFLOPs: 44.02 | 31: iteration 39170/ 52452 | consumed samples: 20055040 | consumed tokens: 41072721920 | elapsed time per iteration (s): 1.75 | learning rate: 4.753E-05 | global batch size: 512 | lm loss: 1.198456E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.034 | TFLOPs: 43.83 | 31: iteration 39180/ 52452 | consumed samples: 20060160 | consumed tokens: 41083207680 | elapsed time per iteration (s): 1.80 | learning rate: 4.749E-05 | global batch size: 512 | lm loss: 1.190897E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.693 | TFLOPs: 42.58 | 31: iteration 39190/ 52452 | consumed samples: 20065280 | consumed tokens: 41093693440 | elapsed time per iteration (s): 1.74 | learning rate: 4.745E-05 | global batch size: 512 | lm loss: 1.203267E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.875 | TFLOPs: 44.26 | 31: iteration 39200/ 52452 | consumed samples: 20070400 | consumed tokens: 41104179200 | elapsed time per iteration (s): 1.75 | learning rate: 4.741E-05 | global batch size: 512 | lm loss: 1.189597E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.651 | TFLOPs: 43.93 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39200 | lm loss value: 4.203640E+00 | lm loss PPL: 6.692951E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39200 | lm loss value: 4.056958E+00 | lm loss PPL: 5.779821E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39210/ 52452 | consumed samples: 20075520 | consumed tokens: 41114664960 | elapsed time per iteration (s): 9.18 | learning rate: 4.737E-05 | global batch size: 512 | lm loss: 1.206132E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.743 | TFLOPs: 8.37 | 31: iteration 39220/ 52452 | consumed samples: 20080640 | consumed tokens: 41125150720 | elapsed time per iteration (s): 1.76 | learning rate: 4.733E-05 | global batch size: 512 | lm loss: 1.200347E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.174 | TFLOPs: 43.55 | 31: iteration 39230/ 52452 | consumed samples: 20085760 | consumed tokens: 41135636480 | elapsed time per iteration (s): 1.75 | learning rate: 4.729E-05 | global batch size: 512 | lm loss: 1.193023E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.007 | TFLOPs: 43.83 | 31: iteration 39240/ 52452 | consumed samples: 20090880 | consumed tokens: 41146122240 | elapsed time per iteration (s): 1.77 | learning rate: 4.725E-05 | global batch size: 512 | lm loss: 1.201034E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.094 | TFLOPs: 43.39 | 31: iteration 39250/ 52452 | consumed samples: 20096000 | consumed tokens: 41156608000 | elapsed time per iteration (s): 1.79 | learning rate: 4.721E-05 | global batch size: 512 | lm loss: 1.205550E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.616 | TFLOPs: 42.87 | 31: iteration 39260/ 52452 | consumed samples: 20101120 | consumed tokens: 41167093760 | elapsed time per iteration (s): 1.76 | learning rate: 4.717E-05 | global batch size: 512 | lm loss: 1.194567E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.371 | TFLOPs: 43.73 | 31: iteration 39270/ 52452 | consumed samples: 20106240 | consumed tokens: 41177579520 | elapsed time per iteration (s): 1.74 | learning rate: 4.714E-05 | global batch size: 512 | lm loss: 1.201642E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.700 | TFLOPs: 44.08 | 31: iteration 39280/ 52452 | consumed samples: 20111360 | consumed tokens: 41188065280 | elapsed time per iteration (s): 1.75 | learning rate: 4.710E-05 | global batch size: 512 | lm loss: 1.206723E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.003 | TFLOPs: 43.83 | 31: iteration 39290/ 52452 | consumed samples: 20116480 | consumed tokens: 41198551040 | elapsed time per iteration (s): 1.73 | learning rate: 4.706E-05 | global batch size: 512 | lm loss: 1.204184E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.678 | TFLOPs: 44.53 | 31: iteration 39300/ 52452 | consumed samples: 20121600 | consumed tokens: 41209036800 | elapsed time per iteration (s): 1.76 | learning rate: 4.702E-05 | global batch size: 512 | lm loss: 1.195028E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.560 | TFLOPs: 43.61 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39300 | lm loss value: 4.180476E+00 | lm loss PPL: 6.539699E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39300 | lm loss value: 4.050705E+00 | lm loss PPL: 5.743796E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39310/ 52452 | consumed samples: 20126720 | consumed tokens: 41219522560 | elapsed time per iteration (s): 9.23 | learning rate: 4.698E-05 | global batch size: 512 | lm loss: 1.198538E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.459 | TFLOPs: 8.32 | 31: iteration 39320/ 52452 | consumed samples: 20131840 | consumed tokens: 41230008320 | elapsed time per iteration (s): 1.82 | learning rate: 4.694E-05 | global batch size: 512 | lm loss: 1.206855E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.349 | TFLOPs: 42.23 | 31: iteration 39330/ 52452 | consumed samples: 20136960 | consumed tokens: 41240494080 | elapsed time per iteration (s): 1.74 | learning rate: 4.690E-05 | global batch size: 512 | lm loss: 1.190557E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.448 | TFLOPs: 44.04 | 31: iteration 39340/ 52452 | consumed samples: 20142080 | consumed tokens: 41250979840 | elapsed time per iteration (s): 1.73 | learning rate: 4.686E-05 | global batch size: 512 | lm loss: 1.192626E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.102 | TFLOPs: 44.44 | 31: iteration 39350/ 52452 | consumed samples: 20147200 | consumed tokens: 41261465600 | elapsed time per iteration (s): 1.78 | learning rate: 4.682E-05 | global batch size: 512 | lm loss: 1.203385E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.137 | TFLOPs: 43.10 | 31: iteration 39360/ 52452 | consumed samples: 20152320 | consumed tokens: 41271951360 | elapsed time per iteration (s): 1.74 | learning rate: 4.679E-05 | global batch size: 512 | lm loss: 1.200084E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.953 | TFLOPs: 44.12 | 31: iteration 39370/ 52452 | consumed samples: 20157440 | consumed tokens: 41282437120 | elapsed time per iteration (s): 1.80 | learning rate: 4.675E-05 | global batch size: 512 | lm loss: 1.198009E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.867 | TFLOPs: 42.76 | 31: iteration 39380/ 52452 | consumed samples: 20162560 | consumed tokens: 41292922880 | elapsed time per iteration (s): 1.82 | learning rate: 4.671E-05 | global batch size: 512 | lm loss: 1.190238E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.767 | TFLOPs: 42.14 | 31: iteration 39390/ 52452 | consumed samples: 20167680 | consumed tokens: 41303408640 | elapsed time per iteration (s): 1.80 | learning rate: 4.667E-05 | global batch size: 512 | lm loss: 1.203589E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.988 | TFLOPs: 42.78 | 31: iteration 39400/ 52452 | consumed samples: 20172800 | consumed tokens: 41313894400 | elapsed time per iteration (s): 1.75 | learning rate: 4.663E-05 | global batch size: 512 | lm loss: 1.194907E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.087 | TFLOPs: 43.99 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39400 | lm loss value: 4.242411E+00 | lm loss PPL: 6.957537E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39400 | lm loss value: 4.101673E+00 | lm loss PPL: 6.044130E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39410/ 52452 | consumed samples: 20177920 | consumed tokens: 41324380160 | elapsed time per iteration (s): 9.19 | learning rate: 4.659E-05 | global batch size: 512 | lm loss: 1.204947E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.692 | TFLOPs: 8.36 | 31: iteration 39420/ 52452 | consumed samples: 20183040 | consumed tokens: 41334865920 | elapsed time per iteration (s): 1.75 | learning rate: 4.655E-05 | global batch size: 512 | lm loss: 1.203500E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.083 | TFLOPs: 43.99 | 31: iteration 39430/ 52452 | consumed samples: 20188160 | consumed tokens: 41345351680 | elapsed time per iteration (s): 1.74 | learning rate: 4.652E-05 | global batch size: 512 | lm loss: 1.198867E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.266 | TFLOPs: 44.17 | 31: iteration 39440/ 52452 | consumed samples: 20193280 | consumed tokens: 41355837440 | elapsed time per iteration (s): 1.76 | learning rate: 4.648E-05 | global batch size: 512 | lm loss: 1.192424E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.706 | TFLOPs: 43.63 | 31: iteration 39450/ 52452 | consumed samples: 20198400 | consumed tokens: 41366323200 | elapsed time per iteration (s): 1.77 | learning rate: 4.644E-05 | global batch size: 512 | lm loss: 1.202809E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.560 | TFLOPs: 43.46 | 31: iteration 39460/ 52452 | consumed samples: 20203520 | consumed tokens: 41376808960 | elapsed time per iteration (s): 1.76 | learning rate: 4.640E-05 | global batch size: 512 | lm loss: 1.196969E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.082 | TFLOPs: 43.69 | 31: iteration 39470/ 52452 | consumed samples: 20208640 | consumed tokens: 41387294720 | elapsed time per iteration (s): 1.75 | learning rate: 4.636E-05 | global batch size: 512 | lm loss: 1.195641E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.993 | TFLOPs: 43.98 | 31: iteration 39480/ 52452 | consumed samples: 20213760 | consumed tokens: 41397780480 | elapsed time per iteration (s): 1.75 | learning rate: 4.632E-05 | global batch size: 512 | lm loss: 1.194854E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.151 | TFLOPs: 43.85 | 31: iteration 39490/ 52452 | consumed samples: 20218880 | consumed tokens: 41408266240 | elapsed time per iteration (s): 1.77 | learning rate: 4.628E-05 | global batch size: 512 | lm loss: 1.197774E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.106 | TFLOPs: 43.39 | 31: iteration 39500/ 52452 | consumed samples: 20224000 | consumed tokens: 41418752000 | elapsed time per iteration (s): 1.73 | learning rate: 4.625E-05 | global batch size: 512 | lm loss: 1.186745E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.931 | TFLOPs: 44.42 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39500 | lm loss value: 4.192023E+00 | lm loss PPL: 6.615651E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39500 | lm loss value: 4.061806E+00 | lm loss PPL: 5.807909E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39510/ 52452 | consumed samples: 20229120 | consumed tokens: 41429237760 | elapsed time per iteration (s): 9.14 | learning rate: 4.621E-05 | global batch size: 512 | lm loss: 1.189899E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.031 | TFLOPs: 8.41 | 31: iteration 39520/ 52452 | consumed samples: 20234240 | consumed tokens: 41439723520 | elapsed time per iteration (s): 1.75 | learning rate: 4.617E-05 | global batch size: 512 | lm loss: 1.203016E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.810 | TFLOPs: 43.95 | 31: iteration 39530/ 52452 | consumed samples: 20239360 | consumed tokens: 41450209280 | elapsed time per iteration (s): 1.78 | learning rate: 4.613E-05 | global batch size: 512 | lm loss: 1.181612E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.059 | TFLOPs: 43.24 | 31: iteration 39540/ 52452 | consumed samples: 20244480 | consumed tokens: 41460695040 | elapsed time per iteration (s): 1.72 | learning rate: 4.609E-05 | global batch size: 512 | lm loss: 1.186921E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.857 | TFLOPs: 44.56 | 31: iteration 39550/ 52452 | consumed samples: 20249600 | consumed tokens: 41471180800 | elapsed time per iteration (s): 1.77 | learning rate: 4.605E-05 | global batch size: 512 | lm loss: 1.210061E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.137 | TFLOPs: 43.40 | 31: iteration 39560/ 52452 | consumed samples: 20254720 | consumed tokens: 41481666560 | elapsed time per iteration (s): 1.79 | learning rate: 4.602E-05 | global batch size: 512 | lm loss: 1.197157E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.506 | TFLOPs: 42.85 | 31: iteration 39570/ 52452 | consumed samples: 20259840 | consumed tokens: 41492152320 | elapsed time per iteration (s): 1.80 | learning rate: 4.598E-05 | global batch size: 512 | lm loss: 1.198296E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.612 | TFLOPs: 42.72 | 31: iteration 39580/ 52452 | consumed samples: 20264960 | consumed tokens: 41502638080 | elapsed time per iteration (s): 1.75 | learning rate: 4.594E-05 | global batch size: 512 | lm loss: 1.207515E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.888 | TFLOPs: 43.81 | 31: iteration 39590/ 52452 | consumed samples: 20270080 | consumed tokens: 41513123840 | elapsed time per iteration (s): 1.82 | learning rate: 4.590E-05 | global batch size: 512 | lm loss: 1.204593E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.726 | TFLOPs: 42.29 | 31: iteration 39600/ 52452 | consumed samples: 20275200 | consumed tokens: 41523609600 | elapsed time per iteration (s): 1.75 | learning rate: 4.586E-05 | global batch size: 512 | lm loss: 1.198766E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.761 | TFLOPs: 43.94 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39600 | lm loss value: 4.306913E+00 | lm loss PPL: 7.421107E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39600 | lm loss value: 4.132548E+00 | lm loss PPL: 6.233658E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39610/ 52452 | consumed samples: 20280320 | consumed tokens: 41534095360 | elapsed time per iteration (s): 9.15 | learning rate: 4.582E-05 | global batch size: 512 | lm loss: 1.191272E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.950 | TFLOPs: 8.40 | 31: iteration 39620/ 52452 | consumed samples: 20285440 | consumed tokens: 41544581120 | elapsed time per iteration (s): 1.72 | learning rate: 4.579E-05 | global batch size: 512 | lm loss: 1.192418E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.981 | TFLOPs: 44.58 | 31: iteration 39630/ 52452 | consumed samples: 20290560 | consumed tokens: 41555066880 | elapsed time per iteration (s): 1.76 | learning rate: 4.575E-05 | global batch size: 512 | lm loss: 1.190536E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.267 | TFLOPs: 43.72 | 31: iteration 39640/ 52452 | consumed samples: 20295680 | consumed tokens: 41565552640 | elapsed time per iteration (s): 1.75 | learning rate: 4.571E-05 | global batch size: 512 | lm loss: 1.190373E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.724 | TFLOPs: 43.94 | 31: iteration 39650/ 52452 | consumed samples: 20300800 | consumed tokens: 41576038400 | elapsed time per iteration (s): 1.79 | learning rate: 4.567E-05 | global batch size: 512 | lm loss: 1.193517E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.592 | TFLOPs: 42.87 | 31: iteration 39660/ 52452 | consumed samples: 20305920 | consumed tokens: 41586524160 | elapsed time per iteration (s): 1.81 | learning rate: 4.563E-05 | global batch size: 512 | lm loss: 1.189346E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.088 | TFLOPs: 42.49 | 31: iteration 39670/ 52452 | consumed samples: 20311040 | consumed tokens: 41597009920 | elapsed time per iteration (s): 1.80 | learning rate: 4.560E-05 | global batch size: 512 | lm loss: 1.182404E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.671 | TFLOPs: 42.58 | 31: iteration 39680/ 52452 | consumed samples: 20316160 | consumed tokens: 41607495680 | elapsed time per iteration (s): 1.77 | learning rate: 4.556E-05 | global batch size: 512 | lm loss: 1.184692E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.486 | TFLOPs: 43.30 | 31: iteration 39690/ 52452 | consumed samples: 20321280 | consumed tokens: 41617981440 | elapsed time per iteration (s): 1.77 | learning rate: 4.552E-05 | global batch size: 512 | lm loss: 1.195156E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.665 | TFLOPs: 43.48 | 31: iteration 39700/ 52452 | consumed samples: 20326400 | consumed tokens: 41628467200 | elapsed time per iteration (s): 1.81 | learning rate: 4.548E-05 | global batch size: 512 | lm loss: 1.195933E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.654 | TFLOPs: 42.42 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39700 | lm loss value: 4.160419E+00 | lm loss PPL: 6.409837E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39700 | lm loss value: 4.026489E+00 | lm loss PPL: 5.606374E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39710/ 52452 | consumed samples: 20331520 | consumed tokens: 41638952960 | elapsed time per iteration (s): 9.24 | learning rate: 4.544E-05 | global batch size: 512 | lm loss: 1.197406E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.392 | TFLOPs: 8.31 | 31: iteration 39720/ 52452 | consumed samples: 20336640 | consumed tokens: 41649438720 | elapsed time per iteration (s): 1.78 | learning rate: 4.541E-05 | global batch size: 512 | lm loss: 1.193038E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.562 | TFLOPs: 43.16 | 31: iteration 39730/ 52452 | consumed samples: 20341760 | consumed tokens: 41659924480 | elapsed time per iteration (s): 1.81 | learning rate: 4.537E-05 | global batch size: 512 | lm loss: 1.183030E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.019 | TFLOPs: 42.48 | 31: iteration 39740/ 52452 | consumed samples: 20346880 | consumed tokens: 41670410240 | elapsed time per iteration (s): 1.79 | learning rate: 4.533E-05 | global batch size: 512 | lm loss: 1.190820E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.252 | TFLOPs: 42.96 | 31: iteration 39750/ 52452 | consumed samples: 20352000 | consumed tokens: 41680896000 | elapsed time per iteration (s): 1.75 | learning rate: 4.529E-05 | global batch size: 512 | lm loss: 1.190533E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.462 | TFLOPs: 43.90 | 31: iteration 39760/ 52452 | consumed samples: 20357120 | consumed tokens: 41691381760 | elapsed time per iteration (s): 1.75 | learning rate: 4.525E-05 | global batch size: 512 | lm loss: 1.197543E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.384 | TFLOPs: 43.89 | 31: iteration 39770/ 52452 | consumed samples: 20362240 | consumed tokens: 41701867520 | elapsed time per iteration (s): 1.74 | learning rate: 4.522E-05 | global batch size: 512 | lm loss: 1.204287E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.926 | TFLOPs: 44.27 | 31: iteration 39780/ 52452 | consumed samples: 20367360 | consumed tokens: 41712353280 | elapsed time per iteration (s): 1.76 | learning rate: 4.518E-05 | global batch size: 512 | lm loss: 1.193883E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.389 | TFLOPs: 43.59 | 31: iteration 39790/ 52452 | consumed samples: 20372480 | consumed tokens: 41722839040 | elapsed time per iteration (s): 1.76 | learning rate: 4.514E-05 | global batch size: 512 | lm loss: 1.192904E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.929 | TFLOPs: 43.67 | 31: iteration 39800/ 52452 | consumed samples: 20377600 | consumed tokens: 41733324800 | elapsed time per iteration (s): 1.77 | learning rate: 4.510E-05 | global batch size: 512 | lm loss: 1.188858E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.836 | TFLOPs: 43.50 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39800 | lm loss value: 4.194746E+00 | lm loss PPL: 6.633691E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39800 | lm loss value: 4.058216E+00 | lm loss PPL: 5.787096E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39810/ 52452 | consumed samples: 20382720 | consumed tokens: 41743810560 | elapsed time per iteration (s): 9.15 | learning rate: 4.507E-05 | global batch size: 512 | lm loss: 1.194945E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.985 | TFLOPs: 8.40 | 31: iteration 39820/ 52452 | consumed samples: 20387840 | consumed tokens: 41754296320 | elapsed time per iteration (s): 1.73 | learning rate: 4.503E-05 | global batch size: 512 | lm loss: 1.181431E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.219 | TFLOPs: 44.31 | 31: iteration 39830/ 52452 | consumed samples: 20392960 | consumed tokens: 41764782080 | elapsed time per iteration (s): 1.74 | learning rate: 4.499E-05 | global batch size: 512 | lm loss: 1.203846E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.382 | TFLOPs: 44.19 | 31: iteration 39840/ 52452 | consumed samples: 20398080 | consumed tokens: 41775267840 | elapsed time per iteration (s): 1.75 | learning rate: 4.495E-05 | global batch size: 512 | lm loss: 1.195134E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.333 | TFLOPs: 43.88 | 31: iteration 39850/ 52452 | consumed samples: 20403200 | consumed tokens: 41785753600 | elapsed time per iteration (s): 1.74 | learning rate: 4.492E-05 | global batch size: 512 | lm loss: 1.197408E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.013 | TFLOPs: 44.13 | 31: iteration 39860/ 52452 | consumed samples: 20408320 | consumed tokens: 41796239360 | elapsed time per iteration (s): 1.74 | learning rate: 4.488E-05 | global batch size: 512 | lm loss: 1.197064E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.684 | TFLOPs: 44.23 | 31: iteration 39870/ 52452 | consumed samples: 20413440 | consumed tokens: 41806725120 | elapsed time per iteration (s): 1.73 | learning rate: 4.484E-05 | global batch size: 512 | lm loss: 1.190742E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.722 | TFLOPs: 44.39 | 31: iteration 39880/ 52452 | consumed samples: 20418560 | consumed tokens: 41817210880 | elapsed time per iteration (s): 1.73 | learning rate: 4.480E-05 | global batch size: 512 | lm loss: 1.186927E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.733 | TFLOPs: 44.39 | 31: iteration 39890/ 52452 | consumed samples: 20423680 | consumed tokens: 41827696640 | elapsed time per iteration (s): 1.77 | learning rate: 4.476E-05 | global batch size: 512 | lm loss: 1.197954E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.518 | TFLOPs: 43.46 | 31: iteration 39900/ 52452 | consumed samples: 20428800 | consumed tokens: 41838182400 | elapsed time per iteration (s): 1.73 | learning rate: 4.473E-05 | global batch size: 512 | lm loss: 1.182372E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.103 | TFLOPs: 44.29 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 39900 | lm loss value: 4.095256E+00 | lm loss PPL: 6.005470E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 39900 | lm loss value: 4.000943E+00 | lm loss PPL: 5.464964E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 39910/ 52452 | consumed samples: 20433920 | consumed tokens: 41848668160 | elapsed time per iteration (s): 9.21 | learning rate: 4.469E-05 | global batch size: 512 | lm loss: 1.190307E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.577 | TFLOPs: 8.34 | 31: iteration 39920/ 52452 | consumed samples: 20439040 | consumed tokens: 41859153920 | elapsed time per iteration (s): 1.76 | learning rate: 4.465E-05 | global batch size: 512 | lm loss: 1.200956E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.865 | TFLOPs: 43.66 | 31: iteration 39930/ 52452 | consumed samples: 20444160 | consumed tokens: 41869639680 | elapsed time per iteration (s): 1.75 | learning rate: 4.461E-05 | global batch size: 512 | lm loss: 1.195489E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.396 | TFLOPs: 44.04 | 31: iteration 39940/ 52452 | consumed samples: 20449280 | consumed tokens: 41880125440 | elapsed time per iteration (s): 1.74 | learning rate: 4.458E-05 | global batch size: 512 | lm loss: 1.183435E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.860 | TFLOPs: 44.11 | 31: iteration 39950/ 52452 | consumed samples: 20454400 | consumed tokens: 41890611200 | elapsed time per iteration (s): 1.74 | learning rate: 4.454E-05 | global batch size: 512 | lm loss: 1.188000E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.577 | TFLOPs: 44.06 | 31: iteration 39960/ 52452 | consumed samples: 20459520 | consumed tokens: 41901096960 | elapsed time per iteration (s): 1.75 | learning rate: 4.450E-05 | global batch size: 512 | lm loss: 1.187574E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.866 | TFLOPs: 43.81 | 31: iteration 39970/ 52452 | consumed samples: 20464640 | consumed tokens: 41911582720 | elapsed time per iteration (s): 1.79 | learning rate: 4.447E-05 | global batch size: 512 | lm loss: 1.200600E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.918 | TFLOPs: 42.91 | 31: iteration 39980/ 52452 | consumed samples: 20469760 | consumed tokens: 41922068480 | elapsed time per iteration (s): 1.78 | learning rate: 4.443E-05 | global batch size: 512 | lm loss: 1.193034E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.481 | TFLOPs: 43.15 | 31: iteration 39990/ 52452 | consumed samples: 20474880 | consumed tokens: 41932554240 | elapsed time per iteration (s): 1.77 | learning rate: 4.439E-05 | global batch size: 512 | lm loss: 1.187456E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.618 | TFLOPs: 43.47 | 0: [2022-12-26 13:07:46,717] [INFO] [logging.py:68:log_dist] [Rank 0] step=40000, skipped=0, lr=[4.4353595346482854e-05, 4.4353595346482854e-05, 4.4353595346482854e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 40000/ 52452 | consumed samples: 20480000 | consumed tokens: 41943040000 | elapsed time per iteration (s): 1.75 | learning rate: 4.435E-05 | global batch size: 512 | lm loss: 1.196408E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.632 | TFLOPs: 43.92 | 0: steps: 40000 loss: 1.1852 iter time (s): 1.767 samples/sec: 289.717 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40000 | lm loss value: 4.260821E+00 | lm loss PPL: 7.086817E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40000 | lm loss value: 4.098565E+00 | lm loss PPL: 6.025374E+01 | 31: ------------------------------------------------------------------------------------------------------ 0: saving checkpoint at iteration 40000 to checkpoints_2b855brealtasky 0: [2022-12-26 13:09:00,667] [INFO] [logging.py:68:log_dist] [Rank 0] [Torch] Checkpoint global_step40000 is begin to save! 0: [2022-12-26 13:09:00,675] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_01-model_00-model_states.pt... 0: [2022-12-26 13:09:01,056] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_01-model_00-model_states.pt. 0: [2022-12-26 13:09:01,056] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_03-model_00-model_states.pt... 0: [2022-12-26 13:09:01,240] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_03-model_00-model_states.pt. 0: [2022-12-26 13:09:01,241] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_04-model_00-model_states.pt... 0: [2022-12-26 13:09:01,421] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_04-model_00-model_states.pt. 0: [2022-12-26 13:09:01,421] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_05-model_00-model_states.pt... 0: [2022-12-26 13:09:01,603] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_05-model_00-model_states.pt. 0: [2022-12-26 13:09:01,604] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_06-model_00-model_states.pt... 0: [2022-12-26 13:09:01,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_06-model_00-model_states.pt. 0: [2022-12-26 13:09:01,791] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_07-model_00-model_states.pt... 0: [2022-12-26 13:09:01,968] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_07-model_00-model_states.pt. 0: [2022-12-26 13:09:01,969] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_08-model_00-model_states.pt... 0: [2022-12-26 13:09:02,152] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_08-model_00-model_states.pt. 0: [2022-12-26 13:09:02,153] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_09-model_00-model_states.pt... 0: [2022-12-26 13:09:02,331] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_09-model_00-model_states.pt. 0: [2022-12-26 13:09:02,332] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_10-model_00-model_states.pt... 0: [2022-12-26 13:09:02,516] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_10-model_00-model_states.pt. 0: [2022-12-26 13:09:02,516] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_11-model_00-model_states.pt... 0: [2022-12-26 13:09:02,693] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_11-model_00-model_states.pt. 0: [2022-12-26 13:09:02,693] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_12-model_00-model_states.pt... 0: [2022-12-26 13:09:02,872] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_12-model_00-model_states.pt. 0: [2022-12-26 13:09:02,873] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_13-model_00-model_states.pt... 0: [2022-12-26 13:09:03,055] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_13-model_00-model_states.pt. 0: [2022-12-26 13:09:03,055] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_14-model_00-model_states.pt... 0: [2022-12-26 13:09:03,235] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_14-model_00-model_states.pt. 0: [2022-12-26 13:09:03,235] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_15-model_00-model_states.pt... 0: [2022-12-26 13:09:03,409] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_15-model_00-model_states.pt. 0: [2022-12-26 13:09:03,409] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_16-model_00-model_states.pt... 0: [2022-12-26 13:09:03,585] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_16-model_00-model_states.pt. 0: [2022-12-26 13:09:03,586] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_17-model_00-model_states.pt... 0: [2022-12-26 13:09:03,764] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_17-model_00-model_states.pt. 0: [2022-12-26 13:09:03,764] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_18-model_00-model_states.pt... 0: [2022-12-26 13:09:03,939] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_18-model_00-model_states.pt. 0: [2022-12-26 13:09:03,939] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_19-model_00-model_states.pt... 0: [2022-12-26 13:09:04,119] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_19-model_00-model_states.pt. 0: [2022-12-26 13:09:04,119] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_20-model_00-model_states.pt... 0: [2022-12-26 13:09:04,291] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_20-model_00-model_states.pt. 0: [2022-12-26 13:09:04,291] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_21-model_00-model_states.pt... 0: [2022-12-26 13:09:04,465] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_21-model_00-model_states.pt. 0: [2022-12-26 13:09:04,466] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_22-model_00-model_states.pt... 0: [2022-12-26 13:09:04,642] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_22-model_00-model_states.pt. 0: [2022-12-26 13:09:04,642] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_23-model_00-model_states.pt... 0: [2022-12-26 13:09:04,815] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_23-model_00-model_states.pt. 0: [2022-12-26 13:09:04,815] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_24-model_00-model_states.pt... 0: [2022-12-26 13:09:04,992] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_24-model_00-model_states.pt. 0: [2022-12-26 13:09:04,993] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_25-model_00-model_states.pt... 0: [2022-12-26 13:09:05,168] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_25-model_00-model_states.pt. 0: [2022-12-26 13:09:05,168] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_26-model_00-model_states.pt... 0: [2022-12-26 13:09:05,337] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_26-model_00-model_states.pt. 0: [2022-12-26 13:09:05,337] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_27-model_00-model_states.pt... 0: [2022-12-26 13:09:05,517] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_27-model_00-model_states.pt. 0: [2022-12-26 13:09:05,517] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_28-model_00-model_states.pt... 0: [2022-12-26 13:09:05,687] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_28-model_00-model_states.pt. 0: [2022-12-26 13:09:05,687] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_29-model_00-model_states.pt... 0: [2022-12-26 13:09:05,864] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_29-model_00-model_states.pt. 0: [2022-12-26 13:09:05,864] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_30-model_00-model_states.pt... 0: [2022-12-26 13:09:06,039] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_30-model_00-model_states.pt. 0: [2022-12-26 13:09:06,039] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_31-model_00-model_states.pt... 0: [2022-12-26 13:09:06,208] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_31-model_00-model_states.pt. 0: [2022-12-26 13:09:06,208] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_32-model_00-model_states.pt... 0: [2022-12-26 13:09:06,385] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_32-model_00-model_states.pt. 0: [2022-12-26 13:09:06,385] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_33-model_00-model_states.pt... 0: [2022-12-26 13:09:06,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_33-model_00-model_states.pt. 0: [2022-12-26 13:09:06,553] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_34-model_00-model_states.pt... 0: [2022-12-26 13:09:06,726] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_34-model_00-model_states.pt. 0: [2022-12-26 13:09:06,726] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_35-model_00-model_states.pt... 0: [2022-12-26 13:09:06,899] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_35-model_00-model_states.pt. 0: [2022-12-26 13:09:06,900] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_36-model_00-model_states.pt... 0: [2022-12-26 13:09:07,072] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_36-model_00-model_states.pt. 0: [2022-12-26 13:09:07,073] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/layer_38-model_00-model_states.pt... 0: [2022-12-26 13:09:07,078] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/layer_38-model_00-model_states.pt. 0: [2022-12-26 13:09:07,079] [INFO] [logging.py:68:log_dist] [Rank 0] Saving model checkpoint: checkpoints_2b855brealtasky/global_step40000/mp_rank_00_model_states.pt 0: [2022-12-26 13:09:07,079] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/mp_rank_00_model_states.pt... 0: [2022-12-26 13:09:07,085] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/mp_rank_00_model_states.pt. 0: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... 0: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... 23: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt... 0: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... 0: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... 16: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt... 16: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt... 14: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt... 14: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt... 14: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt... 12: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt... 12: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt... 12: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt... 12: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt... 12: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt... 3: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt... 3: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt... 3: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt... 3: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt... 3: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt... 18: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt... 18: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt... 18: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt... 18: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt... 18: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt... 8: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt... 8: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt... 8: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt... 8: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt... 8: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt... 5: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt... 4: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt... 4: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt... 4: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt... 4: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt... 7: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt... 7: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt... 7: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt... 7: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt... 7: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt... 9: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt... 13: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt... 13: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt... 13: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt... 13: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt... 13: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt... 24: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt... 24: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt... 24: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt... 24: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt... 17: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt... 17: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt... 17: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt... 17: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt... 17: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt... 28: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt... 28: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt... 28: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt... 28: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt... 28: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt... 21: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt... 30: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt... 30: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt... 30: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt... 30: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt... 26: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt... 26: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt... 26: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt... 19: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt... 25: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt... 25: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt... 25: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt... 25: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt... 25: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt... 31: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt... 31: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt... 31: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt... 31: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt... 29: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt... 29: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt... 29: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt... 29: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt... 11: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt... 11: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt... 11: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt... 11: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt... 11: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt... 10: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt... 10: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt... 10: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt... 10: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt... 23: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt... 23: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt... 23: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt... 23: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt... 23: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt... 20: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt... 20: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt... 20: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt... 2: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt... 2: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt... 27: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt... 27: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt... 27: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt... 22: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt... 22: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt... 22: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt... 22: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt... 22: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt... 1: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... 6: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt... 6: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt... 6: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt... 15: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt... 15: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt... 15: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt... 15: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt... 15: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt... 0: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... 16: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt... 16: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt... 16: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt... 14: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt... 14: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt... 14: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt... 12: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt... 12: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt... 12: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt... 3: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt... 3: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt... 18: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt... 8: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt... 8: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt... 8: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt... 5: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt... 5: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt... 5: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt... 5: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt... 5: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt... 4: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt... 4: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt... 4: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt... 7: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt... 9: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt... 9: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt... 9: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt... 9: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt... 9: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt... 13: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt... 24: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt... 24: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt... 24: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt... 17: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt... 17: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt... 17: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt... 28: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt... 28: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt... 21: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt... 21: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt... 21: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt... 21: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt... 21: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt... 30: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt... 30: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt... 30: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt... 26: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt... 19: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt... 19: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt... 19: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt... 19: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt... 19: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt... 25: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt... 31: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt... 31: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt... 29: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt... 29: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt... 29: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt... 29: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt... 11: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt... 10: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt... 10: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt... 10: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt... 10: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt... 23: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt... 20: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt... 20: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt... 20: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt... 2: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt... 2: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt... 2: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt... 2: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt... 27: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt... 27: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt... 27: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt... 22: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt... 1: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... 1: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... 1: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... 1: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... 1: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... 6: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt... 6: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt... 15: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt... 15: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt... 0: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... 16: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt... 14: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt... 3: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt... 18: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt... 18: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt... 5: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt... 5: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt... 4: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt... 7: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt... 9: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt... 13: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt... 13: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt... 24: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt... 28: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt... 21: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt... 30: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt... 26: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt... 26: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt... 26: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt... 26: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt... 19: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt... 19: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt... 25: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt... 25: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt... 31: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt... 31: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt... 11: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt... 11: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt... 23: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt... 20: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt... 20: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt... 2: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt... 27: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt... 27: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt... 22: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt... 1: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... 1: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... 6: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt... 6: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt... 15: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt... 0: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... 16: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt... 14: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt... 7: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt... 9: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt... 21: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt... 2: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt... 6: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt... 0: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... 16: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt... 22: [2022-12-26 13:09:07,179] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt... 0: [2022-12-26 13:09:07,335] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. 0: [2022-12-26 13:09:07,335] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt 0: [2022-12-26 13:09:07,335] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 0: [2022-12-26 13:09:07,335] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. 0: [2022-12-26 13:09:07,335] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. 0: [2022-12-26 13:09:07,335] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt 0: [2022-12-26 13:09:07,335] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt 0: [2022-12-26 13:09:07,335] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 0: [2022-12-26 13:09:07,335] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 0: [2022-12-26 13:09:07,347] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. 0: [2022-12-26 13:09:07,347] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. 0: [2022-12-26 13:09:07,347] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt 0: [2022-12-26 13:09:07,347] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt 0: [2022-12-26 13:09:07,347] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 0: [2022-12-26 13:09:07,347] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 0: [2022-12-26 13:09:07,367] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. 0: [2022-12-26 13:09:07,367] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. 0: [2022-12-26 13:09:07,367] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt 0: [2022-12-26 13:09:07,367] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 0: [2022-12-26 13:09:07,368] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. 0: [2022-12-26 13:09:07,368] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt 0: [2022-12-26 13:09:07,368] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 3: [2022-12-26 13:09:07,379] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt. 3: [2022-12-26 13:09:07,379] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt. 3: [2022-12-26 13:09:07,379] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt. 3: [2022-12-26 13:09:07,379] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt 3: [2022-12-26 13:09:07,379] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt 3: [2022-12-26 13:09:07,379] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt 3: [2022-12-26 13:09:07,379] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 3: [2022-12-26 13:09:07,379] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 3: [2022-12-26 13:09:07,379] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 21: [2022-12-26 13:09:07,390] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt. 21: [2022-12-26 13:09:07,390] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt. 21: [2022-12-26 13:09:07,390] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt 21: [2022-12-26 13:09:07,390] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt 21: [2022-12-26 13:09:07,390] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 21: [2022-12-26 13:09:07,391] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 21: [2022-12-26 13:09:07,391] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt. 21: [2022-12-26 13:09:07,391] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt 21: [2022-12-26 13:09:07,391] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 21: [2022-12-26 13:09:07,392] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt. 21: [2022-12-26 13:09:07,392] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt 21: [2022-12-26 13:09:07,392] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 21: [2022-12-26 13:09:07,397] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt. 21: [2022-12-26 13:09:07,397] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt 21: [2022-12-26 13:09:07,397] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 10: [2022-12-26 13:09:07,409] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt. 10: [2022-12-26 13:09:07,409] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt. 10: [2022-12-26 13:09:07,409] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt 10: [2022-12-26 13:09:07,409] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt 10: [2022-12-26 13:09:07,409] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 10: [2022-12-26 13:09:07,409] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 10: [2022-12-26 13:09:07,409] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt. 10: [2022-12-26 13:09:07,409] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt 10: [2022-12-26 13:09:07,409] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 10: [2022-12-26 13:09:07,431] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt. 10: [2022-12-26 13:09:07,431] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt 10: [2022-12-26 13:09:07,431] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 16: [2022-12-26 13:09:07,436] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt. 16: [2022-12-26 13:09:07,436] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt. 16: [2022-12-26 13:09:07,436] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt. 16: [2022-12-26 13:09:07,436] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt. 16: [2022-12-26 13:09:07,436] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt. 16: [2022-12-26 13:09:07,436] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt. 16: [2022-12-26 13:09:07,436] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt. 16: [2022-12-26 13:09:07,437] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt 16: [2022-12-26 13:09:07,437] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt 16: [2022-12-26 13:09:07,437] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt 16: [2022-12-26 13:09:07,437] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt 16: [2022-12-26 13:09:07,437] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 16: [2022-12-26 13:09:07,437] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt 16: [2022-12-26 13:09:07,437] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt 16: [2022-12-26 13:09:07,437] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt 16: [2022-12-26 13:09:07,437] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 16: [2022-12-26 13:09:07,437] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 16: [2022-12-26 13:09:07,437] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 16: [2022-12-26 13:09:07,437] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 16: [2022-12-26 13:09:07,437] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 16: [2022-12-26 13:09:07,437] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 10: [2022-12-26 13:09:07,441] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt. 10: [2022-12-26 13:09:07,441] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt 10: [2022-12-26 13:09:07,441] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 21: [2022-12-26 13:09:07,444] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt. 21: [2022-12-26 13:09:07,444] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt 21: [2022-12-26 13:09:07,444] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 15: [2022-12-26 13:09:07,445] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt. 15: [2022-12-26 13:09:07,445] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt. 15: [2022-12-26 13:09:07,445] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt. 15: [2022-12-26 13:09:07,445] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt. 15: [2022-12-26 13:09:07,445] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt 15: [2022-12-26 13:09:07,445] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt 15: [2022-12-26 13:09:07,445] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt 15: [2022-12-26 13:09:07,445] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt 15: [2022-12-26 13:09:07,445] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 15: [2022-12-26 13:09:07,445] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 15: [2022-12-26 13:09:07,445] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 15: [2022-12-26 13:09:07,445] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt. 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt. 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt. 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt. 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt. 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt. 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt. 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt. 12: [2022-12-26 13:09:07,453] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt 12: [2022-12-26 13:09:07,453] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt 12: [2022-12-26 13:09:07,453] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt 12: [2022-12-26 13:09:07,453] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt 12: [2022-12-26 13:09:07,453] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt 12: [2022-12-26 13:09:07,453] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt 12: [2022-12-26 13:09:07,453] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 12: [2022-12-26 13:09:07,453] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 12: [2022-12-26 13:09:07,453] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt. 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt. 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt. 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt. 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt. 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt. 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt. 19: [2022-12-26 13:09:07,458] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt. 19: [2022-12-26 13:09:07,458] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt 19: [2022-12-26 13:09:07,458] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 19: [2022-12-26 13:09:07,458] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt 19: [2022-12-26 13:09:07,458] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt 19: [2022-12-26 13:09:07,458] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt 19: [2022-12-26 13:09:07,458] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 19: [2022-12-26 13:09:07,458] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 19: [2022-12-26 13:09:07,458] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt. 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt. 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt. 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt. 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt. 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt. 17: [2022-12-26 13:09:07,473] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt 17: [2022-12-26 13:09:07,473] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt 17: [2022-12-26 13:09:07,473] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt 17: [2022-12-26 13:09:07,473] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 17: [2022-12-26 13:09:07,473] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 17: [2022-12-26 13:09:07,473] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 17: [2022-12-26 13:09:07,473] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 15: [2022-12-26 13:09:07,477] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt. 15: [2022-12-26 13:09:07,477] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt 15: [2022-12-26 13:09:07,477] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 15: [2022-12-26 13:09:07,477] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt. 15: [2022-12-26 13:09:07,477] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt 15: [2022-12-26 13:09:07,477] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt. 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt. 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt. 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt. 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt. 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt. 18: [2022-12-26 13:09:07,489] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt 18: [2022-12-26 13:09:07,489] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt 18: [2022-12-26 13:09:07,489] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt 18: [2022-12-26 13:09:07,489] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt 18: [2022-12-26 13:09:07,489] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt 18: [2022-12-26 13:09:07,489] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 18: [2022-12-26 13:09:07,489] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 11: [2022-12-26 13:09:07,497] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt. 11: [2022-12-26 13:09:07,497] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt 11: [2022-12-26 13:09:07,497] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 11: [2022-12-26 13:09:07,497] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt. 11: [2022-12-26 13:09:07,497] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt 11: [2022-12-26 13:09:07,497] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 11: [2022-12-26 13:09:07,500] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt. 11: [2022-12-26 13:09:07,500] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt 11: [2022-12-26 13:09:07,500] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt. 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt. 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt. 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt. 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt. 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt. 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt. 9: [2022-12-26 13:09:07,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt 9: [2022-12-26 13:09:07,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt 9: [2022-12-26 13:09:07,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt 9: [2022-12-26 13:09:07,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt 9: [2022-12-26 13:09:07,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt 9: [2022-12-26 13:09:07,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt 9: [2022-12-26 13:09:07,496] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 9: [2022-12-26 13:09:07,496] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 11: [2022-12-26 13:09:07,502] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt. 11: [2022-12-26 13:09:07,502] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt 11: [2022-12-26 13:09:07,502] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 6: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt. 6: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt. 6: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt. 6: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt. 6: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt. 6: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt. 6: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt 6: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt. 6: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt 6: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt 6: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt 6: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt 6: [2022-12-26 13:09:07,507] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 6: [2022-12-26 13:09:07,507] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt 6: [2022-12-26 13:09:07,507] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 6: [2022-12-26 13:09:07,507] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 6: [2022-12-26 13:09:07,507] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 6: [2022-12-26 13:09:07,507] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 6: [2022-12-26 13:09:07,507] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt 6: [2022-12-26 13:09:07,507] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 6: [2022-12-26 13:09:07,507] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt. 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt. 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt. 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt. 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt. 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt. 26: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt 26: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt 26: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt 26: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt. 26: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt. 26: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 26: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt 26: [2022-12-26 13:09:07,506] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 26: [2022-12-26 13:09:07,506] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 11: [2022-12-26 13:09:07,508] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt. 11: [2022-12-26 13:09:07,508] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt 11: [2022-12-26 13:09:07,508] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 27: [2022-12-26 13:09:07,509] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt. 27: [2022-12-26 13:09:07,509] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt. 27: [2022-12-26 13:09:07,509] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt. 27: [2022-12-26 13:09:07,509] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt 27: [2022-12-26 13:09:07,510] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt 27: [2022-12-26 13:09:07,510] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt 27: [2022-12-26 13:09:07,510] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 27: [2022-12-26 13:09:07,510] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 27: [2022-12-26 13:09:07,510] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 27: [2022-12-26 13:09:07,512] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt. 27: [2022-12-26 13:09:07,512] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt. 27: [2022-12-26 13:09:07,512] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt. 27: [2022-12-26 13:09:07,512] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt. 27: [2022-12-26 13:09:07,512] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt 27: [2022-12-26 13:09:07,512] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt 27: [2022-12-26 13:09:07,512] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt 27: [2022-12-26 13:09:07,512] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt 27: [2022-12-26 13:09:07,512] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 27: [2022-12-26 13:09:07,512] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 27: [2022-12-26 13:09:07,512] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 27: [2022-12-26 13:09:07,512] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 18: [2022-12-26 13:09:07,515] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt. 18: [2022-12-26 13:09:07,515] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt 18: [2022-12-26 13:09:07,515] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt. 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt. 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt. 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt. 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt. 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt. 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt. 30: [2022-12-26 13:09:07,514] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt 30: [2022-12-26 13:09:07,514] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt 30: [2022-12-26 13:09:07,514] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt 30: [2022-12-26 13:09:07,514] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt 30: [2022-12-26 13:09:07,514] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt 30: [2022-12-26 13:09:07,514] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt 30: [2022-12-26 13:09:07,514] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 30: [2022-12-26 13:09:07,514] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 31: [2022-12-26 13:09:07,522] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt. 31: [2022-12-26 13:09:07,522] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt. 31: [2022-12-26 13:09:07,522] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt. 0: [2022-12-26 13:09:07,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt 0: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt. 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt. 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt. 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt. 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt. 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt. 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt. 25: [2022-12-26 13:09:07,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt 25: [2022-12-26 13:09:07,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt 25: [2022-12-26 13:09:07,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt 25: [2022-12-26 13:09:07,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt 25: [2022-12-26 13:09:07,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 25: [2022-12-26 13:09:07,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 25: [2022-12-26 13:09:07,531] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 25: [2022-12-26 13:09:07,531] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 11: [2022-12-26 13:09:07,532] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt. 11: [2022-12-26 13:09:07,532] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt 11: [2022-12-26 13:09:07,532] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 23: [2022-12-26 13:09:07,534] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt. 23: [2022-12-26 13:09:07,534] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt. 23: [2022-12-26 13:09:07,534] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt. 23: [2022-12-26 13:09:07,534] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt. 23: [2022-12-26 13:09:07,534] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt. 23: [2022-12-26 13:09:07,534] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt. 23: [2022-12-26 13:09:07,535] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt 23: [2022-12-26 13:09:07,535] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt 23: [2022-12-26 13:09:07,535] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt 23: [2022-12-26 13:09:07,535] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt 23: [2022-12-26 13:09:07,535] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt 23: [2022-12-26 13:09:07,535] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 23: [2022-12-26 13:09:07,535] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt 23: [2022-12-26 13:09:07,535] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 23: [2022-12-26 13:09:07,535] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 23: [2022-12-26 13:09:07,535] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 23: [2022-12-26 13:09:07,535] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 23: [2022-12-26 13:09:07,535] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt. 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt. 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt. 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt. 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt. 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt. 2: [2022-12-26 13:09:07,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt 2: [2022-12-26 13:09:07,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt 2: [2022-12-26 13:09:07,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt 2: [2022-12-26 13:09:07,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt 2: [2022-12-26 13:09:07,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt 2: [2022-12-26 13:09:07,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt. 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 2: [2022-12-26 13:09:07,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt 2: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt. 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt. 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt. 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt. 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt. 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt. 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt. 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt. 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt. 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt. 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt. 29: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt. 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt. 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt. 24: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt 29: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt 29: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt 29: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt 24: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt 24: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt 24: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt 29: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt 29: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 24: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt 24: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 24: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt. 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,541] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt 24: [2022-12-26 13:09:07,541] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 29: [2022-12-26 13:09:07,542] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 15: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt. 15: [2022-12-26 13:09:07,540] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt 15: [2022-12-26 13:09:07,540] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 17: [2022-12-26 13:09:07,543] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt. 23: [2022-12-26 13:09:07,543] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt. 17: [2022-12-26 13:09:07,543] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt 23: [2022-12-26 13:09:07,543] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt 17: [2022-12-26 13:09:07,543] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 23: [2022-12-26 13:09:07,543] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 31: [2022-12-26 13:09:07,522] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt. 31: [2022-12-26 13:09:07,522] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt 31: [2022-12-26 13:09:07,522] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt 31: [2022-12-26 13:09:07,522] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt 31: [2022-12-26 13:09:07,522] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 31: [2022-12-26 13:09:07,522] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 31: [2022-12-26 13:09:07,522] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt 31: [2022-12-26 13:09:07,522] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 31: [2022-12-26 13:09:07,522] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 31: [2022-12-26 13:09:07,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt. 31: [2022-12-26 13:09:07,537] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt 31: [2022-12-26 13:09:07,537] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 31: [2022-12-26 13:09:07,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt. 31: [2022-12-26 13:09:07,537] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt 31: [2022-12-26 13:09:07,537] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 31: [2022-12-26 13:09:07,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt. 31: [2022-12-26 13:09:07,537] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt 31: [2022-12-26 13:09:07,538] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 10: [2022-12-26 13:09:07,549] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt. 10: [2022-12-26 13:09:07,549] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt 10: [2022-12-26 13:09:07,549] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt. 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt. 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt. 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt. 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt. 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt. 13: [2022-12-26 13:09:07,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt 13: [2022-12-26 13:09:07,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt 13: [2022-12-26 13:09:07,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt. 13: [2022-12-26 13:09:07,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 13: [2022-12-26 13:09:07,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt 13: [2022-12-26 13:09:07,552] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 13: [2022-12-26 13:09:07,552] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 13: [2022-12-26 13:09:07,553] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt 13: [2022-12-26 13:09:07,553] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 21: [2022-12-26 13:09:07,553] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt. 21: [2022-12-26 13:09:07,553] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt 21: [2022-12-26 13:09:07,553] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 23: [2022-12-26 13:09:07,558] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt. 23: [2022-12-26 13:09:07,558] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt 23: [2022-12-26 13:09:07,559] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt. 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt. 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt. 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt. 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt. 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt. 14: [2022-12-26 13:09:07,562] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt. 14: [2022-12-26 13:09:07,562] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt 14: [2022-12-26 13:09:07,562] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt 14: [2022-12-26 13:09:07,562] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt 14: [2022-12-26 13:09:07,562] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt 14: [2022-12-26 13:09:07,562] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 14: [2022-12-26 13:09:07,562] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt 14: [2022-12-26 13:09:07,562] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt. 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt. 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt. 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt. 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt. 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt. 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt. 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt. 22: [2022-12-26 13:09:07,568] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt 22: [2022-12-26 13:09:07,568] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt 22: [2022-12-26 13:09:07,568] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt 22: [2022-12-26 13:09:07,568] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt 22: [2022-12-26 13:09:07,568] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt 22: [2022-12-26 13:09:07,568] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 22: [2022-12-26 13:09:07,568] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt 22: [2022-12-26 13:09:07,568] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 22: [2022-12-26 13:09:07,568] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt. 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt. 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt. 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt. 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt. 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt. 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt. 8: [2022-12-26 13:09:07,569] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt 8: [2022-12-26 13:09:07,569] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt 8: [2022-12-26 13:09:07,569] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt 8: [2022-12-26 13:09:07,569] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt 8: [2022-12-26 13:09:07,569] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt 8: [2022-12-26 13:09:07,569] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 8: [2022-12-26 13:09:07,569] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 8: [2022-12-26 13:09:07,569] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 6: [2022-12-26 13:09:07,572] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt. 6: [2022-12-26 13:09:07,573] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt 6: [2022-12-26 13:09:07,573] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 21: [2022-12-26 13:09:07,573] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt. 21: [2022-12-26 13:09:07,573] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt 21: [2022-12-26 13:09:07,573] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 3: [2022-12-26 13:09:07,574] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt. 3: [2022-12-26 13:09:07,574] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt 3: [2022-12-26 13:09:07,574] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt. 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt. 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt. 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt. 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt. 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt. 28: [2022-12-26 13:09:07,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt. 28: [2022-12-26 13:09:07,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 28: [2022-12-26 13:09:07,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt 28: [2022-12-26 13:09:07,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 28: [2022-12-26 13:09:07,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt 28: [2022-12-26 13:09:07,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt 28: [2022-12-26 13:09:07,575] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 28: [2022-12-26 13:09:07,575] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt. 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt. 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt. 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt. 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt. 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt. 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt. 4: [2022-12-26 13:09:07,576] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt 4: [2022-12-26 13:09:07,576] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt 4: [2022-12-26 13:09:07,576] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt 4: [2022-12-26 13:09:07,576] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt 4: [2022-12-26 13:09:07,576] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt 4: [2022-12-26 13:09:07,576] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 4: [2022-12-26 13:09:07,576] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 4: [2022-12-26 13:09:07,576] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 10: [2022-12-26 13:09:07,577] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt. 10: [2022-12-26 13:09:07,577] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt 10: [2022-12-26 13:09:07,577] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 28: [2022-12-26 13:09:07,579] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt. 28: [2022-12-26 13:09:07,579] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt 28: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 20: [2022-12-26 13:09:07,579] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt. 20: [2022-12-26 13:09:07,579] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt. 20: [2022-12-26 13:09:07,579] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt. 20: [2022-12-26 13:09:07,579] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt. 20: [2022-12-26 13:09:07,579] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt. 20: [2022-12-26 13:09:07,579] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt. 20: [2022-12-26 13:09:07,579] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt. 20: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt. 20: [2022-12-26 13:09:07,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt 20: [2022-12-26 13:09:07,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt 20: [2022-12-26 13:09:07,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt 20: [2022-12-26 13:09:07,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt 20: [2022-12-26 13:09:07,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt 20: [2022-12-26 13:09:07,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt 20: [2022-12-26 13:09:07,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt 20: [2022-12-26 13:09:07,580] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt 20: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 20: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 20: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 20: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 20: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 20: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 20: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 20: [2022-12-26 13:09:07,580] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 31: [2022-12-26 13:09:07,584] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt. 31: [2022-12-26 13:09:07,584] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt 31: [2022-12-26 13:09:07,584] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 30: [2022-12-26 13:09:07,586] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt. 30: [2022-12-26 13:09:07,586] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt 30: [2022-12-26 13:09:07,586] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 11: [2022-12-26 13:09:07,587] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt. 11: [2022-12-26 13:09:07,587] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt 11: [2022-12-26 13:09:07,587] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. 1: [2022-12-26 13:09:07,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt 1: [2022-12-26 13:09:07,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt 1: [2022-12-26 13:09:07,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt 1: [2022-12-26 13:09:07,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt 1: [2022-12-26 13:09:07,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt 1: [2022-12-26 13:09:07,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 1: [2022-12-26 13:09:07,588] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 1: [2022-12-26 13:09:07,588] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 1: [2022-12-26 13:09:07,590] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. 1: [2022-12-26 13:09:07,590] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt 1: [2022-12-26 13:09:07,590] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 4: [2022-12-26 13:09:07,594] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt. 4: [2022-12-26 13:09:07,594] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt 4: [2022-12-26 13:09:07,594] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 17: [2022-12-26 13:09:07,594] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt. 17: [2022-12-26 13:09:07,594] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt 17: [2022-12-26 13:09:07,594] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 25: [2022-12-26 13:09:07,594] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt. 25: [2022-12-26 13:09:07,594] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt 25: [2022-12-26 13:09:07,594] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 8: [2022-12-26 13:09:07,596] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt. 8: [2022-12-26 13:09:07,597] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt 8: [2022-12-26 13:09:07,597] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt. 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt. 7: [2022-12-26 13:09:07,603] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt. 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt. 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt. 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt. 7: [2022-12-26 13:09:07,603] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 7: [2022-12-26 13:09:07,603] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 7: [2022-12-26 13:09:07,603] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt 7: [2022-12-26 13:09:07,603] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt 7: [2022-12-26 13:09:07,603] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt. 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt. 7: [2022-12-26 13:09:07,603] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt 7: [2022-12-26 13:09:07,603] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 7: [2022-12-26 13:09:07,603] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 2: [2022-12-26 13:09:07,606] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt. 2: [2022-12-26 13:09:07,606] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt 2: [2022-12-26 13:09:07,607] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 27: [2022-12-26 13:09:07,614] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt. 27: [2022-12-26 13:09:07,614] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt 27: [2022-12-26 13:09:07,614] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt. 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt. 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt. 5: [2022-12-26 13:09:07,615] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt. 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt. 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt. 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt. 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt. 5: [2022-12-26 13:09:07,615] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt 5: [2022-12-26 13:09:07,615] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt 5: [2022-12-26 13:09:07,615] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt 5: [2022-12-26 13:09:07,615] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt 5: [2022-12-26 13:09:07,615] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt 5: [2022-12-26 13:09:07,615] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt 5: [2022-12-26 13:09:07,615] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 5: [2022-12-26 13:09:07,615] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 18: [2022-12-26 13:09:07,617] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt. 18: [2022-12-26 13:09:07,617] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt 18: [2022-12-26 13:09:07,617] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 11: [2022-12-26 13:09:07,631] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt. 11: [2022-12-26 13:09:07,631] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt 11: [2022-12-26 13:09:07,631] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 16: [2022-12-26 13:09:07,639] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt. 16: [2022-12-26 13:09:07,639] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt 16: [2022-12-26 13:09:07,639] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 9: [2022-12-26 13:09:07,640] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt. 9: [2022-12-26 13:09:07,640] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt 9: [2022-12-26 13:09:07,640] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 24: [2022-12-26 13:09:07,648] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt. 24: [2022-12-26 13:09:07,648] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt 24: [2022-12-26 13:09:07,648] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 13: [2022-12-26 13:09:07,650] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt. 13: [2022-12-26 13:09:07,651] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt 13: [2022-12-26 13:09:07,651] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 3: [2022-12-26 13:09:07,658] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt. 3: [2022-12-26 13:09:07,659] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt 3: [2022-12-26 13:09:07,659] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 15: [2022-12-26 13:09:07,659] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt. 15: [2022-12-26 13:09:07,659] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt 15: [2022-12-26 13:09:07,659] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 14: [2022-12-26 13:09:07,663] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt. 14: [2022-12-26 13:09:07,663] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt 14: [2022-12-26 13:09:07,663] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 3: [2022-12-26 13:09:07,676] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt. 3: [2022-12-26 13:09:07,676] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt 3: [2022-12-26 13:09:07,676] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 3: [2022-12-26 13:09:07,692] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt. 3: [2022-12-26 13:09:07,692] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt 3: [2022-12-26 13:09:07,692] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 3: [2022-12-26 13:09:07,704] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt. 3: [2022-12-26 13:09:07,704] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt 3: [2022-12-26 13:09:07,704] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 10: [2022-12-26 13:09:07,744] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt. 10: [2022-12-26 13:09:07,744] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step40000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt 10: [2022-12-26 13:09:07,744] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step40000 is ready now! 0: successfully saved checkpoint at iteration 40000 to checkpoints_2b855brealtasky 31: time (ms) | save-checkpoint: 7115.09 31: iteration 40010/ 52452 | consumed samples: 20485120 | consumed tokens: 41953525760 | elapsed time per iteration (s): 9.87 | learning rate: 4.432E-05 | global batch size: 512 | lm loss: 1.191895E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 51.879 | TFLOPs: 7.79 | 31: iteration 40020/ 52452 | consumed samples: 20490240 | consumed tokens: 41964011520 | elapsed time per iteration (s): 1.77 | learning rate: 4.428E-05 | global batch size: 512 | lm loss: 1.186572E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.065 | TFLOPs: 43.54 | 31: iteration 40030/ 52452 | consumed samples: 20495360 | consumed tokens: 41974497280 | elapsed time per iteration (s): 1.78 | learning rate: 4.424E-05 | global batch size: 512 | lm loss: 1.178593E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.547 | TFLOPs: 43.16 | 31: iteration 40040/ 52452 | consumed samples: 20500480 | consumed tokens: 41984983040 | elapsed time per iteration (s): 1.76 | learning rate: 4.420E-05 | global batch size: 512 | lm loss: 1.186989E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.657 | TFLOPs: 43.63 | 31: iteration 40050/ 52452 | consumed samples: 20505600 | consumed tokens: 41995468800 | elapsed time per iteration (s): 1.74 | learning rate: 4.417E-05 | global batch size: 512 | lm loss: 1.189508E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.747 | TFLOPs: 44.09 | 31: iteration 40060/ 52452 | consumed samples: 20510720 | consumed tokens: 42005954560 | elapsed time per iteration (s): 1.74 | learning rate: 4.413E-05 | global batch size: 512 | lm loss: 1.200585E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.664 | TFLOPs: 44.23 | 31: iteration 40070/ 52452 | consumed samples: 20515840 | consumed tokens: 42016440320 | elapsed time per iteration (s): 1.76 | learning rate: 4.409E-05 | global batch size: 512 | lm loss: 1.191301E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.709 | TFLOPs: 43.63 | 31: iteration 40080/ 52452 | consumed samples: 20520960 | consumed tokens: 42026926080 | elapsed time per iteration (s): 1.76 | learning rate: 4.406E-05 | global batch size: 512 | lm loss: 1.202962E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.806 | TFLOPs: 43.65 | 31: iteration 40090/ 52452 | consumed samples: 20526080 | consumed tokens: 42037411840 | elapsed time per iteration (s): 1.74 | learning rate: 4.402E-05 | global batch size: 512 | lm loss: 1.184496E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.054 | TFLOPs: 44.29 | 31: iteration 40100/ 52452 | consumed samples: 20531200 | consumed tokens: 42047897600 | elapsed time per iteration (s): 1.79 | learning rate: 4.398E-05 | global batch size: 512 | lm loss: 1.200323E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.072 | TFLOPs: 42.94 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40100 | lm loss value: 4.297295E+00 | lm loss PPL: 7.350068E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40100 | lm loss value: 4.131241E+00 | lm loss PPL: 6.225512E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 40110/ 52452 | consumed samples: 20536320 | consumed tokens: 42058383360 | elapsed time per iteration (s): 9.17 | learning rate: 4.395E-05 | global batch size: 512 | lm loss: 1.180485E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.849 | TFLOPs: 8.38 | 31: iteration 40120/ 52452 | consumed samples: 20541440 | consumed tokens: 42068869120 | elapsed time per iteration (s): 1.76 | learning rate: 4.391E-05 | global batch size: 512 | lm loss: 1.201421E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.572 | TFLOPs: 43.76 | 31: iteration 40130/ 52452 | consumed samples: 20546560 | consumed tokens: 42079354880 | elapsed time per iteration (s): 1.74 | learning rate: 4.387E-05 | global batch size: 512 | lm loss: 1.185576E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.130 | TFLOPs: 44.15 | 31: iteration 40140/ 52452 | consumed samples: 20551680 | consumed tokens: 42089840640 | elapsed time per iteration (s): 1.75 | learning rate: 4.383E-05 | global batch size: 512 | lm loss: 1.183735E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.854 | TFLOPs: 43.96 | 31: iteration 40150/ 52452 | consumed samples: 20556800 | consumed tokens: 42100326400 | elapsed time per iteration (s): 1.76 | learning rate: 4.380E-05 | global batch size: 512 | lm loss: 1.188766E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.794 | TFLOPs: 43.65 | 31: iteration 40160/ 52452 | consumed samples: 20561920 | consumed tokens: 42110812160 | elapsed time per iteration (s): 1.79 | learning rate: 4.376E-05 | global batch size: 512 | lm loss: 1.196272E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.810 | TFLOPs: 42.90 | 31: iteration 40170/ 52452 | consumed samples: 20567040 | consumed tokens: 42121297920 | elapsed time per iteration (s): 1.77 | learning rate: 4.372E-05 | global batch size: 512 | lm loss: 1.192160E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.992 | TFLOPs: 43.38 | 31: iteration 40180/ 52452 | consumed samples: 20572160 | consumed tokens: 42131783680 | elapsed time per iteration (s): 1.74 | learning rate: 4.369E-05 | global batch size: 512 | lm loss: 1.205089E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.195 | TFLOPs: 44.16 | 31: iteration 40190/ 52452 | consumed samples: 20577280 | consumed tokens: 42142269440 | elapsed time per iteration (s): 1.80 | learning rate: 4.365E-05 | global batch size: 512 | lm loss: 1.191579E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.171 | TFLOPs: 42.65 | 31: iteration 40200/ 52452 | consumed samples: 20582400 | consumed tokens: 42152755200 | elapsed time per iteration (s): 1.79 | learning rate: 4.361E-05 | global batch size: 512 | lm loss: 1.200899E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.816 | TFLOPs: 43.05 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40200 | lm loss value: 4.212584E+00 | lm loss PPL: 6.753085E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40200 | lm loss value: 4.071606E+00 | lm loss PPL: 5.865109E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 40210/ 52452 | consumed samples: 20587520 | consumed tokens: 42163240960 | elapsed time per iteration (s): 9.15 | learning rate: 4.358E-05 | global batch size: 512 | lm loss: 1.190486E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.966 | TFLOPs: 8.40 | 31: iteration 40220/ 52452 | consumed samples: 20592640 | consumed tokens: 42173726720 | elapsed time per iteration (s): 1.76 | learning rate: 4.354E-05 | global batch size: 512 | lm loss: 1.198689E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.398 | TFLOPs: 43.59 | 31: iteration 40230/ 52452 | consumed samples: 20597760 | consumed tokens: 42184212480 | elapsed time per iteration (s): 1.76 | learning rate: 4.350E-05 | global batch size: 512 | lm loss: 1.189329E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.300 | TFLOPs: 43.72 | 31: iteration 40240/ 52452 | consumed samples: 20602880 | consumed tokens: 42194698240 | elapsed time per iteration (s): 1.74 | learning rate: 4.347E-05 | global batch size: 512 | lm loss: 1.185085E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.668 | TFLOPs: 44.23 | 31: iteration 40250/ 52452 | consumed samples: 20608000 | consumed tokens: 42205184000 | elapsed time per iteration (s): 1.73 | learning rate: 4.343E-05 | global batch size: 512 | lm loss: 1.185432E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.443 | TFLOPs: 44.49 | 31: iteration 40260/ 52452 | consumed samples: 20613120 | consumed tokens: 42215669760 | elapsed time per iteration (s): 1.73 | learning rate: 4.339E-05 | global batch size: 512 | lm loss: 1.205300E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.448 | TFLOPs: 44.35 | 31: iteration 40270/ 52452 | consumed samples: 20618240 | consumed tokens: 42226155520 | elapsed time per iteration (s): 1.74 | learning rate: 4.336E-05 | global batch size: 512 | lm loss: 1.183693E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.513 | TFLOPs: 44.20 | 31: iteration 40280/ 52452 | consumed samples: 20623360 | consumed tokens: 42236641280 | elapsed time per iteration (s): 1.74 | learning rate: 4.332E-05 | global batch size: 512 | lm loss: 1.185923E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.662 | TFLOPs: 44.08 | 31: iteration 40290/ 52452 | consumed samples: 20628480 | consumed tokens: 42247127040 | elapsed time per iteration (s): 1.74 | learning rate: 4.328E-05 | global batch size: 512 | lm loss: 1.192622E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.622 | TFLOPs: 44.07 | 31: iteration 40300/ 52452 | consumed samples: 20633600 | consumed tokens: 42257612800 | elapsed time per iteration (s): 1.77 | learning rate: 4.325E-05 | global batch size: 512 | lm loss: 1.193410E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.840 | TFLOPs: 43.50 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40300 | lm loss value: 4.212080E+00 | lm loss PPL: 6.749682E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40300 | lm loss value: 4.080143E+00 | lm loss PPL: 5.915393E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 40310/ 52452 | consumed samples: 20638720 | consumed tokens: 42268098560 | elapsed time per iteration (s): 9.15 | learning rate: 4.321E-05 | global batch size: 512 | lm loss: 1.195587E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.949 | TFLOPs: 8.40 | 31: iteration 40320/ 52452 | consumed samples: 20643840 | consumed tokens: 42278584320 | elapsed time per iteration (s): 1.74 | learning rate: 4.317E-05 | global batch size: 512 | lm loss: 1.190123E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.634 | TFLOPs: 44.22 | 31: iteration 40330/ 52452 | consumed samples: 20648960 | consumed tokens: 42289070080 | elapsed time per iteration (s): 1.74 | learning rate: 4.314E-05 | global batch size: 512 | lm loss: 1.193578E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.952 | TFLOPs: 44.27 | 31: iteration 40340/ 52452 | consumed samples: 20654080 | consumed tokens: 42299555840 | elapsed time per iteration (s): 1.73 | learning rate: 4.310E-05 | global batch size: 512 | lm loss: 1.175722E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.335 | TFLOPs: 44.33 | 31: iteration 40350/ 52452 | consumed samples: 20659200 | consumed tokens: 42310041600 | elapsed time per iteration (s): 1.76 | learning rate: 4.306E-05 | global batch size: 512 | lm loss: 1.173507E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.227 | TFLOPs: 43.56 | 31: iteration 40360/ 52452 | consumed samples: 20664320 | consumed tokens: 42320527360 | elapsed time per iteration (s): 1.75 | learning rate: 4.303E-05 | global batch size: 512 | lm loss: 1.205762E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.866 | TFLOPs: 43.96 | 31: iteration 40370/ 52452 | consumed samples: 20669440 | consumed tokens: 42331013120 | elapsed time per iteration (s): 1.76 | learning rate: 4.299E-05 | global batch size: 512 | lm loss: 1.192783E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.752 | TFLOPs: 43.64 | 31: iteration 40380/ 52452 | consumed samples: 20674560 | consumed tokens: 42341498880 | elapsed time per iteration (s): 1.73 | learning rate: 4.296E-05 | global batch size: 512 | lm loss: 1.199275E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.057 | TFLOPs: 44.44 | 31: iteration 40390/ 52452 | consumed samples: 20679680 | consumed tokens: 42351984640 | elapsed time per iteration (s): 1.74 | learning rate: 4.292E-05 | global batch size: 512 | lm loss: 1.195037E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.250 | TFLOPs: 44.17 | 31: iteration 40400/ 52452 | consumed samples: 20684800 | consumed tokens: 42362470400 | elapsed time per iteration (s): 1.77 | learning rate: 4.288E-05 | global batch size: 512 | lm loss: 1.195478E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.729 | TFLOPs: 43.34 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40400 | lm loss value: 4.115439E+00 | lm loss PPL: 6.127911E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40400 | lm loss value: 4.007091E+00 | lm loss PPL: 5.498668E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 40410/ 52452 | consumed samples: 20689920 | consumed tokens: 42372956160 | elapsed time per iteration (s): 9.14 | learning rate: 4.285E-05 | global batch size: 512 | lm loss: 1.192406E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.007 | TFLOPs: 8.41 | 31: iteration 40420/ 52452 | consumed samples: 20695040 | consumed tokens: 42383441920 | elapsed time per iteration (s): 1.77 | learning rate: 4.281E-05 | global batch size: 512 | lm loss: 1.189558E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.864 | TFLOPs: 43.51 | 31: iteration 40430/ 52452 | consumed samples: 20700160 | consumed tokens: 42393927680 | elapsed time per iteration (s): 1.76 | learning rate: 4.277E-05 | global batch size: 512 | lm loss: 1.192014E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.581 | TFLOPs: 43.76 | 31: iteration 40440/ 52452 | consumed samples: 20705280 | consumed tokens: 42404413440 | elapsed time per iteration (s): 1.74 | learning rate: 4.274E-05 | global batch size: 512 | lm loss: 1.209460E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.886 | TFLOPs: 44.26 | 31: iteration 40450/ 52452 | consumed samples: 20710400 | consumed tokens: 42414899200 | elapsed time per iteration (s): 1.76 | learning rate: 4.270E-05 | global batch size: 512 | lm loss: 1.190187E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.135 | TFLOPs: 43.70 | 31: iteration 40460/ 52452 | consumed samples: 20715520 | consumed tokens: 42425384960 | elapsed time per iteration (s): 1.74 | learning rate: 4.267E-05 | global batch size: 512 | lm loss: 1.184329E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.426 | TFLOPs: 44.19 | 31: iteration 40470/ 52452 | consumed samples: 20720640 | consumed tokens: 42435870720 | elapsed time per iteration (s): 1.73 | learning rate: 4.263E-05 | global batch size: 512 | lm loss: 1.189138E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.684 | TFLOPs: 44.53 | 31: iteration 40480/ 52452 | consumed samples: 20725760 | consumed tokens: 42446356480 | elapsed time per iteration (s): 1.72 | learning rate: 4.259E-05 | global batch size: 512 | lm loss: 1.188481E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.489 | TFLOPs: 44.65 | 31: iteration 40490/ 52452 | consumed samples: 20730880 | consumed tokens: 42456842240 | elapsed time per iteration (s): 1.73 | learning rate: 4.256E-05 | global batch size: 512 | lm loss: 1.195641E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.185 | TFLOPs: 44.46 | 31: iteration 40500/ 52452 | consumed samples: 20736000 | consumed tokens: 42467328000 | elapsed time per iteration (s): 1.73 | learning rate: 4.252E-05 | global batch size: 512 | lm loss: 1.195887E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.869 | TFLOPs: 44.41 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40500 | lm loss value: 4.166106E+00 | lm loss PPL: 6.446392E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40500 | lm loss value: 4.041894E+00 | lm loss PPL: 5.693410E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 40510/ 52452 | consumed samples: 20741120 | consumed tokens: 42477813760 | elapsed time per iteration (s): 9.17 | learning rate: 4.249E-05 | global batch size: 512 | lm loss: 1.185677E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.816 | TFLOPs: 8.38 | 31: iteration 40520/ 52452 | consumed samples: 20746240 | consumed tokens: 42488299520 | elapsed time per iteration (s): 1.77 | learning rate: 4.245E-05 | global batch size: 512 | lm loss: 1.204148E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.706 | TFLOPs: 43.33 | 31: iteration 40530/ 52452 | consumed samples: 20751360 | consumed tokens: 42498785280 | elapsed time per iteration (s): 1.73 | learning rate: 4.241E-05 | global batch size: 512 | lm loss: 1.180953E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.207 | TFLOPs: 44.31 | 31: iteration 40540/ 52452 | consumed samples: 20756480 | consumed tokens: 42509271040 | elapsed time per iteration (s): 1.76 | learning rate: 4.238E-05 | global batch size: 512 | lm loss: 1.185629E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.737 | TFLOPs: 43.79 | 31: iteration 40550/ 52452 | consumed samples: 20761600 | consumed tokens: 42519756800 | elapsed time per iteration (s): 1.75 | learning rate: 4.234E-05 | global batch size: 512 | lm loss: 1.192705E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.349 | TFLOPs: 43.88 | 31: iteration 40560/ 52452 | consumed samples: 20766720 | consumed tokens: 42530242560 | elapsed time per iteration (s): 1.78 | learning rate: 4.231E-05 | global batch size: 512 | lm loss: 1.193884E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.322 | TFLOPs: 43.13 | 31: iteration 40570/ 52452 | consumed samples: 20771840 | consumed tokens: 42540728320 | elapsed time per iteration (s): 1.79 | learning rate: 4.227E-05 | global batch size: 512 | lm loss: 1.181076E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.427 | TFLOPs: 42.99 | 31: iteration 40580/ 52452 | consumed samples: 20776960 | consumed tokens: 42551214080 | elapsed time per iteration (s): 1.78 | learning rate: 4.223E-05 | global batch size: 512 | lm loss: 1.187843E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.852 | TFLOPs: 43.21 | 31: iteration 40590/ 52452 | consumed samples: 20782080 | consumed tokens: 42561699840 | elapsed time per iteration (s): 1.80 | learning rate: 4.220E-05 | global batch size: 512 | lm loss: 1.199313E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.218 | TFLOPs: 42.66 | 31: iteration 40600/ 52452 | consumed samples: 20787200 | consumed tokens: 42572185600 | elapsed time per iteration (s): 1.78 | learning rate: 4.216E-05 | global batch size: 512 | lm loss: 1.190049E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.366 | TFLOPs: 43.13 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40600 | lm loss value: 4.251756E+00 | lm loss PPL: 7.022864E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40600 | lm loss value: 4.101242E+00 | lm loss PPL: 6.041525E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 40610/ 52452 | consumed samples: 20792320 | consumed tokens: 42582671360 | elapsed time per iteration (s): 9.22 | learning rate: 4.213E-05 | global batch size: 512 | lm loss: 1.193923E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.514 | TFLOPs: 8.33 | 31: iteration 40620/ 52452 | consumed samples: 20797440 | consumed tokens: 42593157120 | elapsed time per iteration (s): 1.74 | learning rate: 4.209E-05 | global batch size: 512 | lm loss: 1.189235E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.780 | TFLOPs: 44.09 | 31: iteration 40630/ 52452 | consumed samples: 20802560 | consumed tokens: 42603642880 | elapsed time per iteration (s): 1.79 | learning rate: 4.206E-05 | global batch size: 512 | lm loss: 1.185467E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.144 | TFLOPs: 42.95 | 31: iteration 40640/ 52452 | consumed samples: 20807680 | consumed tokens: 42614128640 | elapsed time per iteration (s): 1.81 | learning rate: 4.202E-05 | global batch size: 512 | lm loss: 1.192371E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.290 | TFLOPs: 42.52 | 31: iteration 40650/ 52452 | consumed samples: 20812800 | consumed tokens: 42624614400 | elapsed time per iteration (s): 1.79 | learning rate: 4.198E-05 | global batch size: 512 | lm loss: 1.191900E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.279 | TFLOPs: 42.97 | 31: iteration 40660/ 52452 | consumed samples: 20817920 | consumed tokens: 42635100160 | elapsed time per iteration (s): 1.86 | learning rate: 4.195E-05 | global batch size: 512 | lm loss: 1.197079E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.610 | TFLOPs: 41.37 | 31: iteration 40670/ 52452 | consumed samples: 20823040 | consumed tokens: 42645585920 | elapsed time per iteration (s): 1.79 | learning rate: 4.191E-05 | global batch size: 512 | lm loss: 1.184305E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.562 | TFLOPs: 43.01 | 31: iteration 40680/ 52452 | consumed samples: 20828160 | consumed tokens: 42656071680 | elapsed time per iteration (s): 1.74 | learning rate: 4.188E-05 | global batch size: 512 | lm loss: 1.209322E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.535 | TFLOPs: 44.21 | 31: iteration 40690/ 52452 | consumed samples: 20833280 | consumed tokens: 42666557440 | elapsed time per iteration (s): 1.76 | learning rate: 4.184E-05 | global batch size: 512 | lm loss: 1.192961E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.642 | TFLOPs: 43.77 | 31: iteration 40700/ 52452 | consumed samples: 20838400 | consumed tokens: 42677043200 | elapsed time per iteration (s): 1.75 | learning rate: 4.181E-05 | global batch size: 512 | lm loss: 1.183558E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.670 | TFLOPs: 43.93 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40700 | lm loss value: 4.275115E+00 | lm loss PPL: 7.188837E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40700 | lm loss value: 4.129297E+00 | lm loss PPL: 6.213424E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 40710/ 52452 | consumed samples: 20843520 | consumed tokens: 42687528960 | elapsed time per iteration (s): 9.15 | learning rate: 4.177E-05 | global batch size: 512 | lm loss: 1.179892E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.958 | TFLOPs: 8.40 | 31: iteration 40720/ 52452 | consumed samples: 20848640 | consumed tokens: 42698014720 | elapsed time per iteration (s): 1.79 | learning rate: 4.173E-05 | global batch size: 512 | lm loss: 1.189833E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.747 | TFLOPs: 42.89 | 31: iteration 40730/ 52452 | consumed samples: 20853760 | consumed tokens: 42708500480 | elapsed time per iteration (s): 1.74 | learning rate: 4.170E-05 | global batch size: 512 | lm loss: 1.184277E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.675 | TFLOPs: 44.08 | 31: iteration 40740/ 52452 | consumed samples: 20858880 | consumed tokens: 42718986240 | elapsed time per iteration (s): 1.77 | learning rate: 4.166E-05 | global batch size: 512 | lm loss: 1.187169E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.977 | TFLOPs: 43.37 | 31: iteration 40750/ 52452 | consumed samples: 20864000 | consumed tokens: 42729472000 | elapsed time per iteration (s): 1.72 | learning rate: 4.163E-05 | global batch size: 512 | lm loss: 1.180440E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.127 | TFLOPs: 44.60 | 31: iteration 40760/ 52452 | consumed samples: 20869120 | consumed tokens: 42739957760 | elapsed time per iteration (s): 1.78 | learning rate: 4.159E-05 | global batch size: 512 | lm loss: 1.188791E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.632 | TFLOPs: 43.17 | 31: iteration 40770/ 52452 | consumed samples: 20874240 | consumed tokens: 42750443520 | elapsed time per iteration (s): 1.83 | learning rate: 4.156E-05 | global batch size: 512 | lm loss: 1.197135E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.518 | TFLOPs: 41.95 | 31: iteration 40780/ 52452 | consumed samples: 20879360 | consumed tokens: 42760929280 | elapsed time per iteration (s): 1.72 | learning rate: 4.152E-05 | global batch size: 512 | lm loss: 1.183960E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.641 | TFLOPs: 44.67 | 31: iteration 40790/ 52452 | consumed samples: 20884480 | consumed tokens: 42771415040 | elapsed time per iteration (s): 1.76 | learning rate: 4.149E-05 | global batch size: 512 | lm loss: 1.186028E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.364 | TFLOPs: 43.58 | 31: iteration 40800/ 52452 | consumed samples: 20889600 | consumed tokens: 42781900800 | elapsed time per iteration (s): 1.76 | learning rate: 4.145E-05 | global batch size: 512 | lm loss: 1.190931E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.120 | TFLOPs: 43.55 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40800 | lm loss value: 4.158519E+00 | lm loss PPL: 6.397669E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40800 | lm loss value: 4.045516E+00 | lm loss PPL: 5.714064E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 40810/ 52452 | consumed samples: 20894720 | consumed tokens: 42792386560 | elapsed time per iteration (s): 9.20 | learning rate: 4.142E-05 | global batch size: 512 | lm loss: 1.186032E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.626 | TFLOPs: 8.35 | 31: iteration 40820/ 52452 | consumed samples: 20899840 | consumed tokens: 42802872320 | elapsed time per iteration (s): 1.74 | learning rate: 4.138E-05 | global batch size: 512 | lm loss: 1.194207E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.901 | TFLOPs: 44.11 | 31: iteration 40830/ 52452 | consumed samples: 20904960 | consumed tokens: 42813358080 | elapsed time per iteration (s): 1.90 | learning rate: 4.135E-05 | global batch size: 512 | lm loss: 1.177883E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.702 | TFLOPs: 40.48 | 31: iteration 40840/ 52452 | consumed samples: 20910080 | consumed tokens: 42823843840 | elapsed time per iteration (s): 1.75 | learning rate: 4.131E-05 | global batch size: 512 | lm loss: 1.194980E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.468 | TFLOPs: 43.90 | 31: iteration 40850/ 52452 | consumed samples: 20915200 | consumed tokens: 42834329600 | elapsed time per iteration (s): 1.81 | learning rate: 4.128E-05 | global batch size: 512 | lm loss: 1.179948E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.400 | TFLOPs: 42.54 | 31: iteration 40860/ 52452 | consumed samples: 20920320 | consumed tokens: 42844815360 | elapsed time per iteration (s): 1.91 | learning rate: 4.124E-05 | global batch size: 512 | lm loss: 1.197644E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 267.806 | TFLOPs: 40.20 | 31: iteration 40870/ 52452 | consumed samples: 20925440 | consumed tokens: 42855301120 | elapsed time per iteration (s): 1.73 | learning rate: 4.121E-05 | global batch size: 512 | lm loss: 1.190267E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.638 | TFLOPs: 44.37 | 31: iteration 40880/ 52452 | consumed samples: 20930560 | consumed tokens: 42865786880 | elapsed time per iteration (s): 1.79 | learning rate: 4.117E-05 | global batch size: 512 | lm loss: 1.189097E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.082 | TFLOPs: 42.94 | 31: iteration 40890/ 52452 | consumed samples: 20935680 | consumed tokens: 42876272640 | elapsed time per iteration (s): 1.73 | learning rate: 4.114E-05 | global batch size: 512 | lm loss: 1.188622E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.108 | TFLOPs: 44.29 | 31: iteration 40900/ 52452 | consumed samples: 20940800 | consumed tokens: 42886758400 | elapsed time per iteration (s): 1.76 | learning rate: 4.110E-05 | global batch size: 512 | lm loss: 1.183454E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.543 | TFLOPs: 43.76 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 40900 | lm loss value: 4.115735E+00 | lm loss PPL: 6.129723E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 40900 | lm loss value: 4.012917E+00 | lm loss PPL: 5.530797E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 40910/ 52452 | consumed samples: 20945920 | consumed tokens: 42897244160 | elapsed time per iteration (s): 9.38 | learning rate: 4.107E-05 | global batch size: 512 | lm loss: 1.197803E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.561 | TFLOPs: 8.19 | 31: iteration 40920/ 52452 | consumed samples: 20951040 | consumed tokens: 42907729920 | elapsed time per iteration (s): 1.87 | learning rate: 4.103E-05 | global batch size: 512 | lm loss: 1.192146E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.105 | TFLOPs: 41.14 | 31: iteration 40930/ 52452 | consumed samples: 20956160 | consumed tokens: 42918215680 | elapsed time per iteration (s): 1.76 | learning rate: 4.100E-05 | global batch size: 512 | lm loss: 1.186838E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.769 | TFLOPs: 43.64 | 31: iteration 40940/ 52452 | consumed samples: 20961280 | consumed tokens: 42928701440 | elapsed time per iteration (s): 1.75 | learning rate: 4.096E-05 | global batch size: 512 | lm loss: 1.190169E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.617 | TFLOPs: 43.92 | 31: iteration 40950/ 52452 | consumed samples: 20966400 | consumed tokens: 42939187200 | elapsed time per iteration (s): 1.74 | learning rate: 4.093E-05 | global batch size: 512 | lm loss: 1.192748E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.039 | TFLOPs: 44.28 | 31: iteration 40960/ 52452 | consumed samples: 20971520 | consumed tokens: 42949672960 | elapsed time per iteration (s): 1.77 | learning rate: 4.089E-05 | global batch size: 512 | lm loss: 1.183661E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.134 | TFLOPs: 43.40 | 31: iteration 40970/ 52452 | consumed samples: 20976640 | consumed tokens: 42960158720 | elapsed time per iteration (s): 1.75 | learning rate: 4.086E-05 | global batch size: 512 | lm loss: 1.178744E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.914 | TFLOPs: 43.96 | 31: iteration 40980/ 52452 | consumed samples: 20981760 | consumed tokens: 42970644480 | elapsed time per iteration (s): 1.78 | learning rate: 4.082E-05 | global batch size: 512 | lm loss: 1.195413E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.377 | TFLOPs: 43.28 | 31: iteration 40990/ 52452 | consumed samples: 20986880 | consumed tokens: 42981130240 | elapsed time per iteration (s): 1.73 | learning rate: 4.079E-05 | global batch size: 512 | lm loss: 1.176484E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.628 | TFLOPs: 44.37 | 31: iteration 41000/ 52452 | consumed samples: 20992000 | consumed tokens: 42991616000 | elapsed time per iteration (s): 1.73 | learning rate: 4.075E-05 | global batch size: 512 | lm loss: 1.181860E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.059 | TFLOPs: 44.44 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41000 | lm loss value: 4.154788E+00 | lm loss PPL: 6.373848E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41000 | lm loss value: 4.029330E+00 | lm loss PPL: 5.622322E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41010/ 52452 | consumed samples: 20997120 | consumed tokens: 43002101760 | elapsed time per iteration (s): 9.16 | learning rate: 4.072E-05 | global batch size: 512 | lm loss: 1.184659E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.917 | TFLOPs: 8.39 | 31: iteration 41020/ 52452 | consumed samples: 21002240 | consumed tokens: 43012587520 | elapsed time per iteration (s): 1.74 | learning rate: 4.068E-05 | global batch size: 512 | lm loss: 1.203460E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.224 | TFLOPs: 44.16 | 31: iteration 41030/ 52452 | consumed samples: 21007360 | consumed tokens: 43023073280 | elapsed time per iteration (s): 1.77 | learning rate: 4.065E-05 | global batch size: 512 | lm loss: 1.193312E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.388 | TFLOPs: 43.44 | 31: iteration 41040/ 52452 | consumed samples: 21012480 | consumed tokens: 43033559040 | elapsed time per iteration (s): 1.76 | learning rate: 4.061E-05 | global batch size: 512 | lm loss: 1.194620E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.694 | TFLOPs: 43.63 | 31: iteration 41050/ 52452 | consumed samples: 21017600 | consumed tokens: 43044044800 | elapsed time per iteration (s): 1.75 | learning rate: 4.058E-05 | global batch size: 512 | lm loss: 1.179143E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.884 | TFLOPs: 43.81 | 31: iteration 41060/ 52452 | consumed samples: 21022720 | consumed tokens: 43054530560 | elapsed time per iteration (s): 1.77 | learning rate: 4.054E-05 | global batch size: 512 | lm loss: 1.193551E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.813 | TFLOPs: 43.35 | 31: iteration 41070/ 52452 | consumed samples: 21027840 | consumed tokens: 43065016320 | elapsed time per iteration (s): 1.77 | learning rate: 4.051E-05 | global batch size: 512 | lm loss: 1.195213E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.614 | TFLOPs: 43.32 | 31: iteration 41080/ 52452 | consumed samples: 21032960 | consumed tokens: 43075502080 | elapsed time per iteration (s): 1.73 | learning rate: 4.047E-05 | global batch size: 512 | lm loss: 1.183138E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.125 | TFLOPs: 44.30 | 31: iteration 41090/ 52452 | consumed samples: 21038080 | consumed tokens: 43085987840 | elapsed time per iteration (s): 1.80 | learning rate: 4.044E-05 | global batch size: 512 | lm loss: 1.196696E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.953 | TFLOPs: 42.77 | 31: iteration 41100/ 52452 | consumed samples: 21043200 | consumed tokens: 43096473600 | elapsed time per iteration (s): 1.74 | learning rate: 4.040E-05 | global batch size: 512 | lm loss: 1.172491E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.782 | TFLOPs: 44.10 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41100 | lm loss value: 4.117679E+00 | lm loss PPL: 6.141654E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41100 | lm loss value: 4.000982E+00 | lm loss PPL: 5.465178E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41110/ 52452 | consumed samples: 21048320 | consumed tokens: 43106959360 | elapsed time per iteration (s): 9.15 | learning rate: 4.037E-05 | global batch size: 512 | lm loss: 1.193842E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.963 | TFLOPs: 8.40 | 31: iteration 41120/ 52452 | consumed samples: 21053440 | consumed tokens: 43117445120 | elapsed time per iteration (s): 1.76 | learning rate: 4.034E-05 | global batch size: 512 | lm loss: 1.185495E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.796 | TFLOPs: 43.65 | 31: iteration 41130/ 52452 | consumed samples: 21058560 | consumed tokens: 43127930880 | elapsed time per iteration (s): 1.77 | learning rate: 4.030E-05 | global batch size: 512 | lm loss: 1.168825E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.856 | TFLOPs: 43.36 | 31: iteration 41140/ 52452 | consumed samples: 21063680 | consumed tokens: 43138416640 | elapsed time per iteration (s): 1.78 | learning rate: 4.027E-05 | global batch size: 512 | lm loss: 1.188054E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.856 | TFLOPs: 43.21 | 31: iteration 41150/ 52452 | consumed samples: 21068800 | consumed tokens: 43148902400 | elapsed time per iteration (s): 1.79 | learning rate: 4.023E-05 | global batch size: 512 | lm loss: 1.180621E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.572 | TFLOPs: 43.01 | 31: iteration 41160/ 52452 | consumed samples: 21073920 | consumed tokens: 43159388160 | elapsed time per iteration (s): 1.75 | learning rate: 4.020E-05 | global batch size: 512 | lm loss: 1.189488E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.556 | TFLOPs: 43.91 | 31: iteration 41170/ 52452 | consumed samples: 21079040 | consumed tokens: 43169873920 | elapsed time per iteration (s): 1.80 | learning rate: 4.016E-05 | global batch size: 512 | lm loss: 1.180702E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.096 | TFLOPs: 42.79 | 31: iteration 41180/ 52452 | consumed samples: 21084160 | consumed tokens: 43180359680 | elapsed time per iteration (s): 1.76 | learning rate: 4.013E-05 | global batch size: 512 | lm loss: 1.200595E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.609 | TFLOPs: 43.77 | 31: iteration 41190/ 52452 | consumed samples: 21089280 | consumed tokens: 43190845440 | elapsed time per iteration (s): 1.75 | learning rate: 4.009E-05 | global batch size: 512 | lm loss: 1.193228E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.960 | TFLOPs: 43.82 | 31: iteration 41200/ 52452 | consumed samples: 21094400 | consumed tokens: 43201331200 | elapsed time per iteration (s): 1.76 | learning rate: 4.006E-05 | global batch size: 512 | lm loss: 1.199368E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.704 | TFLOPs: 43.63 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41200 | lm loss value: 4.275016E+00 | lm loss PPL: 7.188128E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41200 | lm loss value: 4.114557E+00 | lm loss PPL: 6.122507E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41210/ 52452 | consumed samples: 21099520 | consumed tokens: 43211816960 | elapsed time per iteration (s): 9.18 | learning rate: 4.003E-05 | global batch size: 512 | lm loss: 1.179839E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.753 | TFLOPs: 8.37 | 31: iteration 41220/ 52452 | consumed samples: 21104640 | consumed tokens: 43222302720 | elapsed time per iteration (s): 1.85 | learning rate: 3.999E-05 | global batch size: 512 | lm loss: 1.179128E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.159 | TFLOPs: 41.60 | 31: iteration 41230/ 52452 | consumed samples: 21109760 | consumed tokens: 43232788480 | elapsed time per iteration (s): 1.76 | learning rate: 3.996E-05 | global batch size: 512 | lm loss: 1.190713E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.176 | TFLOPs: 43.70 | 31: iteration 41240/ 52452 | consumed samples: 21114880 | consumed tokens: 43243274240 | elapsed time per iteration (s): 1.82 | learning rate: 3.992E-05 | global batch size: 512 | lm loss: 1.191717E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.545 | TFLOPs: 42.26 | 31: iteration 41250/ 52452 | consumed samples: 21120000 | consumed tokens: 43253760000 | elapsed time per iteration (s): 1.80 | learning rate: 3.989E-05 | global batch size: 512 | lm loss: 1.180264E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.916 | TFLOPs: 42.61 | 31: iteration 41260/ 52452 | consumed samples: 21125120 | consumed tokens: 43264245760 | elapsed time per iteration (s): 1.78 | learning rate: 3.986E-05 | global batch size: 512 | lm loss: 1.178031E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.220 | TFLOPs: 43.11 | 31: iteration 41270/ 52452 | consumed samples: 21130240 | consumed tokens: 43274731520 | elapsed time per iteration (s): 1.81 | learning rate: 3.982E-05 | global batch size: 512 | lm loss: 1.179144E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.440 | TFLOPs: 42.54 | 31: iteration 41280/ 52452 | consumed samples: 21135360 | consumed tokens: 43285217280 | elapsed time per iteration (s): 1.83 | learning rate: 3.979E-05 | global batch size: 512 | lm loss: 1.178505E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.168 | TFLOPs: 42.05 | 31: iteration 41290/ 52452 | consumed samples: 21140480 | consumed tokens: 43295703040 | elapsed time per iteration (s): 1.81 | learning rate: 3.975E-05 | global batch size: 512 | lm loss: 1.171261E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.607 | TFLOPs: 42.42 | 31: iteration 41300/ 52452 | consumed samples: 21145600 | consumed tokens: 43306188800 | elapsed time per iteration (s): 1.78 | learning rate: 3.972E-05 | global batch size: 512 | lm loss: 1.190712E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.882 | TFLOPs: 43.06 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41300 | lm loss value: 4.126954E+00 | lm loss PPL: 6.198879E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41300 | lm loss value: 4.009053E+00 | lm loss PPL: 5.509468E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41310/ 52452 | consumed samples: 21150720 | consumed tokens: 43316674560 | elapsed time per iteration (s): 9.18 | learning rate: 3.969E-05 | global batch size: 512 | lm loss: 1.176717E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.798 | TFLOPs: 8.37 | 31: iteration 41320/ 52452 | consumed samples: 21155840 | consumed tokens: 43327160320 | elapsed time per iteration (s): 1.81 | learning rate: 3.965E-05 | global batch size: 512 | lm loss: 1.185887E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.394 | TFLOPs: 42.54 | 31: iteration 41330/ 52452 | consumed samples: 21160960 | consumed tokens: 43337646080 | elapsed time per iteration (s): 1.81 | learning rate: 3.962E-05 | global batch size: 512 | lm loss: 1.185717E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.135 | TFLOPs: 42.35 | 31: iteration 41340/ 52452 | consumed samples: 21166080 | consumed tokens: 43348131840 | elapsed time per iteration (s): 1.74 | learning rate: 3.958E-05 | global batch size: 512 | lm loss: 1.186396E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.240 | TFLOPs: 44.16 | 31: iteration 41350/ 52452 | consumed samples: 21171200 | consumed tokens: 43358617600 | elapsed time per iteration (s): 1.73 | learning rate: 3.955E-05 | global batch size: 512 | lm loss: 1.187765E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.628 | TFLOPs: 44.37 | 31: iteration 41360/ 52452 | consumed samples: 21176320 | consumed tokens: 43369103360 | elapsed time per iteration (s): 1.79 | learning rate: 3.952E-05 | global batch size: 512 | lm loss: 1.187623E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.619 | TFLOPs: 43.02 | 31: iteration 41370/ 52452 | consumed samples: 21181440 | consumed tokens: 43379589120 | elapsed time per iteration (s): 1.72 | learning rate: 3.948E-05 | global batch size: 512 | lm loss: 1.187565E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.619 | TFLOPs: 44.67 | 31: iteration 41380/ 52452 | consumed samples: 21186560 | consumed tokens: 43390074880 | elapsed time per iteration (s): 1.74 | learning rate: 3.945E-05 | global batch size: 512 | lm loss: 1.171928E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.437 | TFLOPs: 44.19 | 31: iteration 41390/ 52452 | consumed samples: 21191680 | consumed tokens: 43400560640 | elapsed time per iteration (s): 1.74 | learning rate: 3.941E-05 | global batch size: 512 | lm loss: 1.169258E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.258 | TFLOPs: 44.17 | 31: iteration 41400/ 52452 | consumed samples: 21196800 | consumed tokens: 43411046400 | elapsed time per iteration (s): 1.78 | learning rate: 3.938E-05 | global batch size: 512 | lm loss: 1.181040E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.994 | TFLOPs: 43.08 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41400 | lm loss value: 4.278537E+00 | lm loss PPL: 7.213485E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41400 | lm loss value: 4.116100E+00 | lm loss PPL: 6.131962E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41410/ 52452 | consumed samples: 21201920 | consumed tokens: 43421532160 | elapsed time per iteration (s): 9.22 | learning rate: 3.935E-05 | global batch size: 512 | lm loss: 1.185198E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.519 | TFLOPs: 8.33 | 31: iteration 41420/ 52452 | consumed samples: 21207040 | consumed tokens: 43432017920 | elapsed time per iteration (s): 1.76 | learning rate: 3.931E-05 | global batch size: 512 | lm loss: 1.189502E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.763 | TFLOPs: 43.64 | 31: iteration 41430/ 52452 | consumed samples: 21212160 | consumed tokens: 43442503680 | elapsed time per iteration (s): 1.84 | learning rate: 3.928E-05 | global batch size: 512 | lm loss: 1.181048E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.789 | TFLOPs: 41.84 | 31: iteration 41440/ 52452 | consumed samples: 21217280 | consumed tokens: 43452989440 | elapsed time per iteration (s): 1.77 | learning rate: 3.925E-05 | global batch size: 512 | lm loss: 1.188480E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.662 | TFLOPs: 43.33 | 31: iteration 41450/ 52452 | consumed samples: 21222400 | consumed tokens: 43463475200 | elapsed time per iteration (s): 1.75 | learning rate: 3.921E-05 | global batch size: 512 | lm loss: 1.188188E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.330 | TFLOPs: 44.03 | 31: iteration 41460/ 52452 | consumed samples: 21227520 | consumed tokens: 43473960960 | elapsed time per iteration (s): 1.73 | learning rate: 3.918E-05 | global batch size: 512 | lm loss: 1.191239E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.338 | TFLOPs: 44.33 | 31: iteration 41470/ 52452 | consumed samples: 21232640 | consumed tokens: 43484446720 | elapsed time per iteration (s): 1.73 | learning rate: 3.914E-05 | global batch size: 512 | lm loss: 1.182350E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.535 | TFLOPs: 44.51 | 31: iteration 41480/ 52452 | consumed samples: 21237760 | consumed tokens: 43494932480 | elapsed time per iteration (s): 1.77 | learning rate: 3.911E-05 | global batch size: 512 | lm loss: 1.192414E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.056 | TFLOPs: 43.39 | 31: iteration 41490/ 52452 | consumed samples: 21242880 | consumed tokens: 43505418240 | elapsed time per iteration (s): 1.76 | learning rate: 3.908E-05 | global batch size: 512 | lm loss: 1.187640E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.673 | TFLOPs: 43.63 | 31: iteration 41500/ 52452 | consumed samples: 21248000 | consumed tokens: 43515904000 | elapsed time per iteration (s): 1.74 | learning rate: 3.904E-05 | global batch size: 512 | lm loss: 1.181487E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.377 | TFLOPs: 44.18 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41500 | lm loss value: 4.244998E+00 | lm loss PPL: 6.975565E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41500 | lm loss value: 4.106353E+00 | lm loss PPL: 6.072484E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41510/ 52452 | consumed samples: 21253120 | consumed tokens: 43526389760 | elapsed time per iteration (s): 9.13 | learning rate: 3.901E-05 | global batch size: 512 | lm loss: 1.184965E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.081 | TFLOPs: 8.42 | 31: iteration 41520/ 52452 | consumed samples: 21258240 | consumed tokens: 43536875520 | elapsed time per iteration (s): 1.73 | learning rate: 3.898E-05 | global batch size: 512 | lm loss: 1.184957E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.245 | TFLOPs: 44.46 | 31: iteration 41530/ 52452 | consumed samples: 21263360 | consumed tokens: 43547361280 | elapsed time per iteration (s): 1.74 | learning rate: 3.894E-05 | global batch size: 512 | lm loss: 1.190986E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.080 | TFLOPs: 44.14 | 31: iteration 41540/ 52452 | consumed samples: 21268480 | consumed tokens: 43557847040 | elapsed time per iteration (s): 1.78 | learning rate: 3.891E-05 | global batch size: 512 | lm loss: 1.186792E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.089 | TFLOPs: 43.24 | 31: iteration 41550/ 52452 | consumed samples: 21273600 | consumed tokens: 43568332800 | elapsed time per iteration (s): 1.73 | learning rate: 3.888E-05 | global batch size: 512 | lm loss: 1.180440E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.652 | TFLOPs: 44.53 | 31: iteration 41560/ 52452 | consumed samples: 21278720 | consumed tokens: 43578818560 | elapsed time per iteration (s): 1.74 | learning rate: 3.884E-05 | global batch size: 512 | lm loss: 1.188287E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.926 | TFLOPs: 44.12 | 31: iteration 41570/ 52452 | consumed samples: 21283840 | consumed tokens: 43589304320 | elapsed time per iteration (s): 1.73 | learning rate: 3.881E-05 | global batch size: 512 | lm loss: 1.177285E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.605 | TFLOPs: 44.37 | 31: iteration 41580/ 52452 | consumed samples: 21288960 | consumed tokens: 43599790080 | elapsed time per iteration (s): 1.73 | learning rate: 3.878E-05 | global batch size: 512 | lm loss: 1.186906E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.309 | TFLOPs: 44.32 | 31: iteration 41590/ 52452 | consumed samples: 21294080 | consumed tokens: 43610275840 | elapsed time per iteration (s): 1.73 | learning rate: 3.874E-05 | global batch size: 512 | lm loss: 1.180683E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.788 | TFLOPs: 44.55 | 31: iteration 41600/ 52452 | consumed samples: 21299200 | consumed tokens: 43620761600 | elapsed time per iteration (s): 1.75 | learning rate: 3.871E-05 | global batch size: 512 | lm loss: 1.181091E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.156 | TFLOPs: 44.00 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41600 | lm loss value: 4.178495E+00 | lm loss PPL: 6.526758E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41600 | lm loss value: 4.048003E+00 | lm loss PPL: 5.728296E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41610/ 52452 | consumed samples: 21304320 | consumed tokens: 43631247360 | elapsed time per iteration (s): 9.15 | learning rate: 3.868E-05 | global batch size: 512 | lm loss: 1.183843E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.974 | TFLOPs: 8.40 | 31: iteration 41620/ 52452 | consumed samples: 21309440 | consumed tokens: 43641733120 | elapsed time per iteration (s): 1.72 | learning rate: 3.864E-05 | global batch size: 512 | lm loss: 1.185964E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.108 | TFLOPs: 44.59 | 31: iteration 41630/ 52452 | consumed samples: 21314560 | consumed tokens: 43652218880 | elapsed time per iteration (s): 1.81 | learning rate: 3.861E-05 | global batch size: 512 | lm loss: 1.188821E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.252 | TFLOPs: 42.51 | 31: iteration 41640/ 52452 | consumed samples: 21319680 | consumed tokens: 43662704640 | elapsed time per iteration (s): 1.77 | learning rate: 3.858E-05 | global batch size: 512 | lm loss: 1.184571E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.846 | TFLOPs: 43.50 | 31: iteration 41650/ 52452 | consumed samples: 21324800 | consumed tokens: 43673190400 | elapsed time per iteration (s): 1.80 | learning rate: 3.854E-05 | global batch size: 512 | lm loss: 1.188527E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.913 | TFLOPs: 42.61 | 31: iteration 41660/ 52452 | consumed samples: 21329920 | consumed tokens: 43683676160 | elapsed time per iteration (s): 1.74 | learning rate: 3.851E-05 | global batch size: 512 | lm loss: 1.189196E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.957 | TFLOPs: 44.27 | 31: iteration 41670/ 52452 | consumed samples: 21335040 | consumed tokens: 43694161920 | elapsed time per iteration (s): 1.73 | learning rate: 3.848E-05 | global batch size: 512 | lm loss: 1.191094E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.187 | TFLOPs: 44.31 | 31: iteration 41680/ 52452 | consumed samples: 21340160 | consumed tokens: 43704647680 | elapsed time per iteration (s): 1.80 | learning rate: 3.845E-05 | global batch size: 512 | lm loss: 1.183525E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.314 | TFLOPs: 42.67 | 31: iteration 41690/ 52452 | consumed samples: 21345280 | consumed tokens: 43715133440 | elapsed time per iteration (s): 1.79 | learning rate: 3.841E-05 | global batch size: 512 | lm loss: 1.179877E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.396 | TFLOPs: 42.84 | 31: iteration 41700/ 52452 | consumed samples: 21350400 | consumed tokens: 43725619200 | elapsed time per iteration (s): 1.73 | learning rate: 3.838E-05 | global batch size: 512 | lm loss: 1.176398E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.666 | TFLOPs: 44.38 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41700 | lm loss value: 4.165393E+00 | lm loss PPL: 6.441802E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41700 | lm loss value: 4.033140E+00 | lm loss PPL: 5.643786E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41710/ 52452 | consumed samples: 21355520 | consumed tokens: 43736104960 | elapsed time per iteration (s): 9.17 | learning rate: 3.835E-05 | global batch size: 512 | lm loss: 1.185840E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.836 | TFLOPs: 8.38 | 31: iteration 41720/ 52452 | consumed samples: 21360640 | consumed tokens: 43746590720 | elapsed time per iteration (s): 1.80 | learning rate: 3.831E-05 | global batch size: 512 | lm loss: 1.182393E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.811 | TFLOPs: 42.75 | 31: iteration 41730/ 52452 | consumed samples: 21365760 | consumed tokens: 43757076480 | elapsed time per iteration (s): 1.75 | learning rate: 3.828E-05 | global batch size: 512 | lm loss: 1.176104E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.203 | TFLOPs: 43.86 | 31: iteration 41740/ 52452 | consumed samples: 21370880 | consumed tokens: 43767562240 | elapsed time per iteration (s): 1.80 | learning rate: 3.825E-05 | global batch size: 512 | lm loss: 1.177873E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.087 | TFLOPs: 42.79 | 31: iteration 41750/ 52452 | consumed samples: 21376000 | consumed tokens: 43778048000 | elapsed time per iteration (s): 1.76 | learning rate: 3.821E-05 | global batch size: 512 | lm loss: 1.180196E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.248 | TFLOPs: 43.71 | 31: iteration 41760/ 52452 | consumed samples: 21381120 | consumed tokens: 43788533760 | elapsed time per iteration (s): 1.78 | learning rate: 3.818E-05 | global batch size: 512 | lm loss: 1.174637E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.182 | TFLOPs: 43.10 | 31: iteration 41770/ 52452 | consumed samples: 21386240 | consumed tokens: 43799019520 | elapsed time per iteration (s): 1.73 | learning rate: 3.815E-05 | global batch size: 512 | lm loss: 1.172350E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.227 | TFLOPs: 44.46 | 31: iteration 41780/ 52452 | consumed samples: 21391360 | consumed tokens: 43809505280 | elapsed time per iteration (s): 1.78 | learning rate: 3.812E-05 | global batch size: 512 | lm loss: 1.182333E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.392 | TFLOPs: 43.29 | 31: iteration 41790/ 52452 | consumed samples: 21396480 | consumed tokens: 43819991040 | elapsed time per iteration (s): 1.83 | learning rate: 3.808E-05 | global batch size: 512 | lm loss: 1.193544E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.641 | TFLOPs: 41.97 | 31: iteration 41800/ 52452 | consumed samples: 21401600 | consumed tokens: 43830476800 | elapsed time per iteration (s): 1.81 | learning rate: 3.805E-05 | global batch size: 512 | lm loss: 1.185716E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.049 | TFLOPs: 42.48 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41800 | lm loss value: 4.222089E+00 | lm loss PPL: 6.817577E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41800 | lm loss value: 4.072062E+00 | lm loss PPL: 5.867786E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41810/ 52452 | consumed samples: 21406720 | consumed tokens: 43840962560 | elapsed time per iteration (s): 9.19 | learning rate: 3.802E-05 | global batch size: 512 | lm loss: 1.192971E+00 | grad norm: 0.192 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.686 | TFLOPs: 8.36 | 31: iteration 41820/ 52452 | consumed samples: 21411840 | consumed tokens: 43851448320 | elapsed time per iteration (s): 1.73 | learning rate: 3.799E-05 | global batch size: 512 | lm loss: 1.161385E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.580 | TFLOPs: 44.36 | 31: iteration 41830/ 52452 | consumed samples: 21416960 | consumed tokens: 43861934080 | elapsed time per iteration (s): 1.74 | learning rate: 3.795E-05 | global batch size: 512 | lm loss: 1.179924E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.681 | TFLOPs: 44.23 | 31: iteration 41840/ 52452 | consumed samples: 21422080 | consumed tokens: 43872419840 | elapsed time per iteration (s): 1.81 | learning rate: 3.792E-05 | global batch size: 512 | lm loss: 1.176282E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.742 | TFLOPs: 42.44 | 31: iteration 41850/ 52452 | consumed samples: 21427200 | consumed tokens: 43882905600 | elapsed time per iteration (s): 1.75 | learning rate: 3.789E-05 | global batch size: 512 | lm loss: 1.181492E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.787 | TFLOPs: 43.95 | 31: iteration 41860/ 52452 | consumed samples: 21432320 | consumed tokens: 43893391360 | elapsed time per iteration (s): 1.79 | learning rate: 3.786E-05 | global batch size: 512 | lm loss: 1.183772E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.777 | TFLOPs: 43.04 | 31: iteration 41870/ 52452 | consumed samples: 21437440 | consumed tokens: 43903877120 | elapsed time per iteration (s): 1.83 | learning rate: 3.782E-05 | global batch size: 512 | lm loss: 1.183307E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.393 | TFLOPs: 42.09 | 31: iteration 41880/ 52452 | consumed samples: 21442560 | consumed tokens: 43914362880 | elapsed time per iteration (s): 1.76 | learning rate: 3.779E-05 | global batch size: 512 | lm loss: 1.187568E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.156 | TFLOPs: 43.55 | 31: iteration 41890/ 52452 | consumed samples: 21447680 | consumed tokens: 43924848640 | elapsed time per iteration (s): 1.77 | learning rate: 3.776E-05 | global batch size: 512 | lm loss: 1.176761E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.813 | TFLOPs: 43.50 | 31: iteration 41900/ 52452 | consumed samples: 21452800 | consumed tokens: 43935334400 | elapsed time per iteration (s): 1.74 | learning rate: 3.773E-05 | global batch size: 512 | lm loss: 1.182290E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.980 | TFLOPs: 44.12 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 41900 | lm loss value: 4.166005E+00 | lm loss PPL: 6.445744E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 41900 | lm loss value: 4.036713E+00 | lm loss PPL: 5.663987E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 41910/ 52452 | consumed samples: 21457920 | consumed tokens: 43945820160 | elapsed time per iteration (s): 9.15 | learning rate: 3.769E-05 | global batch size: 512 | lm loss: 1.179959E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.980 | TFLOPs: 8.40 | 31: iteration 41920/ 52452 | consumed samples: 21463040 | consumed tokens: 43956305920 | elapsed time per iteration (s): 1.78 | learning rate: 3.766E-05 | global batch size: 512 | lm loss: 1.179553E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.368 | TFLOPs: 43.28 | 31: iteration 41930/ 52452 | consumed samples: 21468160 | consumed tokens: 43966791680 | elapsed time per iteration (s): 1.73 | learning rate: 3.763E-05 | global batch size: 512 | lm loss: 1.162522E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.197 | TFLOPs: 44.31 | 31: iteration 41940/ 52452 | consumed samples: 21473280 | consumed tokens: 43977277440 | elapsed time per iteration (s): 1.74 | learning rate: 3.760E-05 | global batch size: 512 | lm loss: 1.180683E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.880 | TFLOPs: 44.26 | 31: iteration 41950/ 52452 | consumed samples: 21478400 | consumed tokens: 43987763200 | elapsed time per iteration (s): 1.73 | learning rate: 3.756E-05 | global batch size: 512 | lm loss: 1.178328E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.461 | TFLOPs: 44.35 | 31: iteration 41960/ 52452 | consumed samples: 21483520 | consumed tokens: 43998248960 | elapsed time per iteration (s): 1.73 | learning rate: 3.753E-05 | global batch size: 512 | lm loss: 1.188448E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.553 | TFLOPs: 44.36 | 31: iteration 41970/ 52452 | consumed samples: 21488640 | consumed tokens: 44008734720 | elapsed time per iteration (s): 1.78 | learning rate: 3.750E-05 | global batch size: 512 | lm loss: 1.177817E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.422 | TFLOPs: 43.14 | 31: iteration 41980/ 52452 | consumed samples: 21493760 | consumed tokens: 44019220480 | elapsed time per iteration (s): 1.85 | learning rate: 3.747E-05 | global batch size: 512 | lm loss: 1.170754E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.285 | TFLOPs: 41.47 | 31: iteration 41990/ 52452 | consumed samples: 21498880 | consumed tokens: 44029706240 | elapsed time per iteration (s): 1.74 | learning rate: 3.743E-05 | global batch size: 512 | lm loss: 1.184455E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.255 | TFLOPs: 44.17 | 0: [2022-12-26 14:31:26,690] [INFO] [logging.py:68:log_dist] [Rank 0] step=42000, skipped=0, lr=[3.7402130241130454e-05, 3.7402130241130454e-05, 3.7402130241130454e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 42000/ 52452 | consumed samples: 21504000 | consumed tokens: 44040192000 | elapsed time per iteration (s): 1.75 | learning rate: 3.740E-05 | global batch size: 512 | lm loss: 1.188441E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.614 | TFLOPs: 43.92 | 0: steps: 42000 loss: 1.2009 iter time (s): 1.768 samples/sec: 289.606 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42000 | lm loss value: 4.273733E+00 | lm loss PPL: 7.178910E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42000 | lm loss value: 4.119667E+00 | lm loss PPL: 6.153872E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42010/ 52452 | consumed samples: 21509120 | consumed tokens: 44050677760 | elapsed time per iteration (s): 9.49 | learning rate: 3.737E-05 | global batch size: 512 | lm loss: 1.202041E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 53.979 | TFLOPs: 8.10 | 31: iteration 42020/ 52452 | consumed samples: 21514240 | consumed tokens: 44061163520 | elapsed time per iteration (s): 1.82 | learning rate: 3.734E-05 | global batch size: 512 | lm loss: 1.185008E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.660 | TFLOPs: 42.13 | 31: iteration 42030/ 52452 | consumed samples: 21519360 | consumed tokens: 44071649280 | elapsed time per iteration (s): 1.75 | learning rate: 3.731E-05 | global batch size: 512 | lm loss: 1.180489E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.468 | TFLOPs: 43.90 | 31: iteration 42040/ 52452 | consumed samples: 21524480 | consumed tokens: 44082135040 | elapsed time per iteration (s): 1.78 | learning rate: 3.727E-05 | global batch size: 512 | lm loss: 1.169166E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.807 | TFLOPs: 43.20 | 31: iteration 42050/ 52452 | consumed samples: 21529600 | consumed tokens: 44092620800 | elapsed time per iteration (s): 1.78 | learning rate: 3.724E-05 | global batch size: 512 | lm loss: 1.181005E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.024 | TFLOPs: 43.08 | 31: iteration 42060/ 52452 | consumed samples: 21534720 | consumed tokens: 44103106560 | elapsed time per iteration (s): 1.81 | learning rate: 3.721E-05 | global batch size: 512 | lm loss: 1.171379E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.470 | TFLOPs: 42.40 | 31: iteration 42070/ 52452 | consumed samples: 21539840 | consumed tokens: 44113592320 | elapsed time per iteration (s): 1.75 | learning rate: 3.718E-05 | global batch size: 512 | lm loss: 1.170020E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.613 | TFLOPs: 43.92 | 31: iteration 42080/ 52452 | consumed samples: 21544960 | consumed tokens: 44124078080 | elapsed time per iteration (s): 1.76 | learning rate: 3.715E-05 | global batch size: 512 | lm loss: 1.192340E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.697 | TFLOPs: 43.78 | 31: iteration 42090/ 52452 | consumed samples: 21550080 | consumed tokens: 44134563840 | elapsed time per iteration (s): 1.80 | learning rate: 3.711E-05 | global batch size: 512 | lm loss: 1.181891E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.667 | TFLOPs: 42.58 | 31: iteration 42100/ 52452 | consumed samples: 21555200 | consumed tokens: 44145049600 | elapsed time per iteration (s): 1.76 | learning rate: 3.708E-05 | global batch size: 512 | lm loss: 1.176570E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.188 | TFLOPs: 43.56 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42100 | lm loss value: 4.215034E+00 | lm loss PPL: 6.769650E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42100 | lm loss value: 4.087624E+00 | lm loss PPL: 5.959812E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42110/ 52452 | consumed samples: 21560320 | consumed tokens: 44155535360 | elapsed time per iteration (s): 9.11 | learning rate: 3.705E-05 | global batch size: 512 | lm loss: 1.185880E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.176 | TFLOPs: 8.43 | 31: iteration 42120/ 52452 | consumed samples: 21565440 | consumed tokens: 44166021120 | elapsed time per iteration (s): 1.77 | learning rate: 3.702E-05 | global batch size: 512 | lm loss: 1.178687E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.211 | TFLOPs: 43.41 | 31: iteration 42130/ 52452 | consumed samples: 21570560 | consumed tokens: 44176506880 | elapsed time per iteration (s): 1.74 | learning rate: 3.699E-05 | global batch size: 512 | lm loss: 1.181062E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.890 | TFLOPs: 44.26 | 31: iteration 42140/ 52452 | consumed samples: 21575680 | consumed tokens: 44186992640 | elapsed time per iteration (s): 1.76 | learning rate: 3.695E-05 | global batch size: 512 | lm loss: 1.174259E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.190 | TFLOPs: 43.71 | 31: iteration 42150/ 52452 | consumed samples: 21580800 | consumed tokens: 44197478400 | elapsed time per iteration (s): 1.75 | learning rate: 3.692E-05 | global batch size: 512 | lm loss: 1.185030E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.389 | TFLOPs: 44.04 | 31: iteration 42160/ 52452 | consumed samples: 21585920 | consumed tokens: 44207964160 | elapsed time per iteration (s): 1.80 | learning rate: 3.689E-05 | global batch size: 512 | lm loss: 1.170113E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.749 | TFLOPs: 42.74 | 31: iteration 42170/ 52452 | consumed samples: 21591040 | consumed tokens: 44218449920 | elapsed time per iteration (s): 1.78 | learning rate: 3.686E-05 | global batch size: 512 | lm loss: 1.180798E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.244 | TFLOPs: 43.26 | 31: iteration 42180/ 52452 | consumed samples: 21596160 | consumed tokens: 44228935680 | elapsed time per iteration (s): 1.88 | learning rate: 3.683E-05 | global batch size: 512 | lm loss: 1.181644E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.219 | TFLOPs: 40.86 | 31: iteration 42190/ 52452 | consumed samples: 21601280 | consumed tokens: 44239421440 | elapsed time per iteration (s): 1.78 | learning rate: 3.680E-05 | global batch size: 512 | lm loss: 1.175802E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.218 | TFLOPs: 43.26 | 31: iteration 42200/ 52452 | consumed samples: 21606400 | consumed tokens: 44249907200 | elapsed time per iteration (s): 1.79 | learning rate: 3.676E-05 | global batch size: 512 | lm loss: 1.190332E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.493 | TFLOPs: 43.00 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42200 | lm loss value: 4.190632E+00 | lm loss PPL: 6.606452E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42200 | lm loss value: 4.057534E+00 | lm loss PPL: 5.783151E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42210/ 52452 | consumed samples: 21611520 | consumed tokens: 44260392960 | elapsed time per iteration (s): 9.13 | learning rate: 3.673E-05 | global batch size: 512 | lm loss: 1.183806E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.098 | TFLOPs: 8.42 | 31: iteration 42220/ 52452 | consumed samples: 21616640 | consumed tokens: 44270878720 | elapsed time per iteration (s): 1.77 | learning rate: 3.670E-05 | global batch size: 512 | lm loss: 1.181224E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.018 | TFLOPs: 43.38 | 31: iteration 42230/ 52452 | consumed samples: 21621760 | consumed tokens: 44281364480 | elapsed time per iteration (s): 1.75 | learning rate: 3.667E-05 | global batch size: 512 | lm loss: 1.178651E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.560 | TFLOPs: 43.91 | 31: iteration 42240/ 52452 | consumed samples: 21626880 | consumed tokens: 44291850240 | elapsed time per iteration (s): 1.78 | learning rate: 3.664E-05 | global batch size: 512 | lm loss: 1.179967E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.264 | TFLOPs: 43.27 | 31: iteration 42250/ 52452 | consumed samples: 21632000 | consumed tokens: 44302336000 | elapsed time per iteration (s): 1.73 | learning rate: 3.661E-05 | global batch size: 512 | lm loss: 1.171756E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.493 | TFLOPs: 44.35 | 31: iteration 42260/ 52452 | consumed samples: 21637120 | consumed tokens: 44312821760 | elapsed time per iteration (s): 1.76 | learning rate: 3.657E-05 | global batch size: 512 | lm loss: 1.194762E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.582 | TFLOPs: 43.61 | 31: iteration 42270/ 52452 | consumed samples: 21642240 | consumed tokens: 44323307520 | elapsed time per iteration (s): 1.79 | learning rate: 3.654E-05 | global batch size: 512 | lm loss: 1.182897E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.240 | TFLOPs: 42.81 | 31: iteration 42280/ 52452 | consumed samples: 21647360 | consumed tokens: 44333793280 | elapsed time per iteration (s): 1.79 | learning rate: 3.651E-05 | global batch size: 512 | lm loss: 1.170905E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.430 | TFLOPs: 42.99 | 31: iteration 42290/ 52452 | consumed samples: 21652480 | consumed tokens: 44344279040 | elapsed time per iteration (s): 1.75 | learning rate: 3.648E-05 | global batch size: 512 | lm loss: 1.176115E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.387 | TFLOPs: 43.89 | 31: iteration 42300/ 52452 | consumed samples: 21657600 | consumed tokens: 44354764800 | elapsed time per iteration (s): 1.75 | learning rate: 3.645E-05 | global batch size: 512 | lm loss: 1.167896E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.179 | TFLOPs: 43.85 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42300 | lm loss value: 4.298507E+00 | lm loss PPL: 7.358982E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42300 | lm loss value: 4.134445E+00 | lm loss PPL: 6.245493E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42310/ 52452 | consumed samples: 21662720 | consumed tokens: 44365250560 | elapsed time per iteration (s): 9.12 | learning rate: 3.642E-05 | global batch size: 512 | lm loss: 1.179364E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.112 | TFLOPs: 8.42 | 31: iteration 42320/ 52452 | consumed samples: 21667840 | consumed tokens: 44375736320 | elapsed time per iteration (s): 1.75 | learning rate: 3.639E-05 | global batch size: 512 | lm loss: 1.192290E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.244 | TFLOPs: 44.01 | 31: iteration 42330/ 52452 | consumed samples: 21672960 | consumed tokens: 44386222080 | elapsed time per iteration (s): 1.77 | learning rate: 3.635E-05 | global batch size: 512 | lm loss: 1.174635E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.626 | TFLOPs: 43.47 | 31: iteration 42340/ 52452 | consumed samples: 21678080 | consumed tokens: 44396707840 | elapsed time per iteration (s): 1.81 | learning rate: 3.632E-05 | global batch size: 512 | lm loss: 1.178985E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.021 | TFLOPs: 42.48 | 31: iteration 42350/ 52452 | consumed samples: 21683200 | consumed tokens: 44407193600 | elapsed time per iteration (s): 1.78 | learning rate: 3.629E-05 | global batch size: 512 | lm loss: 1.171912E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.414 | TFLOPs: 43.14 | 31: iteration 42360/ 52452 | consumed samples: 21688320 | consumed tokens: 44417679360 | elapsed time per iteration (s): 1.78 | learning rate: 3.626E-05 | global batch size: 512 | lm loss: 1.194811E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.304 | TFLOPs: 43.27 | 31: iteration 42370/ 52452 | consumed samples: 21693440 | consumed tokens: 44428165120 | elapsed time per iteration (s): 1.73 | learning rate: 3.623E-05 | global batch size: 512 | lm loss: 1.182304E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.174 | TFLOPs: 44.45 | 31: iteration 42380/ 52452 | consumed samples: 21698560 | consumed tokens: 44438650880 | elapsed time per iteration (s): 1.75 | learning rate: 3.620E-05 | global batch size: 512 | lm loss: 1.177526E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.745 | TFLOPs: 43.94 | 31: iteration 42390/ 52452 | consumed samples: 21703680 | consumed tokens: 44449136640 | elapsed time per iteration (s): 1.73 | learning rate: 3.617E-05 | global batch size: 512 | lm loss: 1.179128E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.634 | TFLOPs: 44.37 | 31: iteration 42400/ 52452 | consumed samples: 21708800 | consumed tokens: 44459622400 | elapsed time per iteration (s): 1.76 | learning rate: 3.614E-05 | global batch size: 512 | lm loss: 1.167807E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.829 | TFLOPs: 43.65 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42400 | lm loss value: 4.166715E+00 | lm loss PPL: 6.450319E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42400 | lm loss value: 4.044919E+00 | lm loss PPL: 5.710653E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42410/ 52452 | consumed samples: 21713920 | consumed tokens: 44470108160 | elapsed time per iteration (s): 9.13 | learning rate: 3.611E-05 | global batch size: 512 | lm loss: 1.178789E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.055 | TFLOPs: 8.41 | 31: iteration 42420/ 52452 | consumed samples: 21719040 | consumed tokens: 44480593920 | elapsed time per iteration (s): 1.77 | learning rate: 3.607E-05 | global batch size: 512 | lm loss: 1.174244E+00 | grad norm: 0.177 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.021 | TFLOPs: 43.53 | 31: iteration 42430/ 52452 | consumed samples: 21724160 | consumed tokens: 44491079680 | elapsed time per iteration (s): 1.75 | learning rate: 3.604E-05 | global batch size: 512 | lm loss: 1.171266E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.321 | TFLOPs: 43.88 | 31: iteration 42440/ 52452 | consumed samples: 21729280 | consumed tokens: 44501565440 | elapsed time per iteration (s): 1.75 | learning rate: 3.601E-05 | global batch size: 512 | lm loss: 1.170448E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.123 | TFLOPs: 43.85 | 31: iteration 42450/ 52452 | consumed samples: 21734400 | consumed tokens: 44512051200 | elapsed time per iteration (s): 1.77 | learning rate: 3.598E-05 | global batch size: 512 | lm loss: 1.177431E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.119 | TFLOPs: 43.40 | 31: iteration 42460/ 52452 | consumed samples: 21739520 | consumed tokens: 44522536960 | elapsed time per iteration (s): 1.83 | learning rate: 3.595E-05 | global batch size: 512 | lm loss: 1.173390E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.138 | TFLOPs: 42.05 | 31: iteration 42470/ 52452 | consumed samples: 21744640 | consumed tokens: 44533022720 | elapsed time per iteration (s): 1.78 | learning rate: 3.592E-05 | global batch size: 512 | lm loss: 1.163366E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.133 | TFLOPs: 43.25 | 31: iteration 42480/ 52452 | consumed samples: 21749760 | consumed tokens: 44543508480 | elapsed time per iteration (s): 1.78 | learning rate: 3.589E-05 | global batch size: 512 | lm loss: 1.172795E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.936 | TFLOPs: 43.07 | 31: iteration 42490/ 52452 | consumed samples: 21754880 | consumed tokens: 44553994240 | elapsed time per iteration (s): 1.75 | learning rate: 3.586E-05 | global batch size: 512 | lm loss: 1.163713E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.299 | TFLOPs: 43.87 | 31: iteration 42500/ 52452 | consumed samples: 21760000 | consumed tokens: 44564480000 | elapsed time per iteration (s): 1.76 | learning rate: 3.583E-05 | global batch size: 512 | lm loss: 1.174327E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.076 | TFLOPs: 43.69 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42500 | lm loss value: 4.209351E+00 | lm loss PPL: 6.731281E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42500 | lm loss value: 4.063213E+00 | lm loss PPL: 5.816090E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42510/ 52452 | consumed samples: 21765120 | consumed tokens: 44574965760 | elapsed time per iteration (s): 9.17 | learning rate: 3.580E-05 | global batch size: 512 | lm loss: 1.184198E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.840 | TFLOPs: 8.38 | 31: iteration 42520/ 52452 | consumed samples: 21770240 | consumed tokens: 44585451520 | elapsed time per iteration (s): 1.76 | learning rate: 3.576E-05 | global batch size: 512 | lm loss: 1.175454E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.297 | TFLOPs: 43.72 | 31: iteration 42530/ 52452 | consumed samples: 21775360 | consumed tokens: 44595937280 | elapsed time per iteration (s): 1.79 | learning rate: 3.573E-05 | global batch size: 512 | lm loss: 1.175446E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.317 | TFLOPs: 42.82 | 31: iteration 42540/ 52452 | consumed samples: 21780480 | consumed tokens: 44606423040 | elapsed time per iteration (s): 1.79 | learning rate: 3.570E-05 | global batch size: 512 | lm loss: 1.177400E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.324 | TFLOPs: 42.83 | 31: iteration 42550/ 52452 | consumed samples: 21785600 | consumed tokens: 44616908800 | elapsed time per iteration (s): 1.80 | learning rate: 3.567E-05 | global batch size: 512 | lm loss: 1.169299E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.975 | TFLOPs: 42.62 | 31: iteration 42560/ 52452 | consumed samples: 21790720 | consumed tokens: 44627394560 | elapsed time per iteration (s): 1.79 | learning rate: 3.564E-05 | global batch size: 512 | lm loss: 1.174267E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.754 | TFLOPs: 43.04 | 31: iteration 42570/ 52452 | consumed samples: 21795840 | consumed tokens: 44637880320 | elapsed time per iteration (s): 1.77 | learning rate: 3.561E-05 | global batch size: 512 | lm loss: 1.187765E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.201 | TFLOPs: 43.41 | 31: iteration 42580/ 52452 | consumed samples: 21800960 | consumed tokens: 44648366080 | elapsed time per iteration (s): 1.82 | learning rate: 3.558E-05 | global batch size: 512 | lm loss: 1.175975E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.042 | TFLOPs: 42.18 | 31: iteration 42590/ 52452 | consumed samples: 21806080 | consumed tokens: 44658851840 | elapsed time per iteration (s): 1.72 | learning rate: 3.555E-05 | global batch size: 512 | lm loss: 1.165131E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.429 | TFLOPs: 44.64 | 31: iteration 42600/ 52452 | consumed samples: 21811200 | consumed tokens: 44669337600 | elapsed time per iteration (s): 1.74 | learning rate: 3.552E-05 | global batch size: 512 | lm loss: 1.186125E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.439 | TFLOPs: 44.04 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42600 | lm loss value: 4.266081E+00 | lm loss PPL: 7.124191E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42600 | lm loss value: 4.119095E+00 | lm loss PPL: 6.150355E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42610/ 52452 | consumed samples: 21816320 | consumed tokens: 44679823360 | elapsed time per iteration (s): 9.19 | learning rate: 3.549E-05 | global batch size: 512 | lm loss: 1.178800E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.707 | TFLOPs: 8.36 | 31: iteration 42620/ 52452 | consumed samples: 21821440 | consumed tokens: 44690309120 | elapsed time per iteration (s): 1.85 | learning rate: 3.546E-05 | global batch size: 512 | lm loss: 1.177407E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.055 | TFLOPs: 41.58 | 31: iteration 42630/ 52452 | consumed samples: 21826560 | consumed tokens: 44700794880 | elapsed time per iteration (s): 2.04 | learning rate: 3.543E-05 | global batch size: 512 | lm loss: 1.176077E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 250.489 | TFLOPs: 37.60 | 31: iteration 42640/ 52452 | consumed samples: 21831680 | consumed tokens: 44711280640 | elapsed time per iteration (s): 1.78 | learning rate: 3.540E-05 | global batch size: 512 | lm loss: 1.188218E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.304 | TFLOPs: 43.12 | 31: iteration 42650/ 52452 | consumed samples: 21836800 | consumed tokens: 44721766400 | elapsed time per iteration (s): 1.77 | learning rate: 3.537E-05 | global batch size: 512 | lm loss: 1.194945E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.705 | TFLOPs: 43.33 | 31: iteration 42660/ 52452 | consumed samples: 21841920 | consumed tokens: 44732252160 | elapsed time per iteration (s): 1.81 | learning rate: 3.534E-05 | global batch size: 512 | lm loss: 1.164680E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.534 | TFLOPs: 42.56 | 31: iteration 42670/ 52452 | consumed samples: 21847040 | consumed tokens: 44742737920 | elapsed time per iteration (s): 1.74 | learning rate: 3.531E-05 | global batch size: 512 | lm loss: 1.173762E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.646 | TFLOPs: 44.22 | 31: iteration 42680/ 52452 | consumed samples: 21852160 | consumed tokens: 44753223680 | elapsed time per iteration (s): 1.77 | learning rate: 3.528E-05 | global batch size: 512 | lm loss: 1.163903E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.311 | TFLOPs: 43.42 | 31: iteration 42690/ 52452 | consumed samples: 21857280 | consumed tokens: 44763709440 | elapsed time per iteration (s): 1.78 | learning rate: 3.525E-05 | global batch size: 512 | lm loss: 1.178625E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.210 | TFLOPs: 43.11 | 31: iteration 42700/ 52452 | consumed samples: 21862400 | consumed tokens: 44774195200 | elapsed time per iteration (s): 1.74 | learning rate: 3.522E-05 | global batch size: 512 | lm loss: 1.179846E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.641 | TFLOPs: 44.22 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42700 | lm loss value: 4.248641E+00 | lm loss PPL: 7.001024E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42700 | lm loss value: 4.082079E+00 | lm loss PPL: 5.926859E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42710/ 52452 | consumed samples: 21867520 | consumed tokens: 44784680960 | elapsed time per iteration (s): 9.17 | learning rate: 3.518E-05 | global batch size: 512 | lm loss: 1.175937E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.846 | TFLOPs: 8.38 | 31: iteration 42720/ 52452 | consumed samples: 21872640 | consumed tokens: 44795166720 | elapsed time per iteration (s): 1.74 | learning rate: 3.515E-05 | global batch size: 512 | lm loss: 1.169500E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.425 | TFLOPs: 44.04 | 31: iteration 42730/ 52452 | consumed samples: 21877760 | consumed tokens: 44805652480 | elapsed time per iteration (s): 1.83 | learning rate: 3.512E-05 | global batch size: 512 | lm loss: 1.174254E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.449 | TFLOPs: 41.94 | 31: iteration 42740/ 52452 | consumed samples: 21882880 | consumed tokens: 44816138240 | elapsed time per iteration (s): 1.77 | learning rate: 3.509E-05 | global batch size: 512 | lm loss: 1.173995E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.020 | TFLOPs: 43.38 | 31: iteration 42750/ 52452 | consumed samples: 21888000 | consumed tokens: 44826624000 | elapsed time per iteration (s): 1.79 | learning rate: 3.506E-05 | global batch size: 512 | lm loss: 1.183216E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.446 | TFLOPs: 42.84 | 31: iteration 42760/ 52452 | consumed samples: 21893120 | consumed tokens: 44837109760 | elapsed time per iteration (s): 1.78 | learning rate: 3.503E-05 | global batch size: 512 | lm loss: 1.165427E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.657 | TFLOPs: 43.18 | 31: iteration 42770/ 52452 | consumed samples: 21898240 | consumed tokens: 44847595520 | elapsed time per iteration (s): 1.74 | learning rate: 3.500E-05 | global batch size: 512 | lm loss: 1.173961E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.074 | TFLOPs: 44.29 | 31: iteration 42780/ 52452 | consumed samples: 21903360 | consumed tokens: 44858081280 | elapsed time per iteration (s): 1.74 | learning rate: 3.497E-05 | global batch size: 512 | lm loss: 1.181459E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.027 | TFLOPs: 44.13 | 31: iteration 42790/ 52452 | consumed samples: 21908480 | consumed tokens: 44868567040 | elapsed time per iteration (s): 1.76 | learning rate: 3.494E-05 | global batch size: 512 | lm loss: 1.174731E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.189 | TFLOPs: 43.71 | 31: iteration 42800/ 52452 | consumed samples: 21913600 | consumed tokens: 44879052800 | elapsed time per iteration (s): 1.75 | learning rate: 3.491E-05 | global batch size: 512 | lm loss: 1.189567E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.229 | TFLOPs: 43.86 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42800 | lm loss value: 4.156241E+00 | lm loss PPL: 6.383116E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42800 | lm loss value: 4.037108E+00 | lm loss PPL: 5.666224E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42810/ 52452 | consumed samples: 21918720 | consumed tokens: 44889538560 | elapsed time per iteration (s): 9.15 | learning rate: 3.488E-05 | global batch size: 512 | lm loss: 1.175037E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.927 | TFLOPs: 8.39 | 31: iteration 42820/ 52452 | consumed samples: 21923840 | consumed tokens: 44900024320 | elapsed time per iteration (s): 1.81 | learning rate: 3.485E-05 | global batch size: 512 | lm loss: 1.185878E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.124 | TFLOPs: 42.35 | 31: iteration 42830/ 52452 | consumed samples: 21928960 | consumed tokens: 44910510080 | elapsed time per iteration (s): 1.74 | learning rate: 3.482E-05 | global batch size: 512 | lm loss: 1.185852E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.963 | TFLOPs: 44.27 | 31: iteration 42840/ 52452 | consumed samples: 21934080 | consumed tokens: 44920995840 | elapsed time per iteration (s): 1.76 | learning rate: 3.479E-05 | global batch size: 512 | lm loss: 1.178440E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.346 | TFLOPs: 43.73 | 31: iteration 42850/ 52452 | consumed samples: 21939200 | consumed tokens: 44931481600 | elapsed time per iteration (s): 1.79 | learning rate: 3.476E-05 | global batch size: 512 | lm loss: 1.164287E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.948 | TFLOPs: 42.92 | 31: iteration 42860/ 52452 | consumed samples: 21944320 | consumed tokens: 44941967360 | elapsed time per iteration (s): 1.77 | learning rate: 3.473E-05 | global batch size: 512 | lm loss: 1.164149E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.148 | TFLOPs: 43.40 | 31: iteration 42870/ 52452 | consumed samples: 21949440 | consumed tokens: 44952453120 | elapsed time per iteration (s): 1.76 | learning rate: 3.470E-05 | global batch size: 512 | lm loss: 1.161119E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.231 | TFLOPs: 43.71 | 31: iteration 42880/ 52452 | consumed samples: 21954560 | consumed tokens: 44962938880 | elapsed time per iteration (s): 1.77 | learning rate: 3.467E-05 | global batch size: 512 | lm loss: 1.153234E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.598 | TFLOPs: 43.47 | 31: iteration 42890/ 52452 | consumed samples: 21959680 | consumed tokens: 44973424640 | elapsed time per iteration (s): 1.80 | learning rate: 3.464E-05 | global batch size: 512 | lm loss: 1.152055E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.837 | TFLOPs: 42.75 | 31: iteration 42900/ 52452 | consumed samples: 21964800 | consumed tokens: 44983910400 | elapsed time per iteration (s): 1.75 | learning rate: 3.461E-05 | global batch size: 512 | lm loss: 1.168827E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.335 | TFLOPs: 43.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 42900 | lm loss value: 4.172396E+00 | lm loss PPL: 6.487068E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 42900 | lm loss value: 4.041534E+00 | lm loss PPL: 5.691358E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 42910/ 52452 | consumed samples: 21969920 | consumed tokens: 44994396160 | elapsed time per iteration (s): 9.14 | learning rate: 3.459E-05 | global batch size: 512 | lm loss: 1.154941E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.005 | TFLOPs: 8.41 | 31: iteration 42920/ 52452 | consumed samples: 21975040 | consumed tokens: 45004881920 | elapsed time per iteration (s): 1.74 | learning rate: 3.456E-05 | global batch size: 512 | lm loss: 1.166179E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.980 | TFLOPs: 44.12 | 31: iteration 42930/ 52452 | consumed samples: 21980160 | consumed tokens: 45015367680 | elapsed time per iteration (s): 1.76 | learning rate: 3.453E-05 | global batch size: 512 | lm loss: 1.150827E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.895 | TFLOPs: 43.66 | 31: iteration 42940/ 52452 | consumed samples: 21985280 | consumed tokens: 45025853440 | elapsed time per iteration (s): 1.80 | learning rate: 3.450E-05 | global batch size: 512 | lm loss: 1.161074E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.550 | TFLOPs: 42.71 | 31: iteration 42950/ 52452 | consumed samples: 21990400 | consumed tokens: 45036339200 | elapsed time per iteration (s): 1.74 | learning rate: 3.447E-05 | global batch size: 512 | lm loss: 1.162841E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.892 | TFLOPs: 44.11 | 31: iteration 42960/ 52452 | consumed samples: 21995520 | consumed tokens: 45046824960 | elapsed time per iteration (s): 1.85 | learning rate: 3.444E-05 | global batch size: 512 | lm loss: 1.163842E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.402 | TFLOPs: 41.49 | 31: iteration 42970/ 52452 | consumed samples: 22000640 | consumed tokens: 45057310720 | elapsed time per iteration (s): 1.79 | learning rate: 3.441E-05 | global batch size: 512 | lm loss: 1.163635E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.861 | TFLOPs: 42.91 | 31: iteration 42980/ 52452 | consumed samples: 22005760 | consumed tokens: 45067796480 | elapsed time per iteration (s): 1.77 | learning rate: 3.438E-05 | global batch size: 512 | lm loss: 1.153367E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.941 | TFLOPs: 43.52 | 31: iteration 42990/ 52452 | consumed samples: 22010880 | consumed tokens: 45078282240 | elapsed time per iteration (s): 1.76 | learning rate: 3.435E-05 | global batch size: 512 | lm loss: 1.177156E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.823 | TFLOPs: 43.65 | 31: iteration 43000/ 52452 | consumed samples: 22016000 | consumed tokens: 45088768000 | elapsed time per iteration (s): 1.84 | learning rate: 3.432E-05 | global batch size: 512 | lm loss: 1.157984E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.094 | TFLOPs: 41.74 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43000 | lm loss value: 4.259711E+00 | lm loss PPL: 7.078954E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43000 | lm loss value: 4.109636E+00 | lm loss PPL: 6.092453E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43010/ 52452 | consumed samples: 22021120 | consumed tokens: 45099253760 | elapsed time per iteration (s): 9.14 | learning rate: 3.429E-05 | global batch size: 512 | lm loss: 1.159472E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.989 | TFLOPs: 8.40 | 31: iteration 43020/ 52452 | consumed samples: 22026240 | consumed tokens: 45109739520 | elapsed time per iteration (s): 1.75 | learning rate: 3.426E-05 | global batch size: 512 | lm loss: 1.160715E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.151 | TFLOPs: 43.85 | 31: iteration 43030/ 52452 | consumed samples: 22031360 | consumed tokens: 45120225280 | elapsed time per iteration (s): 1.80 | learning rate: 3.423E-05 | global batch size: 512 | lm loss: 1.157736E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.699 | TFLOPs: 42.73 | 31: iteration 43040/ 52452 | consumed samples: 22036480 | consumed tokens: 45130711040 | elapsed time per iteration (s): 1.77 | learning rate: 3.420E-05 | global batch size: 512 | lm loss: 1.163492E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.944 | TFLOPs: 43.52 | 31: iteration 43050/ 52452 | consumed samples: 22041600 | consumed tokens: 45141196800 | elapsed time per iteration (s): 1.78 | learning rate: 3.417E-05 | global batch size: 512 | lm loss: 1.153893E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.315 | TFLOPs: 43.12 | 31: iteration 43060/ 52452 | consumed samples: 22046720 | consumed tokens: 45151682560 | elapsed time per iteration (s): 1.76 | learning rate: 3.414E-05 | global batch size: 512 | lm loss: 1.154934E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.154 | TFLOPs: 43.70 | 31: iteration 43070/ 52452 | consumed samples: 22051840 | consumed tokens: 45162168320 | elapsed time per iteration (s): 1.75 | learning rate: 3.411E-05 | global batch size: 512 | lm loss: 1.166295E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.955 | TFLOPs: 43.82 | 31: iteration 43080/ 52452 | consumed samples: 22056960 | consumed tokens: 45172654080 | elapsed time per iteration (s): 1.76 | learning rate: 3.408E-05 | global batch size: 512 | lm loss: 1.157772E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.678 | TFLOPs: 43.78 | 31: iteration 43090/ 52452 | consumed samples: 22062080 | consumed tokens: 45183139840 | elapsed time per iteration (s): 1.75 | learning rate: 3.405E-05 | global batch size: 512 | lm loss: 1.163406E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.377 | TFLOPs: 43.88 | 31: iteration 43100/ 52452 | consumed samples: 22067200 | consumed tokens: 45193625600 | elapsed time per iteration (s): 1.73 | learning rate: 3.403E-05 | global batch size: 512 | lm loss: 1.168167E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.137 | TFLOPs: 44.45 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43100 | lm loss value: 4.263091E+00 | lm loss PPL: 7.102920E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43100 | lm loss value: 4.108913E+00 | lm loss PPL: 6.088050E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43110/ 52452 | consumed samples: 22072320 | consumed tokens: 45204111360 | elapsed time per iteration (s): 9.26 | learning rate: 3.400E-05 | global batch size: 512 | lm loss: 1.162720E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.321 | TFLOPs: 8.30 | 31: iteration 43120/ 52452 | consumed samples: 22077440 | consumed tokens: 45214597120 | elapsed time per iteration (s): 1.73 | learning rate: 3.397E-05 | global batch size: 512 | lm loss: 1.151248E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.701 | TFLOPs: 44.38 | 31: iteration 43130/ 52452 | consumed samples: 22082560 | consumed tokens: 45225082880 | elapsed time per iteration (s): 1.79 | learning rate: 3.394E-05 | global batch size: 512 | lm loss: 1.173005E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.489 | TFLOPs: 43.00 | 31: iteration 43140/ 52452 | consumed samples: 22087680 | consumed tokens: 45235568640 | elapsed time per iteration (s): 1.76 | learning rate: 3.391E-05 | global batch size: 512 | lm loss: 1.156922E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.440 | TFLOPs: 43.59 | 31: iteration 43150/ 52452 | consumed samples: 22092800 | consumed tokens: 45246054400 | elapsed time per iteration (s): 1.73 | learning rate: 3.388E-05 | global batch size: 512 | lm loss: 1.159257E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.504 | TFLOPs: 44.35 | 31: iteration 43160/ 52452 | consumed samples: 22097920 | consumed tokens: 45256540160 | elapsed time per iteration (s): 1.81 | learning rate: 3.385E-05 | global batch size: 512 | lm loss: 1.150176E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.933 | TFLOPs: 42.47 | 31: iteration 43170/ 52452 | consumed samples: 22103040 | consumed tokens: 45267025920 | elapsed time per iteration (s): 1.78 | learning rate: 3.382E-05 | global batch size: 512 | lm loss: 1.162394E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.372 | TFLOPs: 43.28 | 31: iteration 43180/ 52452 | consumed samples: 22108160 | consumed tokens: 45277511680 | elapsed time per iteration (s): 1.79 | learning rate: 3.379E-05 | global batch size: 512 | lm loss: 1.150368E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.723 | TFLOPs: 42.89 | 31: iteration 43190/ 52452 | consumed samples: 22113280 | consumed tokens: 45287997440 | elapsed time per iteration (s): 1.74 | learning rate: 3.376E-05 | global batch size: 512 | lm loss: 1.162224E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.740 | TFLOPs: 44.24 | 31: iteration 43200/ 52452 | consumed samples: 22118400 | consumed tokens: 45298483200 | elapsed time per iteration (s): 1.82 | learning rate: 3.373E-05 | global batch size: 512 | lm loss: 1.164951E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.577 | TFLOPs: 42.26 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43200 | lm loss value: 4.263462E+00 | lm loss PPL: 7.105552E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43200 | lm loss value: 4.100440E+00 | lm loss PPL: 6.036682E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43210/ 52452 | consumed samples: 22123520 | consumed tokens: 45308968960 | elapsed time per iteration (s): 9.14 | learning rate: 3.371E-05 | global batch size: 512 | lm loss: 1.147626E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.993 | TFLOPs: 8.40 | 31: iteration 43220/ 52452 | consumed samples: 22128640 | consumed tokens: 45319454720 | elapsed time per iteration (s): 1.73 | learning rate: 3.368E-05 | global batch size: 512 | lm loss: 1.162534E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.650 | TFLOPs: 44.53 | 31: iteration 43230/ 52452 | consumed samples: 22133760 | consumed tokens: 45329940480 | elapsed time per iteration (s): 1.79 | learning rate: 3.365E-05 | global batch size: 512 | lm loss: 1.158811E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.681 | TFLOPs: 43.03 | 31: iteration 43240/ 52452 | consumed samples: 22138880 | consumed tokens: 45340426240 | elapsed time per iteration (s): 1.80 | learning rate: 3.362E-05 | global batch size: 512 | lm loss: 1.163108E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.377 | TFLOPs: 42.68 | 31: iteration 43250/ 52452 | consumed samples: 22144000 | consumed tokens: 45350912000 | elapsed time per iteration (s): 1.74 | learning rate: 3.359E-05 | global batch size: 512 | lm loss: 1.160796E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.700 | TFLOPs: 44.08 | 31: iteration 43260/ 52452 | consumed samples: 22149120 | consumed tokens: 45361397760 | elapsed time per iteration (s): 1.75 | learning rate: 3.356E-05 | global batch size: 512 | lm loss: 1.163723E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.260 | TFLOPs: 44.02 | 31: iteration 43270/ 52452 | consumed samples: 22154240 | consumed tokens: 45371883520 | elapsed time per iteration (s): 1.77 | learning rate: 3.353E-05 | global batch size: 512 | lm loss: 1.178254E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.585 | TFLOPs: 43.47 | 31: iteration 43280/ 52452 | consumed samples: 22159360 | consumed tokens: 45382369280 | elapsed time per iteration (s): 1.75 | learning rate: 3.350E-05 | global batch size: 512 | lm loss: 1.166146E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.288 | TFLOPs: 43.87 | 31: iteration 43290/ 52452 | consumed samples: 22164480 | consumed tokens: 45392855040 | elapsed time per iteration (s): 1.75 | learning rate: 3.348E-05 | global batch size: 512 | lm loss: 1.160164E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.744 | TFLOPs: 43.94 | 31: iteration 43300/ 52452 | consumed samples: 22169600 | consumed tokens: 45403340800 | elapsed time per iteration (s): 1.77 | learning rate: 3.345E-05 | global batch size: 512 | lm loss: 1.161379E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.863 | TFLOPs: 43.36 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43300 | lm loss value: 4.138584E+00 | lm loss PPL: 6.271396E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43300 | lm loss value: 4.018801E+00 | lm loss PPL: 5.563435E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43310/ 52452 | consumed samples: 22174720 | consumed tokens: 45413826560 | elapsed time per iteration (s): 9.16 | learning rate: 3.342E-05 | global batch size: 512 | lm loss: 1.153610E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.878 | TFLOPs: 8.39 | 31: iteration 43320/ 52452 | consumed samples: 22179840 | consumed tokens: 45424312320 | elapsed time per iteration (s): 1.76 | learning rate: 3.339E-05 | global batch size: 512 | lm loss: 1.166822E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.336 | TFLOPs: 43.73 | 31: iteration 43330/ 52452 | consumed samples: 22184960 | consumed tokens: 45434798080 | elapsed time per iteration (s): 1.76 | learning rate: 3.336E-05 | global batch size: 512 | lm loss: 1.153661E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.066 | TFLOPs: 43.69 | 31: iteration 43340/ 52452 | consumed samples: 22190080 | consumed tokens: 45445283840 | elapsed time per iteration (s): 1.76 | learning rate: 3.333E-05 | global batch size: 512 | lm loss: 1.163310E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.612 | TFLOPs: 43.62 | 31: iteration 43350/ 52452 | consumed samples: 22195200 | consumed tokens: 45455769600 | elapsed time per iteration (s): 1.77 | learning rate: 3.330E-05 | global batch size: 512 | lm loss: 1.160019E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.012 | TFLOPs: 43.38 | 31: iteration 43360/ 52452 | consumed samples: 22200320 | consumed tokens: 45466255360 | elapsed time per iteration (s): 1.81 | learning rate: 3.328E-05 | global batch size: 512 | lm loss: 1.178522E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.613 | TFLOPs: 42.42 | 31: iteration 43370/ 52452 | consumed samples: 22205440 | consumed tokens: 45476741120 | elapsed time per iteration (s): 1.76 | learning rate: 3.325E-05 | global batch size: 512 | lm loss: 1.164752E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.422 | TFLOPs: 43.59 | 31: iteration 43380/ 52452 | consumed samples: 22210560 | consumed tokens: 45487226880 | elapsed time per iteration (s): 1.77 | learning rate: 3.322E-05 | global batch size: 512 | lm loss: 1.161871E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.235 | TFLOPs: 43.41 | 31: iteration 43390/ 52452 | consumed samples: 22215680 | consumed tokens: 45497712640 | elapsed time per iteration (s): 1.77 | learning rate: 3.319E-05 | global batch size: 512 | lm loss: 1.166411E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.966 | TFLOPs: 43.52 | 31: iteration 43400/ 52452 | consumed samples: 22220800 | consumed tokens: 45508198400 | elapsed time per iteration (s): 1.77 | learning rate: 3.316E-05 | global batch size: 512 | lm loss: 1.176354E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.636 | TFLOPs: 43.47 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43400 | lm loss value: 4.085930E+00 | lm loss PPL: 5.949724E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43400 | lm loss value: 3.992190E+00 | lm loss PPL: 5.417341E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43410/ 52452 | consumed samples: 22225920 | consumed tokens: 45518684160 | elapsed time per iteration (s): 9.19 | learning rate: 3.313E-05 | global batch size: 512 | lm loss: 1.160780E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.701 | TFLOPs: 8.36 | 31: iteration 43420/ 52452 | consumed samples: 22231040 | consumed tokens: 45529169920 | elapsed time per iteration (s): 1.73 | learning rate: 3.311E-05 | global batch size: 512 | lm loss: 1.156974E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.459 | TFLOPs: 44.35 | 31: iteration 43430/ 52452 | consumed samples: 22236160 | consumed tokens: 45539655680 | elapsed time per iteration (s): 1.73 | learning rate: 3.308E-05 | global batch size: 512 | lm loss: 1.164494E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.016 | TFLOPs: 44.43 | 31: iteration 43440/ 52452 | consumed samples: 22241280 | consumed tokens: 45550141440 | elapsed time per iteration (s): 1.79 | learning rate: 3.305E-05 | global batch size: 512 | lm loss: 1.164643E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.002 | TFLOPs: 42.93 | 31: iteration 43450/ 52452 | consumed samples: 22246400 | consumed tokens: 45560627200 | elapsed time per iteration (s): 1.74 | learning rate: 3.302E-05 | global batch size: 512 | lm loss: 1.163984E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.218 | TFLOPs: 44.16 | 31: iteration 43460/ 52452 | consumed samples: 22251520 | consumed tokens: 45571112960 | elapsed time per iteration (s): 1.73 | learning rate: 3.299E-05 | global batch size: 512 | lm loss: 1.165502E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.454 | TFLOPs: 44.50 | 31: iteration 43470/ 52452 | consumed samples: 22256640 | consumed tokens: 45581598720 | elapsed time per iteration (s): 1.78 | learning rate: 3.296E-05 | global batch size: 512 | lm loss: 1.163457E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.529 | TFLOPs: 43.16 | 31: iteration 43480/ 52452 | consumed samples: 22261760 | consumed tokens: 45592084480 | elapsed time per iteration (s): 1.84 | learning rate: 3.294E-05 | global batch size: 512 | lm loss: 1.171147E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.816 | TFLOPs: 41.70 | 31: iteration 43490/ 52452 | consumed samples: 22266880 | consumed tokens: 45602570240 | elapsed time per iteration (s): 1.74 | learning rate: 3.291E-05 | global batch size: 512 | lm loss: 1.165828E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.898 | TFLOPs: 44.26 | 31: iteration 43500/ 52452 | consumed samples: 22272000 | consumed tokens: 45613056000 | elapsed time per iteration (s): 1.78 | learning rate: 3.288E-05 | global batch size: 512 | lm loss: 1.161193E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.725 | TFLOPs: 43.19 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43500 | lm loss value: 4.244025E+00 | lm loss PPL: 6.968776E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43500 | lm loss value: 4.084587E+00 | lm loss PPL: 5.941737E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43510/ 52452 | consumed samples: 22277120 | consumed tokens: 45623541760 | elapsed time per iteration (s): 9.21 | learning rate: 3.285E-05 | global batch size: 512 | lm loss: 1.161348E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.569 | TFLOPs: 8.34 | 31: iteration 43520/ 52452 | consumed samples: 22282240 | consumed tokens: 45634027520 | elapsed time per iteration (s): 1.84 | learning rate: 3.282E-05 | global batch size: 512 | lm loss: 1.160155E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.525 | TFLOPs: 41.81 | 31: iteration 43530/ 52452 | consumed samples: 22287360 | consumed tokens: 45644513280 | elapsed time per iteration (s): 1.80 | learning rate: 3.280E-05 | global batch size: 512 | lm loss: 1.156124E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.906 | TFLOPs: 42.76 | 31: iteration 43540/ 52452 | consumed samples: 22292480 | consumed tokens: 45654999040 | elapsed time per iteration (s): 1.89 | learning rate: 3.277E-05 | global batch size: 512 | lm loss: 1.160990E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.594 | TFLOPs: 40.76 | 31: iteration 43550/ 52452 | consumed samples: 22297600 | consumed tokens: 45665484800 | elapsed time per iteration (s): 1.81 | learning rate: 3.274E-05 | global batch size: 512 | lm loss: 1.155943E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.595 | TFLOPs: 42.42 | 31: iteration 43560/ 52452 | consumed samples: 22302720 | consumed tokens: 45675970560 | elapsed time per iteration (s): 1.84 | learning rate: 3.271E-05 | global batch size: 512 | lm loss: 1.152618E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.710 | TFLOPs: 41.68 | 31: iteration 43570/ 52452 | consumed samples: 22307840 | consumed tokens: 45686456320 | elapsed time per iteration (s): 1.83 | learning rate: 3.268E-05 | global batch size: 512 | lm loss: 1.155238E+00 | grad norm: 0.148 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.803 | TFLOPs: 42.00 | 31: iteration 43580/ 52452 | consumed samples: 22312960 | consumed tokens: 45696942080 | elapsed time per iteration (s): 1.81 | learning rate: 3.266E-05 | global batch size: 512 | lm loss: 1.157419E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.500 | TFLOPs: 42.55 | 31: iteration 43590/ 52452 | consumed samples: 22318080 | consumed tokens: 45707427840 | elapsed time per iteration (s): 1.74 | learning rate: 3.263E-05 | global batch size: 512 | lm loss: 1.154958E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.839 | TFLOPs: 44.10 | 31: iteration 43600/ 52452 | consumed samples: 22323200 | consumed tokens: 45717913600 | elapsed time per iteration (s): 1.80 | learning rate: 3.260E-05 | global batch size: 512 | lm loss: 1.148610E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.056 | TFLOPs: 42.79 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43600 | lm loss value: 4.184860E+00 | lm loss PPL: 6.568429E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43600 | lm loss value: 4.050932E+00 | lm loss PPL: 5.745100E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43610/ 52452 | consumed samples: 22328320 | consumed tokens: 45728399360 | elapsed time per iteration (s): 9.23 | learning rate: 3.257E-05 | global batch size: 512 | lm loss: 1.165293E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.480 | TFLOPs: 8.33 | 31: iteration 43620/ 52452 | consumed samples: 22333440 | consumed tokens: 45738885120 | elapsed time per iteration (s): 1.76 | learning rate: 3.255E-05 | global batch size: 512 | lm loss: 1.158887E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.115 | TFLOPs: 43.69 | 31: iteration 43630/ 52452 | consumed samples: 22338560 | consumed tokens: 45749370880 | elapsed time per iteration (s): 1.76 | learning rate: 3.252E-05 | global batch size: 512 | lm loss: 1.167341E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.012 | TFLOPs: 43.68 | 31: iteration 43640/ 52452 | consumed samples: 22343680 | consumed tokens: 45759856640 | elapsed time per iteration (s): 1.83 | learning rate: 3.249E-05 | global batch size: 512 | lm loss: 1.156963E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.061 | TFLOPs: 41.89 | 31: iteration 43650/ 52452 | consumed samples: 22348800 | consumed tokens: 45770342400 | elapsed time per iteration (s): 1.82 | learning rate: 3.246E-05 | global batch size: 512 | lm loss: 1.163091E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.623 | TFLOPs: 42.27 | 31: iteration 43660/ 52452 | consumed samples: 22353920 | consumed tokens: 45780828160 | elapsed time per iteration (s): 1.81 | learning rate: 3.243E-05 | global batch size: 512 | lm loss: 1.150043E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.055 | TFLOPs: 42.48 | 31: iteration 43670/ 52452 | consumed samples: 22359040 | consumed tokens: 45791313920 | elapsed time per iteration (s): 1.73 | learning rate: 3.241E-05 | global batch size: 512 | lm loss: 1.152600E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.676 | TFLOPs: 44.38 | 31: iteration 43680/ 52452 | consumed samples: 22364160 | consumed tokens: 45801799680 | elapsed time per iteration (s): 1.78 | learning rate: 3.238E-05 | global batch size: 512 | lm loss: 1.155599E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.002 | TFLOPs: 43.23 | 31: iteration 43690/ 52452 | consumed samples: 22369280 | consumed tokens: 45812285440 | elapsed time per iteration (s): 1.81 | learning rate: 3.235E-05 | global batch size: 512 | lm loss: 1.146416E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.431 | TFLOPs: 42.54 | 31: iteration 43700/ 52452 | consumed samples: 22374400 | consumed tokens: 45822771200 | elapsed time per iteration (s): 1.75 | learning rate: 3.232E-05 | global batch size: 512 | lm loss: 1.148758E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.493 | TFLOPs: 43.90 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43700 | lm loss value: 4.196861E+00 | lm loss PPL: 6.647735E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43700 | lm loss value: 4.052476E+00 | lm loss PPL: 5.753977E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43710/ 52452 | consumed samples: 22379520 | consumed tokens: 45833256960 | elapsed time per iteration (s): 9.21 | learning rate: 3.230E-05 | global batch size: 512 | lm loss: 1.164897E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.592 | TFLOPs: 8.34 | 31: iteration 43720/ 52452 | consumed samples: 22384640 | consumed tokens: 45843742720 | elapsed time per iteration (s): 1.77 | learning rate: 3.227E-05 | global batch size: 512 | lm loss: 1.172825E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.912 | TFLOPs: 43.51 | 31: iteration 43730/ 52452 | consumed samples: 22389760 | consumed tokens: 45854228480 | elapsed time per iteration (s): 1.75 | learning rate: 3.224E-05 | global batch size: 512 | lm loss: 1.162368E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.156 | TFLOPs: 44.00 | 31: iteration 43740/ 52452 | consumed samples: 22394880 | consumed tokens: 45864714240 | elapsed time per iteration (s): 1.75 | learning rate: 3.221E-05 | global batch size: 512 | lm loss: 1.150369E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.225 | TFLOPs: 44.01 | 31: iteration 43750/ 52452 | consumed samples: 22400000 | consumed tokens: 45875200000 | elapsed time per iteration (s): 1.82 | learning rate: 3.219E-05 | global batch size: 512 | lm loss: 1.159456E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.321 | TFLOPs: 42.22 | 31: iteration 43760/ 52452 | consumed samples: 22405120 | consumed tokens: 45885685760 | elapsed time per iteration (s): 1.77 | learning rate: 3.216E-05 | global batch size: 512 | lm loss: 1.170530E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.671 | TFLOPs: 43.33 | 31: iteration 43770/ 52452 | consumed samples: 22410240 | consumed tokens: 45896171520 | elapsed time per iteration (s): 1.73 | learning rate: 3.213E-05 | global batch size: 512 | lm loss: 1.164545E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.863 | TFLOPs: 44.41 | 31: iteration 43780/ 52452 | consumed samples: 22415360 | consumed tokens: 45906657280 | elapsed time per iteration (s): 1.79 | learning rate: 3.211E-05 | global batch size: 512 | lm loss: 1.145059E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.259 | TFLOPs: 42.97 | 31: iteration 43790/ 52452 | consumed samples: 22420480 | consumed tokens: 45917143040 | elapsed time per iteration (s): 1.74 | learning rate: 3.208E-05 | global batch size: 512 | lm loss: 1.151075E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.471 | TFLOPs: 44.20 | 31: iteration 43800/ 52452 | consumed samples: 22425600 | consumed tokens: 45927628800 | elapsed time per iteration (s): 1.74 | learning rate: 3.205E-05 | global batch size: 512 | lm loss: 1.166956E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.396 | TFLOPs: 44.19 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43800 | lm loss value: 4.223136E+00 | lm loss PPL: 6.824717E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43800 | lm loss value: 4.078318E+00 | lm loss PPL: 5.904608E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43810/ 52452 | consumed samples: 22430720 | consumed tokens: 45938114560 | elapsed time per iteration (s): 9.19 | learning rate: 3.202E-05 | global batch size: 512 | lm loss: 1.159431E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.728 | TFLOPs: 8.36 | 31: iteration 43820/ 52452 | consumed samples: 22435840 | consumed tokens: 45948600320 | elapsed time per iteration (s): 1.76 | learning rate: 3.200E-05 | global batch size: 512 | lm loss: 1.146399E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.301 | TFLOPs: 43.72 | 31: iteration 43830/ 52452 | consumed samples: 22440960 | consumed tokens: 45959086080 | elapsed time per iteration (s): 1.82 | learning rate: 3.197E-05 | global batch size: 512 | lm loss: 1.157471E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.153 | TFLOPs: 42.20 | 31: iteration 43840/ 52452 | consumed samples: 22446080 | consumed tokens: 45969571840 | elapsed time per iteration (s): 1.75 | learning rate: 3.194E-05 | global batch size: 512 | lm loss: 1.151741E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.762 | TFLOPs: 43.79 | 31: iteration 43850/ 52452 | consumed samples: 22451200 | consumed tokens: 45980057600 | elapsed time per iteration (s): 1.79 | learning rate: 3.192E-05 | global batch size: 512 | lm loss: 1.162927E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.301 | TFLOPs: 42.97 | 31: iteration 43860/ 52452 | consumed samples: 22456320 | consumed tokens: 45990543360 | elapsed time per iteration (s): 1.76 | learning rate: 3.189E-05 | global batch size: 512 | lm loss: 1.160603E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.248 | TFLOPs: 43.71 | 31: iteration 43870/ 52452 | consumed samples: 22461440 | consumed tokens: 46001029120 | elapsed time per iteration (s): 1.77 | learning rate: 3.186E-05 | global batch size: 512 | lm loss: 1.163399E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.851 | TFLOPs: 43.50 | 31: iteration 43880/ 52452 | consumed samples: 22466560 | consumed tokens: 46011514880 | elapsed time per iteration (s): 1.79 | learning rate: 3.183E-05 | global batch size: 512 | lm loss: 1.159636E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.632 | TFLOPs: 43.02 | 31: iteration 43890/ 52452 | consumed samples: 22471680 | consumed tokens: 46022000640 | elapsed time per iteration (s): 1.75 | learning rate: 3.181E-05 | global batch size: 512 | lm loss: 1.158056E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.760 | TFLOPs: 43.79 | 31: iteration 43900/ 52452 | consumed samples: 22476800 | consumed tokens: 46032486400 | elapsed time per iteration (s): 1.75 | learning rate: 3.178E-05 | global batch size: 512 | lm loss: 1.147225E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.349 | TFLOPs: 44.03 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 43900 | lm loss value: 4.284959E+00 | lm loss PPL: 7.259956E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 43900 | lm loss value: 4.121290E+00 | lm loss PPL: 6.163872E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 43910/ 52452 | consumed samples: 22481920 | consumed tokens: 46042972160 | elapsed time per iteration (s): 9.17 | learning rate: 3.175E-05 | global batch size: 512 | lm loss: 1.163148E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.823 | TFLOPs: 8.38 | 31: iteration 43920/ 52452 | consumed samples: 22487040 | consumed tokens: 46053457920 | elapsed time per iteration (s): 1.80 | learning rate: 3.173E-05 | global batch size: 512 | lm loss: 1.169457E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.612 | TFLOPs: 42.72 | 31: iteration 43930/ 52452 | consumed samples: 22492160 | consumed tokens: 46063943680 | elapsed time per iteration (s): 1.81 | learning rate: 3.170E-05 | global batch size: 512 | lm loss: 1.163871E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.654 | TFLOPs: 42.42 | 31: iteration 43940/ 52452 | consumed samples: 22497280 | consumed tokens: 46074429440 | elapsed time per iteration (s): 1.77 | learning rate: 3.167E-05 | global batch size: 512 | lm loss: 1.151009E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.099 | TFLOPs: 43.39 | 31: iteration 43950/ 52452 | consumed samples: 22502400 | consumed tokens: 46084915200 | elapsed time per iteration (s): 1.78 | learning rate: 3.165E-05 | global batch size: 512 | lm loss: 1.155865E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.180 | TFLOPs: 43.10 | 31: iteration 43960/ 52452 | consumed samples: 22507520 | consumed tokens: 46095400960 | elapsed time per iteration (s): 1.80 | learning rate: 3.162E-05 | global batch size: 512 | lm loss: 1.152029E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.023 | TFLOPs: 42.63 | 31: iteration 43970/ 52452 | consumed samples: 22512640 | consumed tokens: 46105886720 | elapsed time per iteration (s): 1.79 | learning rate: 3.159E-05 | global batch size: 512 | lm loss: 1.159112E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.562 | TFLOPs: 42.86 | 31: iteration 43980/ 52452 | consumed samples: 22517760 | consumed tokens: 46116372480 | elapsed time per iteration (s): 1.77 | learning rate: 3.157E-05 | global batch size: 512 | lm loss: 1.162699E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.006 | TFLOPs: 43.38 | 31: iteration 43990/ 52452 | consumed samples: 22522880 | consumed tokens: 46126858240 | elapsed time per iteration (s): 1.80 | learning rate: 3.154E-05 | global batch size: 512 | lm loss: 1.164364E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.073 | TFLOPs: 42.64 | 0: [2022-12-26 15:55:19,948] [INFO] [logging.py:68:log_dist] [Rank 0] step=44000, skipped=0, lr=[3.151225739431621e-05, 3.151225739431621e-05, 3.151225739431621e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 44000/ 52452 | consumed samples: 22528000 | consumed tokens: 46137344000 | elapsed time per iteration (s): 1.77 | learning rate: 3.151E-05 | global batch size: 512 | lm loss: 1.163249E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.834 | TFLOPs: 43.50 | 0: steps: 44000 loss: 1.1547 iter time (s): 1.777 samples/sec: 288.131 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44000 | lm loss value: 4.186176E+00 | lm loss PPL: 6.577079E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44000 | lm loss value: 4.064477E+00 | lm loss PPL: 5.823447E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44010/ 52452 | consumed samples: 22533120 | consumed tokens: 46147829760 | elapsed time per iteration (s): 9.22 | learning rate: 3.149E-05 | global batch size: 512 | lm loss: 1.163988E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.531 | TFLOPs: 8.33 | 31: iteration 44020/ 52452 | consumed samples: 22538240 | consumed tokens: 46158315520 | elapsed time per iteration (s): 1.81 | learning rate: 3.146E-05 | global batch size: 512 | lm loss: 1.159634E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.586 | TFLOPs: 42.56 | 31: iteration 44030/ 52452 | consumed samples: 22543360 | consumed tokens: 46168801280 | elapsed time per iteration (s): 1.78 | learning rate: 3.143E-05 | global batch size: 512 | lm loss: 1.160768E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.207 | TFLOPs: 43.26 | 31: iteration 44040/ 52452 | consumed samples: 22548480 | consumed tokens: 46179287040 | elapsed time per iteration (s): 1.84 | learning rate: 3.141E-05 | global batch size: 512 | lm loss: 1.149298E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.309 | TFLOPs: 41.77 | 31: iteration 44050/ 52452 | consumed samples: 22553600 | consumed tokens: 46189772800 | elapsed time per iteration (s): 1.79 | learning rate: 3.138E-05 | global batch size: 512 | lm loss: 1.158037E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.879 | TFLOPs: 42.91 | 31: iteration 44060/ 52452 | consumed samples: 22558720 | consumed tokens: 46200258560 | elapsed time per iteration (s): 1.86 | learning rate: 3.135E-05 | global batch size: 512 | lm loss: 1.160723E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.635 | TFLOPs: 41.22 | 31: iteration 44070/ 52452 | consumed samples: 22563840 | consumed tokens: 46210744320 | elapsed time per iteration (s): 1.79 | learning rate: 3.133E-05 | global batch size: 512 | lm loss: 1.155106E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.238 | TFLOPs: 42.81 | 31: iteration 44080/ 52452 | consumed samples: 22568960 | consumed tokens: 46221230080 | elapsed time per iteration (s): 1.79 | learning rate: 3.130E-05 | global batch size: 512 | lm loss: 1.164174E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.959 | TFLOPs: 42.92 | 31: iteration 44090/ 52452 | consumed samples: 22574080 | consumed tokens: 46231715840 | elapsed time per iteration (s): 1.87 | learning rate: 3.127E-05 | global batch size: 512 | lm loss: 1.158445E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.303 | TFLOPs: 41.02 | 31: iteration 44100/ 52452 | consumed samples: 22579200 | consumed tokens: 46242201600 | elapsed time per iteration (s): 1.83 | learning rate: 3.125E-05 | global batch size: 512 | lm loss: 1.161239E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.497 | TFLOPs: 41.95 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44100 | lm loss value: 4.221535E+00 | lm loss PPL: 6.813798E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44100 | lm loss value: 4.070755E+00 | lm loss PPL: 5.860116E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44110/ 52452 | consumed samples: 22584320 | consumed tokens: 46252687360 | elapsed time per iteration (s): 9.20 | learning rate: 3.122E-05 | global batch size: 512 | lm loss: 1.149429E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.647 | TFLOPs: 8.35 | 31: iteration 44120/ 52452 | consumed samples: 22589440 | consumed tokens: 46263173120 | elapsed time per iteration (s): 1.73 | learning rate: 3.119E-05 | global batch size: 512 | lm loss: 1.153691E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.005 | TFLOPs: 44.43 | 31: iteration 44130/ 52452 | consumed samples: 22594560 | consumed tokens: 46273658880 | elapsed time per iteration (s): 1.83 | learning rate: 3.117E-05 | global batch size: 512 | lm loss: 1.159919E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.277 | TFLOPs: 41.92 | 31: iteration 44140/ 52452 | consumed samples: 22599680 | consumed tokens: 46284144640 | elapsed time per iteration (s): 1.80 | learning rate: 3.114E-05 | global batch size: 512 | lm loss: 1.156443E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.170 | TFLOPs: 42.65 | 31: iteration 44150/ 52452 | consumed samples: 22604800 | consumed tokens: 46294630400 | elapsed time per iteration (s): 1.84 | learning rate: 3.112E-05 | global batch size: 512 | lm loss: 1.163317E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.425 | TFLOPs: 41.79 | 31: iteration 44160/ 52452 | consumed samples: 22609920 | consumed tokens: 46305116160 | elapsed time per iteration (s): 1.78 | learning rate: 3.109E-05 | global batch size: 512 | lm loss: 1.160250E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.168 | TFLOPs: 43.25 | 31: iteration 44170/ 52452 | consumed samples: 22615040 | consumed tokens: 46315601920 | elapsed time per iteration (s): 1.75 | learning rate: 3.106E-05 | global batch size: 512 | lm loss: 1.162238E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.121 | TFLOPs: 44.00 | 31: iteration 44180/ 52452 | consumed samples: 22620160 | consumed tokens: 46326087680 | elapsed time per iteration (s): 1.79 | learning rate: 3.104E-05 | global batch size: 512 | lm loss: 1.157374E+00 | grad norm: 0.215 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.331 | TFLOPs: 42.83 | 31: iteration 44190/ 52452 | consumed samples: 22625280 | consumed tokens: 46336573440 | elapsed time per iteration (s): 1.80 | learning rate: 3.101E-05 | global batch size: 512 | lm loss: 1.162982E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.664 | TFLOPs: 42.58 | 31: iteration 44200/ 52452 | consumed samples: 22630400 | consumed tokens: 46347059200 | elapsed time per iteration (s): 1.76 | learning rate: 3.099E-05 | global batch size: 512 | lm loss: 1.161440E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.344 | TFLOPs: 43.58 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44200 | lm loss value: 4.250398E+00 | lm loss PPL: 7.013330E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44200 | lm loss value: 4.092467E+00 | lm loss PPL: 5.988744E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44210/ 52452 | consumed samples: 22635520 | consumed tokens: 46357544960 | elapsed time per iteration (s): 9.26 | learning rate: 3.096E-05 | global batch size: 512 | lm loss: 1.162388E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.298 | TFLOPs: 8.30 | 31: iteration 44220/ 52452 | consumed samples: 22640640 | consumed tokens: 46368030720 | elapsed time per iteration (s): 1.79 | learning rate: 3.093E-05 | global batch size: 512 | lm loss: 1.160776E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.356 | TFLOPs: 42.83 | 31: iteration 44230/ 52452 | consumed samples: 22645760 | consumed tokens: 46378516480 | elapsed time per iteration (s): 1.78 | learning rate: 3.091E-05 | global batch size: 512 | lm loss: 1.177566E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.435 | TFLOPs: 43.14 | 31: iteration 44240/ 52452 | consumed samples: 22650880 | consumed tokens: 46389002240 | elapsed time per iteration (s): 1.86 | learning rate: 3.088E-05 | global batch size: 512 | lm loss: 1.160792E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.487 | TFLOPs: 41.35 | 31: iteration 44250/ 52452 | consumed samples: 22656000 | consumed tokens: 46399488000 | elapsed time per iteration (s): 1.78 | learning rate: 3.086E-05 | global batch size: 512 | lm loss: 1.165914E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.055 | TFLOPs: 43.09 | 31: iteration 44260/ 52452 | consumed samples: 22661120 | consumed tokens: 46409973760 | elapsed time per iteration (s): 1.79 | learning rate: 3.083E-05 | global batch size: 512 | lm loss: 1.153297E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.562 | TFLOPs: 42.86 | 31: iteration 44270/ 52452 | consumed samples: 22666240 | consumed tokens: 46420459520 | elapsed time per iteration (s): 1.75 | learning rate: 3.080E-05 | global batch size: 512 | lm loss: 1.158406E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.754 | TFLOPs: 43.94 | 31: iteration 44280/ 52452 | consumed samples: 22671360 | consumed tokens: 46430945280 | elapsed time per iteration (s): 1.77 | learning rate: 3.078E-05 | global batch size: 512 | lm loss: 1.169042E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.365 | TFLOPs: 43.43 | 31: iteration 44290/ 52452 | consumed samples: 22676480 | consumed tokens: 46441431040 | elapsed time per iteration (s): 1.82 | learning rate: 3.075E-05 | global batch size: 512 | lm loss: 1.165842E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.358 | TFLOPs: 42.23 | 31: iteration 44300/ 52452 | consumed samples: 22681600 | consumed tokens: 46451916800 | elapsed time per iteration (s): 1.80 | learning rate: 3.073E-05 | global batch size: 512 | lm loss: 1.162994E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.867 | TFLOPs: 42.61 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44300 | lm loss value: 4.228290E+00 | lm loss PPL: 6.859980E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44300 | lm loss value: 4.093271E+00 | lm loss PPL: 5.993564E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44310/ 52452 | consumed samples: 22686720 | consumed tokens: 46462402560 | elapsed time per iteration (s): 9.22 | learning rate: 3.070E-05 | global batch size: 512 | lm loss: 1.158288E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.527 | TFLOPs: 8.33 | 31: iteration 44320/ 52452 | consumed samples: 22691840 | consumed tokens: 46472888320 | elapsed time per iteration (s): 1.76 | learning rate: 3.067E-05 | global batch size: 512 | lm loss: 1.156822E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.194 | TFLOPs: 43.71 | 31: iteration 44330/ 52452 | consumed samples: 22696960 | consumed tokens: 46483374080 | elapsed time per iteration (s): 1.84 | learning rate: 3.065E-05 | global batch size: 512 | lm loss: 1.155690E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.664 | TFLOPs: 41.83 | 31: iteration 44340/ 52452 | consumed samples: 22702080 | consumed tokens: 46493859840 | elapsed time per iteration (s): 1.84 | learning rate: 3.062E-05 | global batch size: 512 | lm loss: 1.145986E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.186 | TFLOPs: 41.75 | 31: iteration 44350/ 52452 | consumed samples: 22707200 | consumed tokens: 46504345600 | elapsed time per iteration (s): 1.97 | learning rate: 3.060E-05 | global batch size: 512 | lm loss: 1.160653E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 260.015 | TFLOPs: 39.03 | 31: iteration 44360/ 52452 | consumed samples: 22712320 | consumed tokens: 46514831360 | elapsed time per iteration (s): 1.80 | learning rate: 3.057E-05 | global batch size: 512 | lm loss: 1.159761E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.759 | TFLOPs: 42.74 | 31: iteration 44370/ 52452 | consumed samples: 22717440 | consumed tokens: 46525317120 | elapsed time per iteration (s): 1.78 | learning rate: 3.055E-05 | global batch size: 512 | lm loss: 1.161668E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.854 | TFLOPs: 43.06 | 31: iteration 44380/ 52452 | consumed samples: 22722560 | consumed tokens: 46535802880 | elapsed time per iteration (s): 1.83 | learning rate: 3.052E-05 | global batch size: 512 | lm loss: 1.159097E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.192 | TFLOPs: 41.91 | 31: iteration 44390/ 52452 | consumed samples: 22727680 | consumed tokens: 46546288640 | elapsed time per iteration (s): 1.77 | learning rate: 3.050E-05 | global batch size: 512 | lm loss: 1.159067E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.212 | TFLOPs: 43.41 | 31: iteration 44400/ 52452 | consumed samples: 22732800 | consumed tokens: 46556774400 | elapsed time per iteration (s): 1.82 | learning rate: 3.047E-05 | global batch size: 512 | lm loss: 1.163665E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.791 | TFLOPs: 42.30 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44400 | lm loss value: 4.211722E+00 | lm loss PPL: 6.747262E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44400 | lm loss value: 4.072340E+00 | lm loss PPL: 5.869417E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44410/ 52452 | consumed samples: 22737920 | consumed tokens: 46567260160 | elapsed time per iteration (s): 9.19 | learning rate: 3.044E-05 | global batch size: 512 | lm loss: 1.152706E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.734 | TFLOPs: 8.37 | 31: iteration 44420/ 52452 | consumed samples: 22743040 | consumed tokens: 46577745920 | elapsed time per iteration (s): 1.82 | learning rate: 3.042E-05 | global batch size: 512 | lm loss: 1.146134E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.908 | TFLOPs: 42.16 | 31: iteration 44430/ 52452 | consumed samples: 22748160 | consumed tokens: 46588231680 | elapsed time per iteration (s): 1.82 | learning rate: 3.039E-05 | global batch size: 512 | lm loss: 1.160822E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.907 | TFLOPs: 42.31 | 31: iteration 44440/ 52452 | consumed samples: 22753280 | consumed tokens: 46598717440 | elapsed time per iteration (s): 1.83 | learning rate: 3.037E-05 | global batch size: 512 | lm loss: 1.153279E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.824 | TFLOPs: 42.00 | 31: iteration 44450/ 52452 | consumed samples: 22758400 | consumed tokens: 46609203200 | elapsed time per iteration (s): 1.82 | learning rate: 3.034E-05 | global batch size: 512 | lm loss: 1.150101E+00 | grad norm: 0.175 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.043 | TFLOPs: 42.33 | 31: iteration 44460/ 52452 | consumed samples: 22763520 | consumed tokens: 46619688960 | elapsed time per iteration (s): 1.79 | learning rate: 3.032E-05 | global batch size: 512 | lm loss: 1.160289E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.019 | TFLOPs: 42.93 | 31: iteration 44470/ 52452 | consumed samples: 22768640 | consumed tokens: 46630174720 | elapsed time per iteration (s): 1.81 | learning rate: 3.029E-05 | global batch size: 512 | lm loss: 1.150998E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.765 | TFLOPs: 42.44 | 31: iteration 44480/ 52452 | consumed samples: 22773760 | consumed tokens: 46640660480 | elapsed time per iteration (s): 1.81 | learning rate: 3.027E-05 | global batch size: 512 | lm loss: 1.157750E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.378 | TFLOPs: 42.38 | 31: iteration 44490/ 52452 | consumed samples: 22778880 | consumed tokens: 46651146240 | elapsed time per iteration (s): 1.76 | learning rate: 3.024E-05 | global batch size: 512 | lm loss: 1.156963E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.540 | TFLOPs: 43.76 | 31: iteration 44500/ 52452 | consumed samples: 22784000 | consumed tokens: 46661632000 | elapsed time per iteration (s): 1.81 | learning rate: 3.022E-05 | global batch size: 512 | lm loss: 1.155109E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.325 | TFLOPs: 42.38 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44500 | lm loss value: 4.189136E+00 | lm loss PPL: 6.596577E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44500 | lm loss value: 4.061283E+00 | lm loss PPL: 5.804872E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44510/ 52452 | consumed samples: 22789120 | consumed tokens: 46672117760 | elapsed time per iteration (s): 9.19 | learning rate: 3.019E-05 | global batch size: 512 | lm loss: 1.159899E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.704 | TFLOPs: 8.36 | 31: iteration 44520/ 52452 | consumed samples: 22794240 | consumed tokens: 46682603520 | elapsed time per iteration (s): 1.74 | learning rate: 3.017E-05 | global batch size: 512 | lm loss: 1.153118E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.852 | TFLOPs: 44.11 | 31: iteration 44530/ 52452 | consumed samples: 22799360 | consumed tokens: 46693089280 | elapsed time per iteration (s): 1.77 | learning rate: 3.014E-05 | global batch size: 512 | lm loss: 1.165575E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.677 | TFLOPs: 43.33 | 31: iteration 44540/ 52452 | consumed samples: 22804480 | consumed tokens: 46703575040 | elapsed time per iteration (s): 1.78 | learning rate: 3.012E-05 | global batch size: 512 | lm loss: 1.155742E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.005 | TFLOPs: 43.23 | 31: iteration 44550/ 52452 | consumed samples: 22809600 | consumed tokens: 46714060800 | elapsed time per iteration (s): 1.83 | learning rate: 3.009E-05 | global batch size: 512 | lm loss: 1.152276E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.235 | TFLOPs: 42.06 | 31: iteration 44560/ 52452 | consumed samples: 22814720 | consumed tokens: 46724546560 | elapsed time per iteration (s): 1.79 | learning rate: 3.007E-05 | global batch size: 512 | lm loss: 1.160249E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.608 | TFLOPs: 43.02 | 31: iteration 44570/ 52452 | consumed samples: 22819840 | consumed tokens: 46735032320 | elapsed time per iteration (s): 1.82 | learning rate: 3.004E-05 | global batch size: 512 | lm loss: 1.147495E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.671 | TFLOPs: 42.28 | 31: iteration 44580/ 52452 | consumed samples: 22824960 | consumed tokens: 46745518080 | elapsed time per iteration (s): 1.77 | learning rate: 3.002E-05 | global batch size: 512 | lm loss: 1.140135E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.361 | TFLOPs: 43.43 | 31: iteration 44590/ 52452 | consumed samples: 22830080 | consumed tokens: 46756003840 | elapsed time per iteration (s): 1.79 | learning rate: 2.999E-05 | global batch size: 512 | lm loss: 1.152044E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.523 | TFLOPs: 43.01 | 31: iteration 44600/ 52452 | consumed samples: 22835200 | consumed tokens: 46766489600 | elapsed time per iteration (s): 1.78 | learning rate: 2.997E-05 | global batch size: 512 | lm loss: 1.143842E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.979 | TFLOPs: 43.22 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44600 | lm loss value: 4.228580E+00 | lm loss PPL: 6.861976E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44600 | lm loss value: 4.087447E+00 | lm loss PPL: 5.958758E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44610/ 52452 | consumed samples: 22840320 | consumed tokens: 46776975360 | elapsed time per iteration (s): 9.24 | learning rate: 2.994E-05 | global batch size: 512 | lm loss: 1.159743E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.412 | TFLOPs: 8.32 | 31: iteration 44620/ 52452 | consumed samples: 22845440 | consumed tokens: 46787461120 | elapsed time per iteration (s): 1.87 | learning rate: 2.992E-05 | global batch size: 512 | lm loss: 1.156057E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.364 | TFLOPs: 41.18 | 31: iteration 44630/ 52452 | consumed samples: 22850560 | consumed tokens: 46797946880 | elapsed time per iteration (s): 1.79 | learning rate: 2.989E-05 | global batch size: 512 | lm loss: 1.159878E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.519 | TFLOPs: 42.85 | 31: iteration 44640/ 52452 | consumed samples: 22855680 | consumed tokens: 46808432640 | elapsed time per iteration (s): 1.83 | learning rate: 2.987E-05 | global batch size: 512 | lm loss: 1.162513E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.545 | TFLOPs: 42.11 | 31: iteration 44650/ 52452 | consumed samples: 22860800 | consumed tokens: 46818918400 | elapsed time per iteration (s): 1.87 | learning rate: 2.984E-05 | global batch size: 512 | lm loss: 1.170715E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.508 | TFLOPs: 41.20 | 31: iteration 44660/ 52452 | consumed samples: 22865920 | consumed tokens: 46829404160 | elapsed time per iteration (s): 1.79 | learning rate: 2.982E-05 | global batch size: 512 | lm loss: 1.163447E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.615 | TFLOPs: 43.02 | 31: iteration 44670/ 52452 | consumed samples: 22871040 | consumed tokens: 46839889920 | elapsed time per iteration (s): 1.90 | learning rate: 2.979E-05 | global batch size: 512 | lm loss: 1.157068E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.566 | TFLOPs: 40.46 | 31: iteration 44680/ 52452 | consumed samples: 22876160 | consumed tokens: 46850375680 | elapsed time per iteration (s): 1.78 | learning rate: 2.977E-05 | global batch size: 512 | lm loss: 1.156991E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.448 | TFLOPs: 43.14 | 31: iteration 44690/ 52452 | consumed samples: 22881280 | consumed tokens: 46860861440 | elapsed time per iteration (s): 1.86 | learning rate: 2.974E-05 | global batch size: 512 | lm loss: 1.166539E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.675 | TFLOPs: 41.23 | 31: iteration 44700/ 52452 | consumed samples: 22886400 | consumed tokens: 46871347200 | elapsed time per iteration (s): 1.77 | learning rate: 2.972E-05 | global batch size: 512 | lm loss: 1.157527E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.941 | TFLOPs: 43.52 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44700 | lm loss value: 4.219374E+00 | lm loss PPL: 6.799092E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44700 | lm loss value: 4.078819E+00 | lm loss PPL: 5.907568E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44710/ 52452 | consumed samples: 22891520 | consumed tokens: 46881832960 | elapsed time per iteration (s): 9.20 | learning rate: 2.969E-05 | global batch size: 512 | lm loss: 1.162418E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.649 | TFLOPs: 8.35 | 31: iteration 44720/ 52452 | consumed samples: 22896640 | consumed tokens: 46892318720 | elapsed time per iteration (s): 1.77 | learning rate: 2.967E-05 | global batch size: 512 | lm loss: 1.171394E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.834 | TFLOPs: 43.35 | 31: iteration 44730/ 52452 | consumed samples: 22901760 | consumed tokens: 46902804480 | elapsed time per iteration (s): 1.76 | learning rate: 2.964E-05 | global batch size: 512 | lm loss: 1.158374E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.252 | TFLOPs: 43.57 | 31: iteration 44740/ 52452 | consumed samples: 22906880 | consumed tokens: 46913290240 | elapsed time per iteration (s): 1.78 | learning rate: 2.962E-05 | global batch size: 512 | lm loss: 1.157101E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.371 | TFLOPs: 43.13 | 31: iteration 44750/ 52452 | consumed samples: 22912000 | consumed tokens: 46923776000 | elapsed time per iteration (s): 1.77 | learning rate: 2.960E-05 | global batch size: 512 | lm loss: 1.154038E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.877 | TFLOPs: 43.51 | 31: iteration 44760/ 52452 | consumed samples: 22917120 | consumed tokens: 46934261760 | elapsed time per iteration (s): 1.86 | learning rate: 2.957E-05 | global batch size: 512 | lm loss: 1.160871E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.961 | TFLOPs: 41.42 | 31: iteration 44770/ 52452 | consumed samples: 22922240 | consumed tokens: 46944747520 | elapsed time per iteration (s): 1.75 | learning rate: 2.955E-05 | global batch size: 512 | lm loss: 1.160498E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.974 | TFLOPs: 43.82 | 31: iteration 44780/ 52452 | consumed samples: 22927360 | consumed tokens: 46955233280 | elapsed time per iteration (s): 1.78 | learning rate: 2.952E-05 | global batch size: 512 | lm loss: 1.153917E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.431 | TFLOPs: 43.29 | 31: iteration 44790/ 52452 | consumed samples: 22932480 | consumed tokens: 46965719040 | elapsed time per iteration (s): 1.77 | learning rate: 2.950E-05 | global batch size: 512 | lm loss: 1.159935E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.916 | TFLOPs: 43.51 | 31: iteration 44800/ 52452 | consumed samples: 22937600 | consumed tokens: 46976204800 | elapsed time per iteration (s): 1.78 | learning rate: 2.947E-05 | global batch size: 512 | lm loss: 1.164423E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.161 | TFLOPs: 43.10 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44800 | lm loss value: 4.146015E+00 | lm loss PPL: 6.318170E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44800 | lm loss value: 4.031383E+00 | lm loss PPL: 5.633878E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44810/ 52452 | consumed samples: 22942720 | consumed tokens: 46986690560 | elapsed time per iteration (s): 9.14 | learning rate: 2.945E-05 | global batch size: 512 | lm loss: 1.155607E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.036 | TFLOPs: 8.41 | 31: iteration 44820/ 52452 | consumed samples: 22947840 | consumed tokens: 46997176320 | elapsed time per iteration (s): 1.74 | learning rate: 2.942E-05 | global batch size: 512 | lm loss: 1.157800E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.304 | TFLOPs: 44.17 | 31: iteration 44830/ 52452 | consumed samples: 22952960 | consumed tokens: 47007662080 | elapsed time per iteration (s): 1.78 | learning rate: 2.940E-05 | global batch size: 512 | lm loss: 1.154546E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.611 | TFLOPs: 43.17 | 31: iteration 44840/ 52452 | consumed samples: 22958080 | consumed tokens: 47018147840 | elapsed time per iteration (s): 1.76 | learning rate: 2.938E-05 | global batch size: 512 | lm loss: 1.156661E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.725 | TFLOPs: 43.64 | 31: iteration 44850/ 52452 | consumed samples: 22963200 | consumed tokens: 47028633600 | elapsed time per iteration (s): 1.74 | learning rate: 2.935E-05 | global batch size: 512 | lm loss: 1.157579E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.558 | TFLOPs: 44.21 | 31: iteration 44860/ 52452 | consumed samples: 22968320 | consumed tokens: 47039119360 | elapsed time per iteration (s): 1.73 | learning rate: 2.933E-05 | global batch size: 512 | lm loss: 1.144135E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.361 | TFLOPs: 44.33 | 31: iteration 44870/ 52452 | consumed samples: 22973440 | consumed tokens: 47049605120 | elapsed time per iteration (s): 1.76 | learning rate: 2.930E-05 | global batch size: 512 | lm loss: 1.160226E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.960 | TFLOPs: 43.67 | 31: iteration 44880/ 52452 | consumed samples: 22978560 | consumed tokens: 47060090880 | elapsed time per iteration (s): 1.74 | learning rate: 2.928E-05 | global batch size: 512 | lm loss: 1.166750E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.653 | TFLOPs: 44.23 | 31: iteration 44890/ 52452 | consumed samples: 22983680 | consumed tokens: 47070576640 | elapsed time per iteration (s): 1.75 | learning rate: 2.926E-05 | global batch size: 512 | lm loss: 1.140866E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.235 | TFLOPs: 43.86 | 31: iteration 44900/ 52452 | consumed samples: 22988800 | consumed tokens: 47081062400 | elapsed time per iteration (s): 1.77 | learning rate: 2.923E-05 | global batch size: 512 | lm loss: 1.151757E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.143 | TFLOPs: 43.40 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 44900 | lm loss value: 4.178596E+00 | lm loss PPL: 6.527415E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 44900 | lm loss value: 4.043117E+00 | lm loss PPL: 5.700375E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 44910/ 52452 | consumed samples: 22993920 | consumed tokens: 47091548160 | elapsed time per iteration (s): 9.15 | learning rate: 2.921E-05 | global batch size: 512 | lm loss: 1.161324E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.980 | TFLOPs: 8.40 | 31: iteration 44920/ 52452 | consumed samples: 22999040 | consumed tokens: 47102033920 | elapsed time per iteration (s): 1.77 | learning rate: 2.918E-05 | global batch size: 512 | lm loss: 1.141557E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.839 | TFLOPs: 43.35 | 31: iteration 44930/ 52452 | consumed samples: 23004160 | consumed tokens: 47112519680 | elapsed time per iteration (s): 1.75 | learning rate: 2.916E-05 | global batch size: 512 | lm loss: 1.159108E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.247 | TFLOPs: 43.86 | 31: iteration 44940/ 52452 | consumed samples: 23009280 | consumed tokens: 47123005440 | elapsed time per iteration (s): 1.81 | learning rate: 2.914E-05 | global batch size: 512 | lm loss: 1.156640E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.266 | TFLOPs: 42.52 | 31: iteration 44950/ 52452 | consumed samples: 23014400 | consumed tokens: 47133491200 | elapsed time per iteration (s): 1.77 | learning rate: 2.911E-05 | global batch size: 512 | lm loss: 1.153021E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.266 | TFLOPs: 43.42 | 31: iteration 44960/ 52452 | consumed samples: 23019520 | consumed tokens: 47143976960 | elapsed time per iteration (s): 1.83 | learning rate: 2.909E-05 | global batch size: 512 | lm loss: 1.148246E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.082 | TFLOPs: 41.89 | 31: iteration 44970/ 52452 | consumed samples: 23024640 | consumed tokens: 47154462720 | elapsed time per iteration (s): 1.80 | learning rate: 2.906E-05 | global batch size: 512 | lm loss: 1.154090E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.659 | TFLOPs: 42.58 | 31: iteration 44980/ 52452 | consumed samples: 23029760 | consumed tokens: 47164948480 | elapsed time per iteration (s): 1.76 | learning rate: 2.904E-05 | global batch size: 512 | lm loss: 1.166283E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.151 | TFLOPs: 43.70 | 31: iteration 44990/ 52452 | consumed samples: 23034880 | consumed tokens: 47175434240 | elapsed time per iteration (s): 1.74 | learning rate: 2.902E-05 | global batch size: 512 | lm loss: 1.163190E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.032 | TFLOPs: 44.13 | 31: iteration 45000/ 52452 | consumed samples: 23040000 | consumed tokens: 47185920000 | elapsed time per iteration (s): 1.73 | learning rate: 2.899E-05 | global batch size: 512 | lm loss: 1.167491E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.837 | TFLOPs: 44.40 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45000 | lm loss value: 4.182378E+00 | lm loss PPL: 6.552147E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45000 | lm loss value: 4.041381E+00 | lm loss PPL: 5.690489E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45010/ 52452 | consumed samples: 23045120 | consumed tokens: 47196405760 | elapsed time per iteration (s): 9.17 | learning rate: 2.897E-05 | global batch size: 512 | lm loss: 1.152035E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.812 | TFLOPs: 8.38 | 31: iteration 45020/ 52452 | consumed samples: 23050240 | consumed tokens: 47206891520 | elapsed time per iteration (s): 1.77 | learning rate: 2.895E-05 | global batch size: 512 | lm loss: 1.156909E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.595 | TFLOPs: 43.32 | 31: iteration 45030/ 52452 | consumed samples: 23055360 | consumed tokens: 47217377280 | elapsed time per iteration (s): 1.77 | learning rate: 2.892E-05 | global batch size: 512 | lm loss: 1.166073E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.507 | TFLOPs: 43.30 | 31: iteration 45040/ 52452 | consumed samples: 23060480 | consumed tokens: 47227863040 | elapsed time per iteration (s): 1.76 | learning rate: 2.890E-05 | global batch size: 512 | lm loss: 1.150565E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.542 | TFLOPs: 43.76 | 31: iteration 45050/ 52452 | consumed samples: 23065600 | consumed tokens: 47238348800 | elapsed time per iteration (s): 1.75 | learning rate: 2.887E-05 | global batch size: 512 | lm loss: 1.168304E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.167 | TFLOPs: 43.85 | 31: iteration 45060/ 52452 | consumed samples: 23070720 | consumed tokens: 47248834560 | elapsed time per iteration (s): 1.75 | learning rate: 2.885E-05 | global batch size: 512 | lm loss: 1.151673E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.143 | TFLOPs: 43.85 | 31: iteration 45070/ 52452 | consumed samples: 23075840 | consumed tokens: 47259320320 | elapsed time per iteration (s): 1.78 | learning rate: 2.883E-05 | global batch size: 512 | lm loss: 1.152707E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.772 | TFLOPs: 43.19 | 31: iteration 45080/ 52452 | consumed samples: 23080960 | consumed tokens: 47269806080 | elapsed time per iteration (s): 1.77 | learning rate: 2.880E-05 | global batch size: 512 | lm loss: 1.154018E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.931 | TFLOPs: 43.37 | 31: iteration 45090/ 52452 | consumed samples: 23086080 | consumed tokens: 47280291840 | elapsed time per iteration (s): 1.76 | learning rate: 2.878E-05 | global batch size: 512 | lm loss: 1.151950E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.242 | TFLOPs: 43.71 | 31: iteration 45100/ 52452 | consumed samples: 23091200 | consumed tokens: 47290777600 | elapsed time per iteration (s): 1.79 | learning rate: 2.876E-05 | global batch size: 512 | lm loss: 1.154301E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.473 | TFLOPs: 42.85 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45100 | lm loss value: 4.183325E+00 | lm loss PPL: 6.558358E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45100 | lm loss value: 4.057487E+00 | lm loss PPL: 5.782883E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45110/ 52452 | consumed samples: 23096320 | consumed tokens: 47301263360 | elapsed time per iteration (s): 9.16 | learning rate: 2.873E-05 | global batch size: 512 | lm loss: 1.161483E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.903 | TFLOPs: 8.39 | 31: iteration 45120/ 52452 | consumed samples: 23101440 | consumed tokens: 47311749120 | elapsed time per iteration (s): 1.79 | learning rate: 2.871E-05 | global batch size: 512 | lm loss: 1.159563E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.437 | TFLOPs: 42.84 | 31: iteration 45130/ 52452 | consumed samples: 23106560 | consumed tokens: 47322234880 | elapsed time per iteration (s): 1.79 | learning rate: 2.869E-05 | global batch size: 512 | lm loss: 1.149541E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.194 | TFLOPs: 42.96 | 31: iteration 45140/ 52452 | consumed samples: 23111680 | consumed tokens: 47332720640 | elapsed time per iteration (s): 1.85 | learning rate: 2.866E-05 | global batch size: 512 | lm loss: 1.150673E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.567 | TFLOPs: 41.51 | 31: iteration 45150/ 52452 | consumed samples: 23116800 | consumed tokens: 47343206400 | elapsed time per iteration (s): 1.78 | learning rate: 2.864E-05 | global batch size: 512 | lm loss: 1.161920E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.433 | TFLOPs: 43.29 | 31: iteration 45160/ 52452 | consumed samples: 23121920 | consumed tokens: 47353692160 | elapsed time per iteration (s): 1.74 | learning rate: 2.862E-05 | global batch size: 512 | lm loss: 1.149148E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.096 | TFLOPs: 44.29 | 31: iteration 45170/ 52452 | consumed samples: 23127040 | consumed tokens: 47364177920 | elapsed time per iteration (s): 1.76 | learning rate: 2.859E-05 | global batch size: 512 | lm loss: 1.144265E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.353 | TFLOPs: 43.58 | 31: iteration 45180/ 52452 | consumed samples: 23132160 | consumed tokens: 47374663680 | elapsed time per iteration (s): 1.81 | learning rate: 2.857E-05 | global batch size: 512 | lm loss: 1.144142E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.017 | TFLOPs: 42.48 | 31: iteration 45190/ 52452 | consumed samples: 23137280 | consumed tokens: 47385149440 | elapsed time per iteration (s): 1.83 | learning rate: 2.855E-05 | global batch size: 512 | lm loss: 1.153815E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.310 | TFLOPs: 41.92 | 31: iteration 45200/ 52452 | consumed samples: 23142400 | consumed tokens: 47395635200 | elapsed time per iteration (s): 1.84 | learning rate: 2.852E-05 | global batch size: 512 | lm loss: 1.139011E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.617 | TFLOPs: 41.67 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45200 | lm loss value: 4.158463E+00 | lm loss PPL: 6.397315E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45200 | lm loss value: 4.037096E+00 | lm loss PPL: 5.666153E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45210/ 52452 | consumed samples: 23147520 | consumed tokens: 47406120960 | elapsed time per iteration (s): 9.21 | learning rate: 2.850E-05 | global batch size: 512 | lm loss: 1.153973E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.606 | TFLOPs: 8.35 | 31: iteration 45220/ 52452 | consumed samples: 23152640 | consumed tokens: 47416606720 | elapsed time per iteration (s): 1.79 | learning rate: 2.848E-05 | global batch size: 512 | lm loss: 1.165093E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.879 | TFLOPs: 42.91 | 31: iteration 45230/ 52452 | consumed samples: 23157760 | consumed tokens: 47427092480 | elapsed time per iteration (s): 1.78 | learning rate: 2.846E-05 | global batch size: 512 | lm loss: 1.152290E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.940 | TFLOPs: 43.22 | 31: iteration 45240/ 52452 | consumed samples: 23162880 | consumed tokens: 47437578240 | elapsed time per iteration (s): 1.82 | learning rate: 2.843E-05 | global batch size: 512 | lm loss: 1.157252E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.934 | TFLOPs: 42.17 | 31: iteration 45250/ 52452 | consumed samples: 23168000 | consumed tokens: 47448064000 | elapsed time per iteration (s): 1.88 | learning rate: 2.841E-05 | global batch size: 512 | lm loss: 1.162074E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.909 | TFLOPs: 40.96 | 31: iteration 45260/ 52452 | consumed samples: 23173120 | consumed tokens: 47458549760 | elapsed time per iteration (s): 1.79 | learning rate: 2.839E-05 | global batch size: 512 | lm loss: 1.152173E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.780 | TFLOPs: 42.89 | 31: iteration 45270/ 52452 | consumed samples: 23178240 | consumed tokens: 47469035520 | elapsed time per iteration (s): 1.80 | learning rate: 2.836E-05 | global batch size: 512 | lm loss: 1.154213E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.935 | TFLOPs: 42.62 | 31: iteration 45280/ 52452 | consumed samples: 23183360 | consumed tokens: 47479521280 | elapsed time per iteration (s): 1.78 | learning rate: 2.834E-05 | global batch size: 512 | lm loss: 1.162836E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.178 | TFLOPs: 43.10 | 31: iteration 45290/ 52452 | consumed samples: 23188480 | consumed tokens: 47490007040 | elapsed time per iteration (s): 1.74 | learning rate: 2.832E-05 | global batch size: 512 | lm loss: 1.156394E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.588 | TFLOPs: 44.07 | 31: iteration 45300/ 52452 | consumed samples: 23193600 | consumed tokens: 47500492800 | elapsed time per iteration (s): 1.74 | learning rate: 2.829E-05 | global batch size: 512 | lm loss: 1.147774E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.225 | TFLOPs: 44.16 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45300 | lm loss value: 4.223147E+00 | lm loss PPL: 6.824795E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45300 | lm loss value: 4.076174E+00 | lm loss PPL: 5.891963E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45310/ 52452 | consumed samples: 23198720 | consumed tokens: 47510978560 | elapsed time per iteration (s): 9.18 | learning rate: 2.827E-05 | global batch size: 512 | lm loss: 1.159092E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.764 | TFLOPs: 8.37 | 31: iteration 45320/ 52452 | consumed samples: 23203840 | consumed tokens: 47521464320 | elapsed time per iteration (s): 1.75 | learning rate: 2.825E-05 | global batch size: 512 | lm loss: 1.140646E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.274 | TFLOPs: 44.02 | 31: iteration 45330/ 52452 | consumed samples: 23208960 | consumed tokens: 47531950080 | elapsed time per iteration (s): 1.80 | learning rate: 2.823E-05 | global batch size: 512 | lm loss: 1.138968E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.341 | TFLOPs: 42.68 | 31: iteration 45340/ 52452 | consumed samples: 23214080 | consumed tokens: 47542435840 | elapsed time per iteration (s): 1.92 | learning rate: 2.820E-05 | global batch size: 512 | lm loss: 1.155542E+00 | grad norm: 0.150 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 266.992 | TFLOPs: 40.07 | 31: iteration 45350/ 52452 | consumed samples: 23219200 | consumed tokens: 47552921600 | elapsed time per iteration (s): 1.83 | learning rate: 2.818E-05 | global batch size: 512 | lm loss: 1.149762E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.436 | TFLOPs: 42.09 | 31: iteration 45360/ 52452 | consumed samples: 23224320 | consumed tokens: 47563407360 | elapsed time per iteration (s): 1.76 | learning rate: 2.816E-05 | global batch size: 512 | lm loss: 1.156928E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.137 | TFLOPs: 43.70 | 31: iteration 45370/ 52452 | consumed samples: 23229440 | consumed tokens: 47573893120 | elapsed time per iteration (s): 1.79 | learning rate: 2.814E-05 | global batch size: 512 | lm loss: 1.165524E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.404 | TFLOPs: 42.99 | 31: iteration 45380/ 52452 | consumed samples: 23234560 | consumed tokens: 47584378880 | elapsed time per iteration (s): 1.82 | learning rate: 2.811E-05 | global batch size: 512 | lm loss: 1.158835E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.624 | TFLOPs: 42.27 | 31: iteration 45390/ 52452 | consumed samples: 23239680 | consumed tokens: 47594864640 | elapsed time per iteration (s): 1.77 | learning rate: 2.809E-05 | global batch size: 512 | lm loss: 1.153835E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.022 | TFLOPs: 43.38 | 31: iteration 45400/ 52452 | consumed samples: 23244800 | consumed tokens: 47605350400 | elapsed time per iteration (s): 1.79 | learning rate: 2.807E-05 | global batch size: 512 | lm loss: 1.153205E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.301 | TFLOPs: 42.97 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45400 | lm loss value: 4.234613E+00 | lm loss PPL: 6.903495E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45400 | lm loss value: 4.078285E+00 | lm loss PPL: 5.904414E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45410/ 52452 | consumed samples: 23249920 | consumed tokens: 47615836160 | elapsed time per iteration (s): 9.30 | learning rate: 2.805E-05 | global batch size: 512 | lm loss: 1.158006E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.076 | TFLOPs: 8.27 | 31: iteration 45420/ 52452 | consumed samples: 23255040 | consumed tokens: 47626321920 | elapsed time per iteration (s): 1.78 | learning rate: 2.802E-05 | global batch size: 512 | lm loss: 1.149894E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.732 | TFLOPs: 43.19 | 31: iteration 45430/ 52452 | consumed samples: 23260160 | consumed tokens: 47636807680 | elapsed time per iteration (s): 1.76 | learning rate: 2.800E-05 | global batch size: 512 | lm loss: 1.149464E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.194 | TFLOPs: 43.71 | 31: iteration 45440/ 52452 | consumed samples: 23265280 | consumed tokens: 47647293440 | elapsed time per iteration (s): 1.86 | learning rate: 2.798E-05 | global batch size: 512 | lm loss: 1.153023E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.727 | TFLOPs: 41.39 | 31: iteration 45450/ 52452 | consumed samples: 23270400 | consumed tokens: 47657779200 | elapsed time per iteration (s): 1.75 | learning rate: 2.796E-05 | global batch size: 512 | lm loss: 1.157113E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.410 | TFLOPs: 43.89 | 31: iteration 45460/ 52452 | consumed samples: 23275520 | consumed tokens: 47668264960 | elapsed time per iteration (s): 1.79 | learning rate: 2.793E-05 | global batch size: 512 | lm loss: 1.155618E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.705 | TFLOPs: 43.03 | 31: iteration 45470/ 52452 | consumed samples: 23280640 | consumed tokens: 47678750720 | elapsed time per iteration (s): 1.81 | learning rate: 2.791E-05 | global batch size: 512 | lm loss: 1.154179E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.922 | TFLOPs: 42.47 | 31: iteration 45480/ 52452 | consumed samples: 23285760 | consumed tokens: 47689236480 | elapsed time per iteration (s): 1.76 | learning rate: 2.789E-05 | global batch size: 512 | lm loss: 1.159323E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.495 | TFLOPs: 43.60 | 31: iteration 45490/ 52452 | consumed samples: 23290880 | consumed tokens: 47699722240 | elapsed time per iteration (s): 1.82 | learning rate: 2.787E-05 | global batch size: 512 | lm loss: 1.146036E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.566 | TFLOPs: 42.26 | 31: iteration 45500/ 52452 | consumed samples: 23296000 | consumed tokens: 47710208000 | elapsed time per iteration (s): 1.81 | learning rate: 2.784E-05 | global batch size: 512 | lm loss: 1.150151E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.273 | TFLOPs: 42.37 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45500 | lm loss value: 4.232523E+00 | lm loss PPL: 6.889082E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45500 | lm loss value: 4.082748E+00 | lm loss PPL: 5.930822E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45510/ 52452 | consumed samples: 23301120 | consumed tokens: 47720693760 | elapsed time per iteration (s): 9.28 | learning rate: 2.782E-05 | global batch size: 512 | lm loss: 1.157230E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.171 | TFLOPs: 8.28 | 31: iteration 45520/ 52452 | consumed samples: 23306240 | consumed tokens: 47731179520 | elapsed time per iteration (s): 1.78 | learning rate: 2.780E-05 | global batch size: 512 | lm loss: 1.162688E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.269 | TFLOPs: 43.12 | 31: iteration 45530/ 52452 | consumed samples: 23311360 | consumed tokens: 47741665280 | elapsed time per iteration (s): 1.86 | learning rate: 2.778E-05 | global batch size: 512 | lm loss: 1.149805E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.879 | TFLOPs: 41.26 | 31: iteration 45540/ 52452 | consumed samples: 23316480 | consumed tokens: 47752151040 | elapsed time per iteration (s): 1.82 | learning rate: 2.776E-05 | global batch size: 512 | lm loss: 1.160529E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.847 | TFLOPs: 42.30 | 31: iteration 45550/ 52452 | consumed samples: 23321600 | consumed tokens: 47762636800 | elapsed time per iteration (s): 1.79 | learning rate: 2.773E-05 | global batch size: 512 | lm loss: 1.140919E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.456 | TFLOPs: 42.85 | 31: iteration 45560/ 52452 | consumed samples: 23326720 | consumed tokens: 47773122560 | elapsed time per iteration (s): 1.83 | learning rate: 2.771E-05 | global batch size: 512 | lm loss: 1.152270E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.071 | TFLOPs: 41.89 | 31: iteration 45570/ 52452 | consumed samples: 23331840 | consumed tokens: 47783608320 | elapsed time per iteration (s): 1.73 | learning rate: 2.769E-05 | global batch size: 512 | lm loss: 1.153313E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.558 | TFLOPs: 44.51 | 31: iteration 45580/ 52452 | consumed samples: 23336960 | consumed tokens: 47794094080 | elapsed time per iteration (s): 1.79 | learning rate: 2.767E-05 | global batch size: 512 | lm loss: 1.154405E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.858 | TFLOPs: 42.91 | 31: iteration 45590/ 52452 | consumed samples: 23342080 | consumed tokens: 47804579840 | elapsed time per iteration (s): 1.75 | learning rate: 2.765E-05 | global batch size: 512 | lm loss: 1.144996E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.211 | TFLOPs: 44.01 | 31: iteration 45600/ 52452 | consumed samples: 23347200 | consumed tokens: 47815065600 | elapsed time per iteration (s): 1.86 | learning rate: 2.762E-05 | global batch size: 512 | lm loss: 1.156123E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.586 | TFLOPs: 41.21 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45600 | lm loss value: 4.239396E+00 | lm loss PPL: 6.936595E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45600 | lm loss value: 4.088467E+00 | lm loss PPL: 5.964836E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45610/ 52452 | consumed samples: 23352320 | consumed tokens: 47825551360 | elapsed time per iteration (s): 9.24 | learning rate: 2.760E-05 | global batch size: 512 | lm loss: 1.147319E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.429 | TFLOPs: 8.32 | 31: iteration 45620/ 52452 | consumed samples: 23357440 | consumed tokens: 47836037120 | elapsed time per iteration (s): 1.82 | learning rate: 2.758E-05 | global batch size: 512 | lm loss: 1.149266E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.619 | TFLOPs: 42.12 | 31: iteration 45630/ 52452 | consumed samples: 23362560 | consumed tokens: 47846522880 | elapsed time per iteration (s): 1.85 | learning rate: 2.756E-05 | global batch size: 512 | lm loss: 1.152910E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.039 | TFLOPs: 41.58 | 31: iteration 45640/ 52452 | consumed samples: 23367680 | consumed tokens: 47857008640 | elapsed time per iteration (s): 1.82 | learning rate: 2.754E-05 | global batch size: 512 | lm loss: 1.160136E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.310 | TFLOPs: 42.22 | 31: iteration 45650/ 52452 | consumed samples: 23372800 | consumed tokens: 47867494400 | elapsed time per iteration (s): 1.86 | learning rate: 2.751E-05 | global batch size: 512 | lm loss: 1.147513E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.138 | TFLOPs: 41.30 | 31: iteration 45660/ 52452 | consumed samples: 23377920 | consumed tokens: 47877980160 | elapsed time per iteration (s): 1.75 | learning rate: 2.749E-05 | global batch size: 512 | lm loss: 1.148653E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.044 | TFLOPs: 43.98 | 31: iteration 45670/ 52452 | consumed samples: 23383040 | consumed tokens: 47888465920 | elapsed time per iteration (s): 1.78 | learning rate: 2.747E-05 | global batch size: 512 | lm loss: 1.159172E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.470 | TFLOPs: 43.15 | 31: iteration 45680/ 52452 | consumed samples: 23388160 | consumed tokens: 47898951680 | elapsed time per iteration (s): 1.76 | learning rate: 2.745E-05 | global batch size: 512 | lm loss: 1.151804E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.341 | TFLOPs: 43.58 | 31: iteration 45690/ 52452 | consumed samples: 23393280 | consumed tokens: 47909437440 | elapsed time per iteration (s): 1.87 | learning rate: 2.743E-05 | global batch size: 512 | lm loss: 1.158148E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.394 | TFLOPs: 41.03 | 31: iteration 45700/ 52452 | consumed samples: 23398400 | consumed tokens: 47919923200 | elapsed time per iteration (s): 1.74 | learning rate: 2.741E-05 | global batch size: 512 | lm loss: 1.150115E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.662 | TFLOPs: 44.23 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45700 | lm loss value: 4.230582E+00 | lm loss PPL: 6.875722E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45700 | lm loss value: 4.070342E+00 | lm loss PPL: 5.857697E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45710/ 52452 | consumed samples: 23403520 | consumed tokens: 47930408960 | elapsed time per iteration (s): 9.13 | learning rate: 2.738E-05 | global batch size: 512 | lm loss: 1.159582E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.107 | TFLOPs: 8.42 | 31: iteration 45720/ 52452 | consumed samples: 23408640 | consumed tokens: 47940894720 | elapsed time per iteration (s): 1.85 | learning rate: 2.736E-05 | global batch size: 512 | lm loss: 1.155099E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.369 | TFLOPs: 41.63 | 31: iteration 45730/ 52452 | consumed samples: 23413760 | consumed tokens: 47951380480 | elapsed time per iteration (s): 1.80 | learning rate: 2.734E-05 | global batch size: 512 | lm loss: 1.146841E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.082 | TFLOPs: 42.64 | 31: iteration 45740/ 52452 | consumed samples: 23418880 | consumed tokens: 47961866240 | elapsed time per iteration (s): 1.77 | learning rate: 2.732E-05 | global batch size: 512 | lm loss: 1.154104E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.345 | TFLOPs: 43.43 | 31: iteration 45750/ 52452 | consumed samples: 23424000 | consumed tokens: 47972352000 | elapsed time per iteration (s): 1.83 | learning rate: 2.730E-05 | global batch size: 512 | lm loss: 1.156721E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.103 | TFLOPs: 42.04 | 31: iteration 45760/ 52452 | consumed samples: 23429120 | consumed tokens: 47982837760 | elapsed time per iteration (s): 1.76 | learning rate: 2.728E-05 | global batch size: 512 | lm loss: 1.151740E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.507 | TFLOPs: 43.75 | 31: iteration 45770/ 52452 | consumed samples: 23434240 | consumed tokens: 47993323520 | elapsed time per iteration (s): 1.75 | learning rate: 2.725E-05 | global batch size: 512 | lm loss: 1.160585E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.815 | TFLOPs: 43.80 | 31: iteration 45780/ 52452 | consumed samples: 23439360 | consumed tokens: 48003809280 | elapsed time per iteration (s): 1.74 | learning rate: 2.723E-05 | global batch size: 512 | lm loss: 1.157308E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.782 | TFLOPs: 44.25 | 31: iteration 45790/ 52452 | consumed samples: 23444480 | consumed tokens: 48014295040 | elapsed time per iteration (s): 1.85 | learning rate: 2.721E-05 | global batch size: 512 | lm loss: 1.144455E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.157 | TFLOPs: 41.45 | 31: iteration 45800/ 52452 | consumed samples: 23449600 | consumed tokens: 48024780800 | elapsed time per iteration (s): 1.77 | learning rate: 2.719E-05 | global batch size: 512 | lm loss: 1.161447E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.662 | TFLOPs: 43.33 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45800 | lm loss value: 4.244157E+00 | lm loss PPL: 6.969700E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45800 | lm loss value: 4.081888E+00 | lm loss PPL: 5.925725E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45810/ 52452 | consumed samples: 23454720 | consumed tokens: 48035266560 | elapsed time per iteration (s): 9.13 | learning rate: 2.717E-05 | global batch size: 512 | lm loss: 1.158215E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.063 | TFLOPs: 8.41 | 31: iteration 45820/ 52452 | consumed samples: 23459840 | consumed tokens: 48045752320 | elapsed time per iteration (s): 1.78 | learning rate: 2.715E-05 | global batch size: 512 | lm loss: 1.148417E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.267 | TFLOPs: 43.27 | 31: iteration 45830/ 52452 | consumed samples: 23464960 | consumed tokens: 48056238080 | elapsed time per iteration (s): 1.74 | learning rate: 2.713E-05 | global batch size: 512 | lm loss: 1.150860E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.708 | TFLOPs: 44.23 | 31: iteration 45840/ 52452 | consumed samples: 23470080 | consumed tokens: 48066723840 | elapsed time per iteration (s): 1.81 | learning rate: 2.711E-05 | global batch size: 512 | lm loss: 1.156834E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.813 | TFLOPs: 42.45 | 31: iteration 45850/ 52452 | consumed samples: 23475200 | consumed tokens: 48077209600 | elapsed time per iteration (s): 1.80 | learning rate: 2.708E-05 | global batch size: 512 | lm loss: 1.151562E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.721 | TFLOPs: 42.74 | 31: iteration 45860/ 52452 | consumed samples: 23480320 | consumed tokens: 48087695360 | elapsed time per iteration (s): 1.86 | learning rate: 2.706E-05 | global batch size: 512 | lm loss: 1.145632E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.004 | TFLOPs: 41.28 | 31: iteration 45870/ 52452 | consumed samples: 23485440 | consumed tokens: 48098181120 | elapsed time per iteration (s): 1.75 | learning rate: 2.704E-05 | global batch size: 512 | lm loss: 1.153870E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.267 | TFLOPs: 43.87 | 31: iteration 45880/ 52452 | consumed samples: 23490560 | consumed tokens: 48108666880 | elapsed time per iteration (s): 1.78 | learning rate: 2.702E-05 | global batch size: 512 | lm loss: 1.142712E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.442 | TFLOPs: 43.14 | 31: iteration 45890/ 52452 | consumed samples: 23495680 | consumed tokens: 48119152640 | elapsed time per iteration (s): 1.88 | learning rate: 2.700E-05 | global batch size: 512 | lm loss: 1.152229E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.384 | TFLOPs: 40.88 | 31: iteration 45900/ 52452 | consumed samples: 23500800 | consumed tokens: 48129638400 | elapsed time per iteration (s): 1.79 | learning rate: 2.698E-05 | global batch size: 512 | lm loss: 1.153510E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.401 | TFLOPs: 42.84 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 45900 | lm loss value: 4.210485E+00 | lm loss PPL: 6.738925E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 45900 | lm loss value: 4.056120E+00 | lm loss PPL: 5.774983E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 45910/ 52452 | consumed samples: 23505920 | consumed tokens: 48140124160 | elapsed time per iteration (s): 9.21 | learning rate: 2.696E-05 | global batch size: 512 | lm loss: 1.156774E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.599 | TFLOPs: 8.35 | 31: iteration 45920/ 52452 | consumed samples: 23511040 | consumed tokens: 48150609920 | elapsed time per iteration (s): 1.82 | learning rate: 2.694E-05 | global batch size: 512 | lm loss: 1.147987E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.878 | TFLOPs: 42.31 | 31: iteration 45930/ 52452 | consumed samples: 23516160 | consumed tokens: 48161095680 | elapsed time per iteration (s): 1.77 | learning rate: 2.692E-05 | global batch size: 512 | lm loss: 1.162962E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.778 | TFLOPs: 43.34 | 31: iteration 45940/ 52452 | consumed samples: 23521280 | consumed tokens: 48171581440 | elapsed time per iteration (s): 1.74 | learning rate: 2.689E-05 | global batch size: 512 | lm loss: 1.122337E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.916 | TFLOPs: 44.27 | 31: iteration 45950/ 52452 | consumed samples: 23526400 | consumed tokens: 48182067200 | elapsed time per iteration (s): 1.80 | learning rate: 2.687E-05 | global batch size: 512 | lm loss: 1.157740E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.055 | TFLOPs: 42.64 | 31: iteration 45960/ 52452 | consumed samples: 23531520 | consumed tokens: 48192552960 | elapsed time per iteration (s): 1.76 | learning rate: 2.685E-05 | global batch size: 512 | lm loss: 1.157156E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.718 | TFLOPs: 43.79 | 31: iteration 45970/ 52452 | consumed samples: 23536640 | consumed tokens: 48203038720 | elapsed time per iteration (s): 1.78 | learning rate: 2.683E-05 | global batch size: 512 | lm loss: 1.144230E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.442 | TFLOPs: 43.29 | 31: iteration 45980/ 52452 | consumed samples: 23541760 | consumed tokens: 48213524480 | elapsed time per iteration (s): 1.74 | learning rate: 2.681E-05 | global batch size: 512 | lm loss: 1.151730E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.639 | TFLOPs: 44.07 | 31: iteration 45990/ 52452 | consumed samples: 23546880 | consumed tokens: 48224010240 | elapsed time per iteration (s): 1.80 | learning rate: 2.679E-05 | global batch size: 512 | lm loss: 1.150213E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.793 | TFLOPs: 42.75 | 0: [2022-12-26 17:19:49,188] [INFO] [logging.py:68:log_dist] [Rank 0] step=46000, skipped=0, lr=[2.6770103898180908e-05, 2.6770103898180908e-05, 2.6770103898180908e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 46000/ 52452 | consumed samples: 23552000 | consumed tokens: 48234496000 | elapsed time per iteration (s): 1.79 | learning rate: 2.677E-05 | global batch size: 512 | lm loss: 1.158686E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.459 | TFLOPs: 42.85 | 0: steps: 46000 loss: 1.1900 iter time (s): 1.795 samples/sec: 285.164 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46000 | lm loss value: 4.226581E+00 | lm loss PPL: 6.848270E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46000 | lm loss value: 4.076621E+00 | lm loss PPL: 5.894596E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46010/ 52452 | consumed samples: 23557120 | consumed tokens: 48244981760 | elapsed time per iteration (s): 9.13 | learning rate: 2.675E-05 | global batch size: 512 | lm loss: 1.143544E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.085 | TFLOPs: 8.42 | 31: iteration 46020/ 52452 | consumed samples: 23562240 | consumed tokens: 48255467520 | elapsed time per iteration (s): 1.75 | learning rate: 2.673E-05 | global batch size: 512 | lm loss: 1.152997E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.839 | TFLOPs: 43.80 | 31: iteration 46030/ 52452 | consumed samples: 23567360 | consumed tokens: 48265953280 | elapsed time per iteration (s): 1.77 | learning rate: 2.671E-05 | global batch size: 512 | lm loss: 1.164315E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.827 | TFLOPs: 43.35 | 31: iteration 46040/ 52452 | consumed samples: 23572480 | consumed tokens: 48276439040 | elapsed time per iteration (s): 1.77 | learning rate: 2.669E-05 | global batch size: 512 | lm loss: 1.153135E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.075 | TFLOPs: 43.39 | 31: iteration 46050/ 52452 | consumed samples: 23577600 | consumed tokens: 48286924800 | elapsed time per iteration (s): 1.79 | learning rate: 2.667E-05 | global batch size: 512 | lm loss: 1.157354E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.210 | TFLOPs: 42.96 | 31: iteration 46060/ 52452 | consumed samples: 23582720 | consumed tokens: 48297410560 | elapsed time per iteration (s): 1.79 | learning rate: 2.665E-05 | global batch size: 512 | lm loss: 1.146957E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.631 | TFLOPs: 43.02 | 31: iteration 46070/ 52452 | consumed samples: 23587840 | consumed tokens: 48307896320 | elapsed time per iteration (s): 1.78 | learning rate: 2.663E-05 | global batch size: 512 | lm loss: 1.145889E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.089 | TFLOPs: 43.24 | 31: iteration 46080/ 52452 | consumed samples: 23592960 | consumed tokens: 48318382080 | elapsed time per iteration (s): 1.88 | learning rate: 2.661E-05 | global batch size: 512 | lm loss: 1.161172E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.538 | TFLOPs: 40.91 | 31: iteration 46090/ 52452 | consumed samples: 23598080 | consumed tokens: 48328867840 | elapsed time per iteration (s): 1.79 | learning rate: 2.658E-05 | global batch size: 512 | lm loss: 1.156440E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.539 | TFLOPs: 43.01 | 31: iteration 46100/ 52452 | consumed samples: 23603200 | consumed tokens: 48339353600 | elapsed time per iteration (s): 1.85 | learning rate: 2.656E-05 | global batch size: 512 | lm loss: 1.145206E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.075 | TFLOPs: 41.59 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46100 | lm loss value: 4.228792E+00 | lm loss PPL: 6.863428E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46100 | lm loss value: 4.081688E+00 | lm loss PPL: 5.924542E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46110/ 52452 | consumed samples: 23608320 | consumed tokens: 48349839360 | elapsed time per iteration (s): 9.12 | learning rate: 2.654E-05 | global batch size: 512 | lm loss: 1.149294E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.151 | TFLOPs: 8.43 | 31: iteration 46120/ 52452 | consumed samples: 23613440 | consumed tokens: 48360325120 | elapsed time per iteration (s): 1.75 | learning rate: 2.652E-05 | global batch size: 512 | lm loss: 1.147449E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.369 | TFLOPs: 44.03 | 31: iteration 46130/ 52452 | consumed samples: 23618560 | consumed tokens: 48370810880 | elapsed time per iteration (s): 1.83 | learning rate: 2.650E-05 | global batch size: 512 | lm loss: 1.163198E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.278 | TFLOPs: 42.07 | 31: iteration 46140/ 52452 | consumed samples: 23623680 | consumed tokens: 48381296640 | elapsed time per iteration (s): 1.72 | learning rate: 2.648E-05 | global batch size: 512 | lm loss: 1.157464E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.396 | TFLOPs: 44.64 | 31: iteration 46150/ 52452 | consumed samples: 23628800 | consumed tokens: 48391782400 | elapsed time per iteration (s): 1.85 | learning rate: 2.646E-05 | global batch size: 512 | lm loss: 1.159941E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.777 | TFLOPs: 41.54 | 31: iteration 46160/ 52452 | consumed samples: 23633920 | consumed tokens: 48402268160 | elapsed time per iteration (s): 1.80 | learning rate: 2.644E-05 | global batch size: 512 | lm loss: 1.160001E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.921 | TFLOPs: 42.77 | 31: iteration 46170/ 52452 | consumed samples: 23639040 | consumed tokens: 48412753920 | elapsed time per iteration (s): 1.74 | learning rate: 2.642E-05 | global batch size: 512 | lm loss: 1.162682E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.072 | TFLOPs: 44.29 | 31: iteration 46180/ 52452 | consumed samples: 23644160 | consumed tokens: 48423239680 | elapsed time per iteration (s): 1.77 | learning rate: 2.640E-05 | global batch size: 512 | lm loss: 1.157850E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.960 | TFLOPs: 43.52 | 31: iteration 46190/ 52452 | consumed samples: 23649280 | consumed tokens: 48433725440 | elapsed time per iteration (s): 1.75 | learning rate: 2.638E-05 | global batch size: 512 | lm loss: 1.139668E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.710 | TFLOPs: 43.93 | 31: iteration 46200/ 52452 | consumed samples: 23654400 | consumed tokens: 48444211200 | elapsed time per iteration (s): 1.76 | learning rate: 2.636E-05 | global batch size: 512 | lm loss: 1.156854E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.576 | TFLOPs: 43.61 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46200 | lm loss value: 4.322865E+00 | lm loss PPL: 7.540432E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46200 | lm loss value: 4.133936E+00 | lm loss PPL: 6.242313E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46210/ 52452 | consumed samples: 23659520 | consumed tokens: 48454696960 | elapsed time per iteration (s): 9.19 | learning rate: 2.634E-05 | global batch size: 512 | lm loss: 1.147378E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.703 | TFLOPs: 8.36 | 31: iteration 46220/ 52452 | consumed samples: 23664640 | consumed tokens: 48465182720 | elapsed time per iteration (s): 1.79 | learning rate: 2.632E-05 | global batch size: 512 | lm loss: 1.163354E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.255 | TFLOPs: 42.97 | 31: iteration 46230/ 52452 | consumed samples: 23669760 | consumed tokens: 48475668480 | elapsed time per iteration (s): 1.72 | learning rate: 2.630E-05 | global batch size: 512 | lm loss: 1.144186E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.037 | TFLOPs: 44.58 | 31: iteration 46240/ 52452 | consumed samples: 23674880 | consumed tokens: 48486154240 | elapsed time per iteration (s): 1.77 | learning rate: 2.628E-05 | global batch size: 512 | lm loss: 1.144150E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.913 | TFLOPs: 43.36 | 31: iteration 46250/ 52452 | consumed samples: 23680000 | consumed tokens: 48496640000 | elapsed time per iteration (s): 1.76 | learning rate: 2.626E-05 | global batch size: 512 | lm loss: 1.152973E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.404 | TFLOPs: 43.59 | 31: iteration 46260/ 52452 | consumed samples: 23685120 | consumed tokens: 48507125760 | elapsed time per iteration (s): 1.76 | learning rate: 2.624E-05 | global batch size: 512 | lm loss: 1.154079E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.819 | TFLOPs: 43.65 | 31: iteration 46270/ 52452 | consumed samples: 23690240 | consumed tokens: 48517611520 | elapsed time per iteration (s): 1.74 | learning rate: 2.622E-05 | global batch size: 512 | lm loss: 1.149644E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.732 | TFLOPs: 44.09 | 31: iteration 46280/ 52452 | consumed samples: 23695360 | consumed tokens: 48528097280 | elapsed time per iteration (s): 1.73 | learning rate: 2.620E-05 | global batch size: 512 | lm loss: 1.154293E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.175 | TFLOPs: 44.30 | 31: iteration 46290/ 52452 | consumed samples: 23700480 | consumed tokens: 48538583040 | elapsed time per iteration (s): 1.74 | learning rate: 2.618E-05 | global batch size: 512 | lm loss: 1.143930E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.044 | TFLOPs: 44.13 | 31: iteration 46300/ 52452 | consumed samples: 23705600 | consumed tokens: 48549068800 | elapsed time per iteration (s): 1.76 | learning rate: 2.616E-05 | global batch size: 512 | lm loss: 1.144170E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.625 | TFLOPs: 43.62 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46300 | lm loss value: 4.249628E+00 | lm loss PPL: 7.007934E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46300 | lm loss value: 4.087184E+00 | lm loss PPL: 5.957190E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46310/ 52452 | consumed samples: 23710720 | consumed tokens: 48559554560 | elapsed time per iteration (s): 9.18 | learning rate: 2.614E-05 | global batch size: 512 | lm loss: 1.167098E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.772 | TFLOPs: 8.37 | 31: iteration 46320/ 52452 | consumed samples: 23715840 | consumed tokens: 48570040320 | elapsed time per iteration (s): 1.78 | learning rate: 2.612E-05 | global batch size: 512 | lm loss: 1.162325E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.660 | TFLOPs: 43.18 | 31: iteration 46330/ 52452 | consumed samples: 23720960 | consumed tokens: 48580526080 | elapsed time per iteration (s): 1.81 | learning rate: 2.610E-05 | global batch size: 512 | lm loss: 1.163070E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.337 | TFLOPs: 42.38 | 31: iteration 46340/ 52452 | consumed samples: 23726080 | consumed tokens: 48591011840 | elapsed time per iteration (s): 1.85 | learning rate: 2.608E-05 | global batch size: 512 | lm loss: 1.160935E+00 | grad norm: 0.149 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.811 | TFLOPs: 41.55 | 31: iteration 46350/ 52452 | consumed samples: 23731200 | consumed tokens: 48601497600 | elapsed time per iteration (s): 1.82 | learning rate: 2.606E-05 | global batch size: 512 | lm loss: 1.155387E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.659 | TFLOPs: 42.28 | 31: iteration 46360/ 52452 | consumed samples: 23736320 | consumed tokens: 48611983360 | elapsed time per iteration (s): 1.80 | learning rate: 2.604E-05 | global batch size: 512 | lm loss: 1.157808E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.883 | TFLOPs: 42.76 | 31: iteration 46370/ 52452 | consumed samples: 23741440 | consumed tokens: 48622469120 | elapsed time per iteration (s): 1.82 | learning rate: 2.602E-05 | global batch size: 512 | lm loss: 1.149713E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.672 | TFLOPs: 42.13 | 31: iteration 46380/ 52452 | consumed samples: 23746560 | consumed tokens: 48632954880 | elapsed time per iteration (s): 1.77 | learning rate: 2.600E-05 | global batch size: 512 | lm loss: 1.147035E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.874 | TFLOPs: 43.36 | 31: iteration 46390/ 52452 | consumed samples: 23751680 | consumed tokens: 48643440640 | elapsed time per iteration (s): 1.77 | learning rate: 2.599E-05 | global batch size: 512 | lm loss: 1.156771E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.915 | TFLOPs: 43.51 | 31: iteration 46400/ 52452 | consumed samples: 23756800 | consumed tokens: 48653926400 | elapsed time per iteration (s): 1.75 | learning rate: 2.597E-05 | global batch size: 512 | lm loss: 1.159450E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.295 | TFLOPs: 43.87 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46400 | lm loss value: 4.216643E+00 | lm loss PPL: 6.780550E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46400 | lm loss value: 4.062834E+00 | lm loss PPL: 5.813886E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46410/ 52452 | consumed samples: 23761920 | consumed tokens: 48664412160 | elapsed time per iteration (s): 9.12 | learning rate: 2.595E-05 | global batch size: 512 | lm loss: 1.153659E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.148 | TFLOPs: 8.43 | 31: iteration 46420/ 52452 | consumed samples: 23767040 | consumed tokens: 48674897920 | elapsed time per iteration (s): 1.76 | learning rate: 2.593E-05 | global batch size: 512 | lm loss: 1.151006E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.048 | TFLOPs: 43.68 | 31: iteration 46430/ 52452 | consumed samples: 23772160 | consumed tokens: 48685383680 | elapsed time per iteration (s): 1.79 | learning rate: 2.591E-05 | global batch size: 512 | lm loss: 1.153484E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.266 | TFLOPs: 42.82 | 31: iteration 46440/ 52452 | consumed samples: 23777280 | consumed tokens: 48695869440 | elapsed time per iteration (s): 1.78 | learning rate: 2.589E-05 | global batch size: 512 | lm loss: 1.146527E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.084 | TFLOPs: 43.24 | 31: iteration 46450/ 52452 | consumed samples: 23782400 | consumed tokens: 48706355200 | elapsed time per iteration (s): 1.77 | learning rate: 2.587E-05 | global batch size: 512 | lm loss: 1.148782E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.769 | TFLOPs: 43.49 | 31: iteration 46460/ 52452 | consumed samples: 23787520 | consumed tokens: 48716840960 | elapsed time per iteration (s): 1.76 | learning rate: 2.585E-05 | global batch size: 512 | lm loss: 1.152745E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.382 | TFLOPs: 43.73 | 31: iteration 46470/ 52452 | consumed samples: 23792640 | consumed tokens: 48727326720 | elapsed time per iteration (s): 1.81 | learning rate: 2.583E-05 | global batch size: 512 | lm loss: 1.150430E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.594 | TFLOPs: 42.57 | 31: iteration 46480/ 52452 | consumed samples: 23797760 | consumed tokens: 48737812480 | elapsed time per iteration (s): 1.81 | learning rate: 2.581E-05 | global batch size: 512 | lm loss: 1.148334E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.907 | TFLOPs: 42.46 | 31: iteration 46490/ 52452 | consumed samples: 23802880 | consumed tokens: 48748298240 | elapsed time per iteration (s): 1.77 | learning rate: 2.579E-05 | global batch size: 512 | lm loss: 1.149981E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.748 | TFLOPs: 43.34 | 31: iteration 46500/ 52452 | consumed samples: 23808000 | consumed tokens: 48758784000 | elapsed time per iteration (s): 1.76 | learning rate: 2.577E-05 | global batch size: 512 | lm loss: 1.143394E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.470 | TFLOPs: 43.75 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46500 | lm loss value: 4.234266E+00 | lm loss PPL: 6.901099E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46500 | lm loss value: 4.077162E+00 | lm loss PPL: 5.897787E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46510/ 52452 | consumed samples: 23813120 | consumed tokens: 48769269760 | elapsed time per iteration (s): 9.24 | learning rate: 2.575E-05 | global batch size: 512 | lm loss: 1.151763E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.402 | TFLOPs: 8.32 | 31: iteration 46520/ 52452 | consumed samples: 23818240 | consumed tokens: 48779755520 | elapsed time per iteration (s): 1.88 | learning rate: 2.573E-05 | global batch size: 512 | lm loss: 1.150770E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.624 | TFLOPs: 40.77 | 31: iteration 46530/ 52452 | consumed samples: 23823360 | consumed tokens: 48790241280 | elapsed time per iteration (s): 1.91 | learning rate: 2.571E-05 | global batch size: 512 | lm loss: 1.155005E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 268.685 | TFLOPs: 40.33 | 31: iteration 46540/ 52452 | consumed samples: 23828480 | consumed tokens: 48800727040 | elapsed time per iteration (s): 1.77 | learning rate: 2.570E-05 | global batch size: 512 | lm loss: 1.154411E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.751 | TFLOPs: 43.49 | 31: iteration 46550/ 52452 | consumed samples: 23833600 | consumed tokens: 48811212800 | elapsed time per iteration (s): 1.79 | learning rate: 2.568E-05 | global batch size: 512 | lm loss: 1.144975E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.094 | TFLOPs: 42.94 | 31: iteration 46560/ 52452 | consumed samples: 23838720 | consumed tokens: 48821698560 | elapsed time per iteration (s): 1.85 | learning rate: 2.566E-05 | global batch size: 512 | lm loss: 1.154396E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.413 | TFLOPs: 41.64 | 31: iteration 46570/ 52452 | consumed samples: 23843840 | consumed tokens: 48832184320 | elapsed time per iteration (s): 1.79 | learning rate: 2.564E-05 | global batch size: 512 | lm loss: 1.145812E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.635 | TFLOPs: 42.87 | 31: iteration 46580/ 52452 | consumed samples: 23848960 | consumed tokens: 48842670080 | elapsed time per iteration (s): 1.76 | learning rate: 2.562E-05 | global batch size: 512 | lm loss: 1.143604E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.322 | TFLOPs: 43.58 | 31: iteration 46590/ 52452 | consumed samples: 23854080 | consumed tokens: 48853155840 | elapsed time per iteration (s): 1.88 | learning rate: 2.560E-05 | global batch size: 512 | lm loss: 1.155436E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.602 | TFLOPs: 40.92 | 31: iteration 46600/ 52452 | consumed samples: 23859200 | consumed tokens: 48863641600 | elapsed time per iteration (s): 1.88 | learning rate: 2.558E-05 | global batch size: 512 | lm loss: 1.157111E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 271.642 | TFLOPs: 40.77 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46600 | lm loss value: 4.238675E+00 | lm loss PPL: 6.931592E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46600 | lm loss value: 4.082715E+00 | lm loss PPL: 5.930624E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46610/ 52452 | consumed samples: 23864320 | consumed tokens: 48874127360 | elapsed time per iteration (s): 9.23 | learning rate: 2.556E-05 | global batch size: 512 | lm loss: 1.152951E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.462 | TFLOPs: 8.32 | 31: iteration 46620/ 52452 | consumed samples: 23869440 | consumed tokens: 48884613120 | elapsed time per iteration (s): 1.74 | learning rate: 2.554E-05 | global batch size: 512 | lm loss: 1.143304E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.622 | TFLOPs: 44.07 | 31: iteration 46630/ 52452 | consumed samples: 23874560 | consumed tokens: 48895098880 | elapsed time per iteration (s): 1.84 | learning rate: 2.553E-05 | global batch size: 512 | lm loss: 1.133116E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.587 | TFLOPs: 41.81 | 31: iteration 46640/ 52452 | consumed samples: 23879680 | consumed tokens: 48905584640 | elapsed time per iteration (s): 1.77 | learning rate: 2.551E-05 | global batch size: 512 | lm loss: 1.155591E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.851 | TFLOPs: 43.35 | 31: iteration 46650/ 52452 | consumed samples: 23884800 | consumed tokens: 48916070400 | elapsed time per iteration (s): 1.76 | learning rate: 2.549E-05 | global batch size: 512 | lm loss: 1.154708E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.655 | TFLOPs: 43.63 | 31: iteration 46660/ 52452 | consumed samples: 23889920 | consumed tokens: 48926556160 | elapsed time per iteration (s): 1.80 | learning rate: 2.547E-05 | global batch size: 512 | lm loss: 1.157278E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.895 | TFLOPs: 42.76 | 31: iteration 46670/ 52452 | consumed samples: 23895040 | consumed tokens: 48937041920 | elapsed time per iteration (s): 1.90 | learning rate: 2.545E-05 | global batch size: 512 | lm loss: 1.153091E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.742 | TFLOPs: 40.49 | 31: iteration 46680/ 52452 | consumed samples: 23900160 | consumed tokens: 48947527680 | elapsed time per iteration (s): 1.74 | learning rate: 2.543E-05 | global batch size: 512 | lm loss: 1.155973E+00 | grad norm: 0.173 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.596 | TFLOPs: 44.22 | 31: iteration 46690/ 52452 | consumed samples: 23905280 | consumed tokens: 48958013440 | elapsed time per iteration (s): 1.84 | learning rate: 2.541E-05 | global batch size: 512 | lm loss: 1.157346E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.000 | TFLOPs: 41.73 | 31: iteration 46700/ 52452 | consumed samples: 23910400 | consumed tokens: 48968499200 | elapsed time per iteration (s): 1.80 | learning rate: 2.539E-05 | global batch size: 512 | lm loss: 1.159538E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.968 | TFLOPs: 42.62 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46700 | lm loss value: 4.204564E+00 | lm loss PPL: 6.699136E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46700 | lm loss value: 4.051750E+00 | lm loss PPL: 5.749800E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46710/ 52452 | consumed samples: 23915520 | consumed tokens: 48978984960 | elapsed time per iteration (s): 9.20 | learning rate: 2.538E-05 | global batch size: 512 | lm loss: 1.150274E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.675 | TFLOPs: 8.36 | 31: iteration 46720/ 52452 | consumed samples: 23920640 | consumed tokens: 48989470720 | elapsed time per iteration (s): 1.76 | learning rate: 2.536E-05 | global batch size: 512 | lm loss: 1.157093E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.757 | TFLOPs: 43.64 | 31: iteration 46730/ 52452 | consumed samples: 23925760 | consumed tokens: 48999956480 | elapsed time per iteration (s): 1.86 | learning rate: 2.534E-05 | global batch size: 512 | lm loss: 1.144199E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.416 | TFLOPs: 41.34 | 31: iteration 46740/ 52452 | consumed samples: 23930880 | consumed tokens: 49010442240 | elapsed time per iteration (s): 1.84 | learning rate: 2.532E-05 | global batch size: 512 | lm loss: 1.142123E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.579 | TFLOPs: 41.81 | 31: iteration 46750/ 52452 | consumed samples: 23936000 | consumed tokens: 49020928000 | elapsed time per iteration (s): 1.77 | learning rate: 2.530E-05 | global batch size: 512 | lm loss: 1.144575E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.671 | TFLOPs: 43.48 | 31: iteration 46760/ 52452 | consumed samples: 23941120 | consumed tokens: 49031413760 | elapsed time per iteration (s): 1.77 | learning rate: 2.528E-05 | global batch size: 512 | lm loss: 1.148306E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.650 | TFLOPs: 43.47 | 31: iteration 46770/ 52452 | consumed samples: 23946240 | consumed tokens: 49041899520 | elapsed time per iteration (s): 1.79 | learning rate: 2.527E-05 | global batch size: 512 | lm loss: 1.140260E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.396 | TFLOPs: 42.84 | 31: iteration 46780/ 52452 | consumed samples: 23951360 | consumed tokens: 49052385280 | elapsed time per iteration (s): 1.75 | learning rate: 2.525E-05 | global batch size: 512 | lm loss: 1.157373E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.907 | TFLOPs: 43.81 | 31: iteration 46790/ 52452 | consumed samples: 23956480 | consumed tokens: 49062871040 | elapsed time per iteration (s): 1.87 | learning rate: 2.523E-05 | global batch size: 512 | lm loss: 1.149014E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.422 | TFLOPs: 41.19 | 31: iteration 46800/ 52452 | consumed samples: 23961600 | consumed tokens: 49073356800 | elapsed time per iteration (s): 1.85 | learning rate: 2.521E-05 | global batch size: 512 | lm loss: 1.157061E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.026 | TFLOPs: 41.58 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46800 | lm loss value: 4.239625E+00 | lm loss PPL: 6.938186E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46800 | lm loss value: 4.088151E+00 | lm loss PPL: 5.962953E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46810/ 52452 | consumed samples: 23966720 | consumed tokens: 49083842560 | elapsed time per iteration (s): 9.15 | learning rate: 2.519E-05 | global batch size: 512 | lm loss: 1.138901E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.951 | TFLOPs: 8.40 | 31: iteration 46820/ 52452 | consumed samples: 23971840 | consumed tokens: 49094328320 | elapsed time per iteration (s): 1.78 | learning rate: 2.517E-05 | global batch size: 512 | lm loss: 1.164478E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.432 | TFLOPs: 43.14 | 31: iteration 46830/ 52452 | consumed samples: 23976960 | consumed tokens: 49104814080 | elapsed time per iteration (s): 1.79 | learning rate: 2.516E-05 | global batch size: 512 | lm loss: 1.143957E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.379 | TFLOPs: 42.83 | 31: iteration 46840/ 52452 | consumed samples: 23982080 | consumed tokens: 49115299840 | elapsed time per iteration (s): 1.76 | learning rate: 2.514E-05 | global batch size: 512 | lm loss: 1.141931E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.025 | TFLOPs: 43.68 | 31: iteration 46850/ 52452 | consumed samples: 23987200 | consumed tokens: 49125785600 | elapsed time per iteration (s): 1.82 | learning rate: 2.512E-05 | global batch size: 512 | lm loss: 1.153776E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.133 | TFLOPs: 42.20 | 31: iteration 46860/ 52452 | consumed samples: 23992320 | consumed tokens: 49136271360 | elapsed time per iteration (s): 1.78 | learning rate: 2.510E-05 | global batch size: 512 | lm loss: 1.146503E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.463 | TFLOPs: 43.15 | 31: iteration 46870/ 52452 | consumed samples: 23997440 | consumed tokens: 49146757120 | elapsed time per iteration (s): 1.79 | learning rate: 2.508E-05 | global batch size: 512 | lm loss: 1.142688E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.298 | TFLOPs: 42.97 | 31: iteration 46880/ 52452 | consumed samples: 24002560 | consumed tokens: 49157242880 | elapsed time per iteration (s): 1.87 | learning rate: 2.507E-05 | global batch size: 512 | lm loss: 1.149389E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.434 | TFLOPs: 41.19 | 31: iteration 46890/ 52452 | consumed samples: 24007680 | consumed tokens: 49167728640 | elapsed time per iteration (s): 1.77 | learning rate: 2.505E-05 | global batch size: 512 | lm loss: 1.148594E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.810 | TFLOPs: 43.50 | 31: iteration 46900/ 52452 | consumed samples: 24012800 | consumed tokens: 49178214400 | elapsed time per iteration (s): 1.83 | learning rate: 2.503E-05 | global batch size: 512 | lm loss: 1.141878E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.053 | TFLOPs: 41.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 46900 | lm loss value: 4.277195E+00 | lm loss PPL: 7.203809E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 46900 | lm loss value: 4.109196E+00 | lm loss PPL: 6.089775E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 46910/ 52452 | consumed samples: 24017920 | consumed tokens: 49188700160 | elapsed time per iteration (s): 9.14 | learning rate: 2.501E-05 | global batch size: 512 | lm loss: 1.148437E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.020 | TFLOPs: 8.41 | 31: iteration 46920/ 52452 | consumed samples: 24023040 | consumed tokens: 49199185920 | elapsed time per iteration (s): 1.76 | learning rate: 2.499E-05 | global batch size: 512 | lm loss: 1.148894E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.406 | TFLOPs: 43.74 | 31: iteration 46930/ 52452 | consumed samples: 24028160 | consumed tokens: 49209671680 | elapsed time per iteration (s): 1.81 | learning rate: 2.498E-05 | global batch size: 512 | lm loss: 1.143891E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.603 | TFLOPs: 42.42 | 31: iteration 46940/ 52452 | consumed samples: 24033280 | consumed tokens: 49220157440 | elapsed time per iteration (s): 1.81 | learning rate: 2.496E-05 | global batch size: 512 | lm loss: 1.150074E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.121 | TFLOPs: 42.34 | 31: iteration 46950/ 52452 | consumed samples: 24038400 | consumed tokens: 49230643200 | elapsed time per iteration (s): 1.78 | learning rate: 2.494E-05 | global batch size: 512 | lm loss: 1.143716E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.911 | TFLOPs: 43.06 | 31: iteration 46960/ 52452 | consumed samples: 24043520 | consumed tokens: 49241128960 | elapsed time per iteration (s): 1.74 | learning rate: 2.492E-05 | global batch size: 512 | lm loss: 1.156064E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.012 | TFLOPs: 44.28 | 31: iteration 46970/ 52452 | consumed samples: 24048640 | consumed tokens: 49251614720 | elapsed time per iteration (s): 1.74 | learning rate: 2.490E-05 | global batch size: 512 | lm loss: 1.141208E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.743 | TFLOPs: 44.24 | 31: iteration 46980/ 52452 | consumed samples: 24053760 | consumed tokens: 49262100480 | elapsed time per iteration (s): 1.75 | learning rate: 2.489E-05 | global batch size: 512 | lm loss: 1.146279E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.738 | TFLOPs: 43.94 | 31: iteration 46990/ 52452 | consumed samples: 24058880 | consumed tokens: 49272586240 | elapsed time per iteration (s): 1.75 | learning rate: 2.487E-05 | global batch size: 512 | lm loss: 1.141922E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.224 | TFLOPs: 43.86 | 31: iteration 47000/ 52452 | consumed samples: 24064000 | consumed tokens: 49283072000 | elapsed time per iteration (s): 1.78 | learning rate: 2.485E-05 | global batch size: 512 | lm loss: 1.134651E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.065 | TFLOPs: 43.09 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47000 | lm loss value: 4.260774E+00 | lm loss PPL: 7.086482E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47000 | lm loss value: 4.100439E+00 | lm loss PPL: 6.036676E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47010/ 52452 | consumed samples: 24069120 | consumed tokens: 49293557760 | elapsed time per iteration (s): 9.20 | learning rate: 2.483E-05 | global batch size: 512 | lm loss: 1.150941E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.640 | TFLOPs: 8.35 | 31: iteration 47020/ 52452 | consumed samples: 24074240 | consumed tokens: 49304043520 | elapsed time per iteration (s): 1.74 | learning rate: 2.482E-05 | global batch size: 512 | lm loss: 1.147196E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.875 | TFLOPs: 44.26 | 31: iteration 47030/ 52452 | consumed samples: 24079360 | consumed tokens: 49314529280 | elapsed time per iteration (s): 1.75 | learning rate: 2.480E-05 | global batch size: 512 | lm loss: 1.145858E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.511 | TFLOPs: 43.90 | 31: iteration 47040/ 52452 | consumed samples: 24084480 | consumed tokens: 49325015040 | elapsed time per iteration (s): 1.79 | learning rate: 2.478E-05 | global batch size: 512 | lm loss: 1.159121E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.006 | TFLOPs: 42.93 | 31: iteration 47050/ 52452 | consumed samples: 24089600 | consumed tokens: 49335500800 | elapsed time per iteration (s): 2.03 | learning rate: 2.476E-05 | global batch size: 512 | lm loss: 1.158120E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 252.286 | TFLOPs: 37.87 | 31: iteration 47060/ 52452 | consumed samples: 24094720 | consumed tokens: 49345986560 | elapsed time per iteration (s): 1.81 | learning rate: 2.475E-05 | global batch size: 512 | lm loss: 1.144277E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.216 | TFLOPs: 42.51 | 31: iteration 47070/ 52452 | consumed samples: 24099840 | consumed tokens: 49356472320 | elapsed time per iteration (s): 1.76 | learning rate: 2.473E-05 | global batch size: 512 | lm loss: 1.158276E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.312 | TFLOPs: 43.57 | 31: iteration 47080/ 52452 | consumed samples: 24104960 | consumed tokens: 49366958080 | elapsed time per iteration (s): 1.85 | learning rate: 2.471E-05 | global batch size: 512 | lm loss: 1.149920E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.997 | TFLOPs: 41.58 | 31: iteration 47090/ 52452 | consumed samples: 24110080 | consumed tokens: 49377443840 | elapsed time per iteration (s): 1.78 | learning rate: 2.469E-05 | global batch size: 512 | lm loss: 1.149501E+00 | grad norm: 0.176 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.025 | TFLOPs: 43.08 | 31: iteration 47100/ 52452 | consumed samples: 24115200 | consumed tokens: 49387929600 | elapsed time per iteration (s): 1.75 | learning rate: 2.468E-05 | global batch size: 512 | lm loss: 1.159834E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.866 | TFLOPs: 43.81 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47100 | lm loss value: 4.275230E+00 | lm loss PPL: 7.189667E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47100 | lm loss value: 4.102158E+00 | lm loss PPL: 6.047062E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47110/ 52452 | consumed samples: 24120320 | consumed tokens: 49398415360 | elapsed time per iteration (s): 9.23 | learning rate: 2.466E-05 | global batch size: 512 | lm loss: 1.147146E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.473 | TFLOPs: 8.33 | 31: iteration 47120/ 52452 | consumed samples: 24125440 | consumed tokens: 49408901120 | elapsed time per iteration (s): 1.78 | learning rate: 2.464E-05 | global batch size: 512 | lm loss: 1.144034E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.548 | TFLOPs: 43.16 | 31: iteration 47130/ 52452 | consumed samples: 24130560 | consumed tokens: 49419386880 | elapsed time per iteration (s): 1.78 | learning rate: 2.463E-05 | global batch size: 512 | lm loss: 1.155322E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.728 | TFLOPs: 43.19 | 31: iteration 47140/ 52452 | consumed samples: 24135680 | consumed tokens: 49429872640 | elapsed time per iteration (s): 1.75 | learning rate: 2.461E-05 | global batch size: 512 | lm loss: 1.167295E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.025 | TFLOPs: 43.98 | 31: iteration 47150/ 52452 | consumed samples: 24140800 | consumed tokens: 49440358400 | elapsed time per iteration (s): 1.82 | learning rate: 2.459E-05 | global batch size: 512 | lm loss: 1.154457E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.736 | TFLOPs: 42.14 | 31: iteration 47160/ 52452 | consumed samples: 24145920 | consumed tokens: 49450844160 | elapsed time per iteration (s): 1.74 | learning rate: 2.457E-05 | global batch size: 512 | lm loss: 1.149019E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.548 | TFLOPs: 44.21 | 31: iteration 47170/ 52452 | consumed samples: 24151040 | consumed tokens: 49461329920 | elapsed time per iteration (s): 1.77 | learning rate: 2.456E-05 | global batch size: 512 | lm loss: 1.141284E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.951 | TFLOPs: 43.52 | 31: iteration 47180/ 52452 | consumed samples: 24156160 | consumed tokens: 49471815680 | elapsed time per iteration (s): 1.82 | learning rate: 2.454E-05 | global batch size: 512 | lm loss: 1.150633E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.896 | TFLOPs: 42.31 | 31: iteration 47190/ 52452 | consumed samples: 24161280 | consumed tokens: 49482301440 | elapsed time per iteration (s): 1.81 | learning rate: 2.452E-05 | global batch size: 512 | lm loss: 1.145268E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.564 | TFLOPs: 42.41 | 31: iteration 47200/ 52452 | consumed samples: 24166400 | consumed tokens: 49492787200 | elapsed time per iteration (s): 1.79 | learning rate: 2.451E-05 | global batch size: 512 | lm loss: 1.157562E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.688 | TFLOPs: 42.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47200 | lm loss value: 4.260273E+00 | lm loss PPL: 7.082932E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47200 | lm loss value: 4.099752E+00 | lm loss PPL: 6.032532E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47210/ 52452 | consumed samples: 24171520 | consumed tokens: 49503272960 | elapsed time per iteration (s): 9.13 | learning rate: 2.449E-05 | global batch size: 512 | lm loss: 1.138966E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.072 | TFLOPs: 8.42 | 31: iteration 47220/ 52452 | consumed samples: 24176640 | consumed tokens: 49513758720 | elapsed time per iteration (s): 1.94 | learning rate: 2.447E-05 | global batch size: 512 | lm loss: 1.160565E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 264.456 | TFLOPs: 39.69 | 31: iteration 47230/ 52452 | consumed samples: 24181760 | consumed tokens: 49524244480 | elapsed time per iteration (s): 1.75 | learning rate: 2.445E-05 | global batch size: 512 | lm loss: 1.141016E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.412 | TFLOPs: 43.89 | 31: iteration 47240/ 52452 | consumed samples: 24186880 | consumed tokens: 49534730240 | elapsed time per iteration (s): 1.76 | learning rate: 2.444E-05 | global batch size: 512 | lm loss: 1.148303E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.511 | TFLOPs: 43.75 | 31: iteration 47250/ 52452 | consumed samples: 24192000 | consumed tokens: 49545216000 | elapsed time per iteration (s): 1.79 | learning rate: 2.442E-05 | global batch size: 512 | lm loss: 1.140747E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.121 | TFLOPs: 42.95 | 31: iteration 47260/ 52452 | consumed samples: 24197120 | consumed tokens: 49555701760 | elapsed time per iteration (s): 1.75 | learning rate: 2.440E-05 | global batch size: 512 | lm loss: 1.150375E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.622 | TFLOPs: 43.92 | 31: iteration 47270/ 52452 | consumed samples: 24202240 | consumed tokens: 49566187520 | elapsed time per iteration (s): 1.81 | learning rate: 2.439E-05 | global batch size: 512 | lm loss: 1.150462E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.474 | TFLOPs: 42.40 | 31: iteration 47280/ 52452 | consumed samples: 24207360 | consumed tokens: 49576673280 | elapsed time per iteration (s): 1.72 | learning rate: 2.437E-05 | global batch size: 512 | lm loss: 1.154138E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.154 | TFLOPs: 44.60 | 31: iteration 47290/ 52452 | consumed samples: 24212480 | consumed tokens: 49587159040 | elapsed time per iteration (s): 1.75 | learning rate: 2.435E-05 | global batch size: 512 | lm loss: 1.148269E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.415 | TFLOPs: 43.89 | 31: iteration 47300/ 52452 | consumed samples: 24217600 | consumed tokens: 49597644800 | elapsed time per iteration (s): 1.75 | learning rate: 2.434E-05 | global batch size: 512 | lm loss: 1.151596E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.930 | TFLOPs: 43.82 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47300 | lm loss value: 4.330033E+00 | lm loss PPL: 7.594678E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47300 | lm loss value: 4.135849E+00 | lm loss PPL: 6.254267E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47310/ 52452 | consumed samples: 24222720 | consumed tokens: 49608130560 | elapsed time per iteration (s): 9.18 | learning rate: 2.432E-05 | global batch size: 512 | lm loss: 1.157645E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.790 | TFLOPs: 8.37 | 31: iteration 47320/ 52452 | consumed samples: 24227840 | consumed tokens: 49618616320 | elapsed time per iteration (s): 1.75 | learning rate: 2.430E-05 | global batch size: 512 | lm loss: 1.139764E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.844 | TFLOPs: 43.95 | 31: iteration 47330/ 52452 | consumed samples: 24232960 | consumed tokens: 49629102080 | elapsed time per iteration (s): 1.76 | learning rate: 2.429E-05 | global batch size: 512 | lm loss: 1.143633E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.886 | TFLOPs: 43.66 | 31: iteration 47340/ 52452 | consumed samples: 24238080 | consumed tokens: 49639587840 | elapsed time per iteration (s): 1.76 | learning rate: 2.427E-05 | global batch size: 512 | lm loss: 1.146642E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.064 | TFLOPs: 43.69 | 31: iteration 47350/ 52452 | consumed samples: 24243200 | consumed tokens: 49650073600 | elapsed time per iteration (s): 1.79 | learning rate: 2.425E-05 | global batch size: 512 | lm loss: 1.139012E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.772 | TFLOPs: 42.89 | 31: iteration 47360/ 52452 | consumed samples: 24248320 | consumed tokens: 49660559360 | elapsed time per iteration (s): 1.73 | learning rate: 2.424E-05 | global batch size: 512 | lm loss: 1.140861E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.854 | TFLOPs: 44.41 | 31: iteration 47370/ 52452 | consumed samples: 24253440 | consumed tokens: 49671045120 | elapsed time per iteration (s): 1.75 | learning rate: 2.422E-05 | global batch size: 512 | lm loss: 1.141468E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.709 | TFLOPs: 43.93 | 31: iteration 47380/ 52452 | consumed samples: 24258560 | consumed tokens: 49681530880 | elapsed time per iteration (s): 1.76 | learning rate: 2.420E-05 | global batch size: 512 | lm loss: 1.147592E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.789 | TFLOPs: 43.65 | 31: iteration 47390/ 52452 | consumed samples: 24263680 | consumed tokens: 49692016640 | elapsed time per iteration (s): 1.78 | learning rate: 2.419E-05 | global batch size: 512 | lm loss: 1.154859E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.981 | TFLOPs: 43.22 | 31: iteration 47400/ 52452 | consumed samples: 24268800 | consumed tokens: 49702502400 | elapsed time per iteration (s): 1.74 | learning rate: 2.417E-05 | global batch size: 512 | lm loss: 1.131529E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.872 | TFLOPs: 44.11 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47400 | lm loss value: 4.301600E+00 | lm loss PPL: 7.381784E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47400 | lm loss value: 4.137629E+00 | lm loss PPL: 6.265406E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47410/ 52452 | consumed samples: 24273920 | consumed tokens: 49712988160 | elapsed time per iteration (s): 9.10 | learning rate: 2.415E-05 | global batch size: 512 | lm loss: 1.145138E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.286 | TFLOPs: 8.45 | 31: iteration 47420/ 52452 | consumed samples: 24279040 | consumed tokens: 49723473920 | elapsed time per iteration (s): 1.74 | learning rate: 2.414E-05 | global batch size: 512 | lm loss: 1.142743E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.533 | TFLOPs: 44.06 | 31: iteration 47430/ 52452 | consumed samples: 24284160 | consumed tokens: 49733959680 | elapsed time per iteration (s): 1.81 | learning rate: 2.412E-05 | global batch size: 512 | lm loss: 1.150991E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.989 | TFLOPs: 42.48 | 31: iteration 47440/ 52452 | consumed samples: 24289280 | consumed tokens: 49744445440 | elapsed time per iteration (s): 1.82 | learning rate: 2.411E-05 | global batch size: 512 | lm loss: 1.150588E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.932 | TFLOPs: 42.32 | 31: iteration 47450/ 52452 | consumed samples: 24294400 | consumed tokens: 49754931200 | elapsed time per iteration (s): 1.81 | learning rate: 2.409E-05 | global batch size: 512 | lm loss: 1.144238E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.302 | TFLOPs: 42.52 | 31: iteration 47460/ 52452 | consumed samples: 24299520 | consumed tokens: 49765416960 | elapsed time per iteration (s): 1.85 | learning rate: 2.407E-05 | global batch size: 512 | lm loss: 1.153927E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.969 | TFLOPs: 41.57 | 31: iteration 47470/ 52452 | consumed samples: 24304640 | consumed tokens: 49775902720 | elapsed time per iteration (s): 1.81 | learning rate: 2.406E-05 | global batch size: 512 | lm loss: 1.146755E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.196 | TFLOPs: 42.51 | 31: iteration 47480/ 52452 | consumed samples: 24309760 | consumed tokens: 49786388480 | elapsed time per iteration (s): 1.78 | learning rate: 2.404E-05 | global batch size: 512 | lm loss: 1.148732E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.557 | TFLOPs: 43.16 | 31: iteration 47490/ 52452 | consumed samples: 24314880 | consumed tokens: 49796874240 | elapsed time per iteration (s): 1.78 | learning rate: 2.403E-05 | global batch size: 512 | lm loss: 1.145557E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.722 | TFLOPs: 43.19 | 31: iteration 47500/ 52452 | consumed samples: 24320000 | consumed tokens: 49807360000 | elapsed time per iteration (s): 1.76 | learning rate: 2.401E-05 | global batch size: 512 | lm loss: 1.144390E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.610 | TFLOPs: 43.77 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47500 | lm loss value: 4.317968E+00 | lm loss PPL: 7.503603E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47500 | lm loss value: 4.141616E+00 | lm loss PPL: 6.290441E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47510/ 52452 | consumed samples: 24325120 | consumed tokens: 49817845760 | elapsed time per iteration (s): 9.12 | learning rate: 2.399E-05 | global batch size: 512 | lm loss: 1.156425E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.114 | TFLOPs: 8.42 | 31: iteration 47520/ 52452 | consumed samples: 24330240 | consumed tokens: 49828331520 | elapsed time per iteration (s): 1.76 | learning rate: 2.398E-05 | global batch size: 512 | lm loss: 1.149828E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.655 | TFLOPs: 43.63 | 31: iteration 47530/ 52452 | consumed samples: 24335360 | consumed tokens: 49838817280 | elapsed time per iteration (s): 1.85 | learning rate: 2.396E-05 | global batch size: 512 | lm loss: 1.153311E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.017 | TFLOPs: 41.43 | 31: iteration 47540/ 52452 | consumed samples: 24340480 | consumed tokens: 49849303040 | elapsed time per iteration (s): 1.77 | learning rate: 2.395E-05 | global batch size: 512 | lm loss: 1.156626E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.058 | TFLOPs: 43.39 | 31: iteration 47550/ 52452 | consumed samples: 24345600 | consumed tokens: 49859788800 | elapsed time per iteration (s): 1.79 | learning rate: 2.393E-05 | global batch size: 512 | lm loss: 1.155077E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.016 | TFLOPs: 42.93 | 31: iteration 47560/ 52452 | consumed samples: 24350720 | consumed tokens: 49870274560 | elapsed time per iteration (s): 1.81 | learning rate: 2.391E-05 | global batch size: 512 | lm loss: 1.154493E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.926 | TFLOPs: 42.47 | 31: iteration 47570/ 52452 | consumed samples: 24355840 | consumed tokens: 49880760320 | elapsed time per iteration (s): 1.72 | learning rate: 2.390E-05 | global batch size: 512 | lm loss: 1.143541E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.423 | TFLOPs: 44.64 | 31: iteration 47580/ 52452 | consumed samples: 24360960 | consumed tokens: 49891246080 | elapsed time per iteration (s): 1.77 | learning rate: 2.388E-05 | global batch size: 512 | lm loss: 1.142134E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.921 | TFLOPs: 43.37 | 31: iteration 47590/ 52452 | consumed samples: 24366080 | consumed tokens: 49901731840 | elapsed time per iteration (s): 1.81 | learning rate: 2.387E-05 | global batch size: 512 | lm loss: 1.149365E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.073 | TFLOPs: 42.49 | 31: iteration 47600/ 52452 | consumed samples: 24371200 | consumed tokens: 49912217600 | elapsed time per iteration (s): 1.79 | learning rate: 2.385E-05 | global batch size: 512 | lm loss: 1.144184E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.797 | TFLOPs: 42.90 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47600 | lm loss value: 4.234516E+00 | lm loss PPL: 6.902827E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47600 | lm loss value: 4.066887E+00 | lm loss PPL: 5.837495E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47610/ 52452 | consumed samples: 24376320 | consumed tokens: 49922703360 | elapsed time per iteration (s): 9.26 | learning rate: 2.383E-05 | global batch size: 512 | lm loss: 1.137890E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.283 | TFLOPs: 8.30 | 31: iteration 47620/ 52452 | consumed samples: 24381440 | consumed tokens: 49933189120 | elapsed time per iteration (s): 1.77 | learning rate: 2.382E-05 | global batch size: 512 | lm loss: 1.141436E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.484 | TFLOPs: 43.45 | 31: iteration 47630/ 52452 | consumed samples: 24386560 | consumed tokens: 49943674880 | elapsed time per iteration (s): 1.93 | learning rate: 2.380E-05 | global batch size: 512 | lm loss: 1.136411E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 265.009 | TFLOPs: 39.78 | 31: iteration 47640/ 52452 | consumed samples: 24391680 | consumed tokens: 49954160640 | elapsed time per iteration (s): 1.81 | learning rate: 2.379E-05 | global batch size: 512 | lm loss: 1.149322E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.499 | TFLOPs: 42.55 | 31: iteration 47650/ 52452 | consumed samples: 24396800 | consumed tokens: 49964646400 | elapsed time per iteration (s): 2.08 | learning rate: 2.377E-05 | global batch size: 512 | lm loss: 1.140104E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 245.960 | TFLOPs: 36.92 | 31: iteration 47660/ 52452 | consumed samples: 24401920 | consumed tokens: 49975132160 | elapsed time per iteration (s): 1.84 | learning rate: 2.376E-05 | global batch size: 512 | lm loss: 1.149047E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.998 | TFLOPs: 41.88 | 31: iteration 47670/ 52452 | consumed samples: 24407040 | consumed tokens: 49985617920 | elapsed time per iteration (s): 1.89 | learning rate: 2.374E-05 | global batch size: 512 | lm loss: 1.151056E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.878 | TFLOPs: 40.66 | 31: iteration 47680/ 52452 | consumed samples: 24412160 | consumed tokens: 49996103680 | elapsed time per iteration (s): 1.78 | learning rate: 2.372E-05 | global batch size: 512 | lm loss: 1.149102E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.206 | TFLOPs: 43.26 | 31: iteration 47690/ 52452 | consumed samples: 24417280 | consumed tokens: 50006589440 | elapsed time per iteration (s): 1.75 | learning rate: 2.371E-05 | global batch size: 512 | lm loss: 1.143262E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.994 | TFLOPs: 43.98 | 31: iteration 47700/ 52452 | consumed samples: 24422400 | consumed tokens: 50017075200 | elapsed time per iteration (s): 1.78 | learning rate: 2.369E-05 | global batch size: 512 | lm loss: 1.133169E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.659 | TFLOPs: 43.18 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47700 | lm loss value: 4.271297E+00 | lm loss PPL: 7.161449E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47700 | lm loss value: 4.096730E+00 | lm loss PPL: 6.014328E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47710/ 52452 | consumed samples: 24427520 | consumed tokens: 50027560960 | elapsed time per iteration (s): 9.17 | learning rate: 2.368E-05 | global batch size: 512 | lm loss: 1.144385E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.823 | TFLOPs: 8.38 | 31: iteration 47720/ 52452 | consumed samples: 24432640 | consumed tokens: 50038046720 | elapsed time per iteration (s): 1.79 | learning rate: 2.366E-05 | global batch size: 512 | lm loss: 1.148798E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.823 | TFLOPs: 42.90 | 31: iteration 47730/ 52452 | consumed samples: 24437760 | consumed tokens: 50048532480 | elapsed time per iteration (s): 1.93 | learning rate: 2.365E-05 | global batch size: 512 | lm loss: 1.139783E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 264.925 | TFLOPs: 39.76 | 31: iteration 47740/ 52452 | consumed samples: 24442880 | consumed tokens: 50059018240 | elapsed time per iteration (s): 1.82 | learning rate: 2.363E-05 | global batch size: 512 | lm loss: 1.140226E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.652 | TFLOPs: 42.12 | 31: iteration 47750/ 52452 | consumed samples: 24448000 | consumed tokens: 50069504000 | elapsed time per iteration (s): 1.77 | learning rate: 2.362E-05 | global batch size: 512 | lm loss: 1.139121E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.938 | TFLOPs: 43.52 | 31: iteration 47760/ 52452 | consumed samples: 24453120 | consumed tokens: 50079989760 | elapsed time per iteration (s): 1.77 | learning rate: 2.360E-05 | global batch size: 512 | lm loss: 1.144821E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.695 | TFLOPs: 43.33 | 31: iteration 47770/ 52452 | consumed samples: 24458240 | consumed tokens: 50090475520 | elapsed time per iteration (s): 1.77 | learning rate: 2.359E-05 | global batch size: 512 | lm loss: 1.147860E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.127 | TFLOPs: 43.40 | 31: iteration 47780/ 52452 | consumed samples: 24463360 | consumed tokens: 50100961280 | elapsed time per iteration (s): 1.78 | learning rate: 2.357E-05 | global batch size: 512 | lm loss: 1.137415E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.572 | TFLOPs: 43.16 | 31: iteration 47790/ 52452 | consumed samples: 24468480 | consumed tokens: 50111447040 | elapsed time per iteration (s): 1.76 | learning rate: 2.356E-05 | global batch size: 512 | lm loss: 1.145635E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.583 | TFLOPs: 43.77 | 31: iteration 47800/ 52452 | consumed samples: 24473600 | consumed tokens: 50121932800 | elapsed time per iteration (s): 1.79 | learning rate: 2.354E-05 | global batch size: 512 | lm loss: 1.141655E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.282 | TFLOPs: 42.82 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47800 | lm loss value: 4.259113E+00 | lm loss PPL: 7.074719E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47800 | lm loss value: 4.091879E+00 | lm loss PPL: 5.985224E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47810/ 52452 | consumed samples: 24478720 | consumed tokens: 50132418560 | elapsed time per iteration (s): 9.18 | learning rate: 2.353E-05 | global batch size: 512 | lm loss: 1.150141E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.773 | TFLOPs: 8.37 | 31: iteration 47820/ 52452 | consumed samples: 24483840 | consumed tokens: 50142904320 | elapsed time per iteration (s): 1.88 | learning rate: 2.351E-05 | global batch size: 512 | lm loss: 1.149592E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.830 | TFLOPs: 40.95 | 31: iteration 47830/ 52452 | consumed samples: 24488960 | consumed tokens: 50153390080 | elapsed time per iteration (s): 1.79 | learning rate: 2.350E-05 | global batch size: 512 | lm loss: 1.142496E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.908 | TFLOPs: 42.91 | 31: iteration 47840/ 52452 | consumed samples: 24494080 | consumed tokens: 50163875840 | elapsed time per iteration (s): 1.78 | learning rate: 2.348E-05 | global batch size: 512 | lm loss: 1.147488E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.061 | TFLOPs: 43.24 | 31: iteration 47850/ 52452 | consumed samples: 24499200 | consumed tokens: 50174361600 | elapsed time per iteration (s): 1.81 | learning rate: 2.347E-05 | global batch size: 512 | lm loss: 1.133323E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.154 | TFLOPs: 42.35 | 31: iteration 47860/ 52452 | consumed samples: 24504320 | consumed tokens: 50184847360 | elapsed time per iteration (s): 1.79 | learning rate: 2.345E-05 | global batch size: 512 | lm loss: 1.135466E+00 | grad norm: 0.151 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.493 | TFLOPs: 42.85 | 31: iteration 47870/ 52452 | consumed samples: 24509440 | consumed tokens: 50195333120 | elapsed time per iteration (s): 1.76 | learning rate: 2.344E-05 | global batch size: 512 | lm loss: 1.146378E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.829 | TFLOPs: 43.65 | 31: iteration 47880/ 52452 | consumed samples: 24514560 | consumed tokens: 50205818880 | elapsed time per iteration (s): 1.84 | learning rate: 2.342E-05 | global batch size: 512 | lm loss: 1.152416E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.355 | TFLOPs: 41.78 | 31: iteration 47890/ 52452 | consumed samples: 24519680 | consumed tokens: 50216304640 | elapsed time per iteration (s): 1.83 | learning rate: 2.341E-05 | global batch size: 512 | lm loss: 1.152385E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.474 | TFLOPs: 41.95 | 31: iteration 47900/ 52452 | consumed samples: 24524800 | consumed tokens: 50226790400 | elapsed time per iteration (s): 1.74 | learning rate: 2.339E-05 | global batch size: 512 | lm loss: 1.148345E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.878 | TFLOPs: 44.26 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 47900 | lm loss value: 4.274110E+00 | lm loss PPL: 7.181618E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 47900 | lm loss value: 4.108943E+00 | lm loss PPL: 6.088233E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 47910/ 52452 | consumed samples: 24529920 | consumed tokens: 50237276160 | elapsed time per iteration (s): 9.21 | learning rate: 2.338E-05 | global batch size: 512 | lm loss: 1.153275E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.597 | TFLOPs: 8.34 | 31: iteration 47920/ 52452 | consumed samples: 24535040 | consumed tokens: 50247761920 | elapsed time per iteration (s): 1.84 | learning rate: 2.336E-05 | global batch size: 512 | lm loss: 1.154943E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.533 | TFLOPs: 41.81 | 31: iteration 47930/ 52452 | consumed samples: 24540160 | consumed tokens: 50258247680 | elapsed time per iteration (s): 1.83 | learning rate: 2.335E-05 | global batch size: 512 | lm loss: 1.152220E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.253 | TFLOPs: 42.06 | 31: iteration 47940/ 52452 | consumed samples: 24545280 | consumed tokens: 50268733440 | elapsed time per iteration (s): 1.81 | learning rate: 2.333E-05 | global batch size: 512 | lm loss: 1.142372E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.509 | TFLOPs: 42.55 | 31: iteration 47950/ 52452 | consumed samples: 24550400 | consumed tokens: 50279219200 | elapsed time per iteration (s): 1.78 | learning rate: 2.332E-05 | global batch size: 512 | lm loss: 1.151217E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.163 | TFLOPs: 43.25 | 31: iteration 47960/ 52452 | consumed samples: 24555520 | consumed tokens: 50289704960 | elapsed time per iteration (s): 1.89 | learning rate: 2.330E-05 | global batch size: 512 | lm loss: 1.132676E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.718 | TFLOPs: 40.63 | 31: iteration 47970/ 52452 | consumed samples: 24560640 | consumed tokens: 50300190720 | elapsed time per iteration (s): 1.82 | learning rate: 2.329E-05 | global batch size: 512 | lm loss: 1.142865E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.997 | TFLOPs: 42.33 | 31: iteration 47980/ 52452 | consumed samples: 24565760 | consumed tokens: 50310676480 | elapsed time per iteration (s): 1.80 | learning rate: 2.327E-05 | global batch size: 512 | lm loss: 1.139639E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.738 | TFLOPs: 42.59 | 31: iteration 47990/ 52452 | consumed samples: 24570880 | consumed tokens: 50321162240 | elapsed time per iteration (s): 1.79 | learning rate: 2.326E-05 | global batch size: 512 | lm loss: 1.141969E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.581 | TFLOPs: 42.86 | 0: [2022-12-26 18:44:13,455] [INFO] [logging.py:68:log_dist] [Rank 0] step=48000, skipped=0, lr=[2.3245013845836666e-05, 2.3245013845836666e-05, 2.3245013845836666e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 48000/ 52452 | consumed samples: 24576000 | consumed tokens: 50331648000 | elapsed time per iteration (s): 1.82 | learning rate: 2.325E-05 | global batch size: 512 | lm loss: 1.144155E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.957 | TFLOPs: 42.17 | 0: steps: 48000 loss: 1.1560 iter time (s): 1.792 samples/sec: 285.657 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48000 | lm loss value: 4.199605E+00 | lm loss PPL: 6.665999E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48000 | lm loss value: 4.048732E+00 | lm loss PPL: 5.732471E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48010/ 52452 | consumed samples: 24581120 | consumed tokens: 50342133760 | elapsed time per iteration (s): 9.18 | learning rate: 2.323E-05 | global batch size: 512 | lm loss: 1.156370E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.776 | TFLOPs: 8.37 | 31: iteration 48020/ 52452 | consumed samples: 24586240 | consumed tokens: 50352619520 | elapsed time per iteration (s): 1.79 | learning rate: 2.322E-05 | global batch size: 512 | lm loss: 1.146294E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.326 | TFLOPs: 42.83 | 31: iteration 48030/ 52452 | consumed samples: 24591360 | consumed tokens: 50363105280 | elapsed time per iteration (s): 1.79 | learning rate: 2.320E-05 | global batch size: 512 | lm loss: 1.140801E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.380 | TFLOPs: 42.83 | 31: iteration 48040/ 52452 | consumed samples: 24596480 | consumed tokens: 50373591040 | elapsed time per iteration (s): 1.74 | learning rate: 2.319E-05 | global batch size: 512 | lm loss: 1.146551E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.654 | TFLOPs: 44.08 | 31: iteration 48050/ 52452 | consumed samples: 24601600 | consumed tokens: 50384076800 | elapsed time per iteration (s): 1.78 | learning rate: 2.317E-05 | global batch size: 512 | lm loss: 1.143773E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.324 | TFLOPs: 43.28 | 31: iteration 48060/ 52452 | consumed samples: 24606720 | consumed tokens: 50394562560 | elapsed time per iteration (s): 1.80 | learning rate: 2.316E-05 | global batch size: 512 | lm loss: 1.152756E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.650 | TFLOPs: 42.72 | 31: iteration 48070/ 52452 | consumed samples: 24611840 | consumed tokens: 50405048320 | elapsed time per iteration (s): 1.85 | learning rate: 2.314E-05 | global batch size: 512 | lm loss: 1.149787E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.330 | TFLOPs: 41.48 | 31: iteration 48080/ 52452 | consumed samples: 24616960 | consumed tokens: 50415534080 | elapsed time per iteration (s): 1.84 | learning rate: 2.313E-05 | global batch size: 512 | lm loss: 1.142271E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.961 | TFLOPs: 41.87 | 31: iteration 48090/ 52452 | consumed samples: 24622080 | consumed tokens: 50426019840 | elapsed time per iteration (s): 1.79 | learning rate: 2.312E-05 | global batch size: 512 | lm loss: 1.141071E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.245 | TFLOPs: 42.81 | 31: iteration 48100/ 52452 | consumed samples: 24627200 | consumed tokens: 50436505600 | elapsed time per iteration (s): 1.77 | learning rate: 2.310E-05 | global batch size: 512 | lm loss: 1.135942E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.974 | TFLOPs: 43.37 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48100 | lm loss value: 4.192166E+00 | lm loss PPL: 6.616594E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48100 | lm loss value: 4.051873E+00 | lm loss PPL: 5.750508E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48110/ 52452 | consumed samples: 24632320 | consumed tokens: 50446991360 | elapsed time per iteration (s): 9.18 | learning rate: 2.309E-05 | global batch size: 512 | lm loss: 1.144858E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.764 | TFLOPs: 8.37 | 31: iteration 48120/ 52452 | consumed samples: 24637440 | consumed tokens: 50457477120 | elapsed time per iteration (s): 1.79 | learning rate: 2.307E-05 | global batch size: 512 | lm loss: 1.158928E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.648 | TFLOPs: 42.87 | 31: iteration 48130/ 52452 | consumed samples: 24642560 | consumed tokens: 50467962880 | elapsed time per iteration (s): 1.77 | learning rate: 2.306E-05 | global batch size: 512 | lm loss: 1.151719E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.641 | TFLOPs: 43.32 | 31: iteration 48140/ 52452 | consumed samples: 24647680 | consumed tokens: 50478448640 | elapsed time per iteration (s): 1.76 | learning rate: 2.305E-05 | global batch size: 512 | lm loss: 1.141351E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.378 | TFLOPs: 43.73 | 31: iteration 48150/ 52452 | consumed samples: 24652800 | consumed tokens: 50488934400 | elapsed time per iteration (s): 1.77 | learning rate: 2.303E-05 | global batch size: 512 | lm loss: 1.150774E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.404 | TFLOPs: 43.44 | 31: iteration 48160/ 52452 | consumed samples: 24657920 | consumed tokens: 50499420160 | elapsed time per iteration (s): 1.79 | learning rate: 2.302E-05 | global batch size: 512 | lm loss: 1.145932E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.163 | TFLOPs: 42.95 | 31: iteration 48170/ 52452 | consumed samples: 24663040 | consumed tokens: 50509905920 | elapsed time per iteration (s): 1.86 | learning rate: 2.300E-05 | global batch size: 512 | lm loss: 1.128559E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.714 | TFLOPs: 41.23 | 31: iteration 48180/ 52452 | consumed samples: 24668160 | consumed tokens: 50520391680 | elapsed time per iteration (s): 1.75 | learning rate: 2.299E-05 | global batch size: 512 | lm loss: 1.147823E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.005 | TFLOPs: 43.98 | 31: iteration 48190/ 52452 | consumed samples: 24673280 | consumed tokens: 50530877440 | elapsed time per iteration (s): 1.74 | learning rate: 2.298E-05 | global batch size: 512 | lm loss: 1.141455E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.387 | TFLOPs: 44.19 | 31: iteration 48200/ 52452 | consumed samples: 24678400 | consumed tokens: 50541363200 | elapsed time per iteration (s): 1.74 | learning rate: 2.296E-05 | global batch size: 512 | lm loss: 1.144498E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.510 | TFLOPs: 44.05 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48200 | lm loss value: 4.193388E+00 | lm loss PPL: 6.624689E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48200 | lm loss value: 4.047225E+00 | lm loss PPL: 5.723837E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48210/ 52452 | consumed samples: 24683520 | consumed tokens: 50551848960 | elapsed time per iteration (s): 9.14 | learning rate: 2.295E-05 | global batch size: 512 | lm loss: 1.144977E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.029 | TFLOPs: 8.41 | 31: iteration 48220/ 52452 | consumed samples: 24688640 | consumed tokens: 50562334720 | elapsed time per iteration (s): 1.73 | learning rate: 2.293E-05 | global batch size: 512 | lm loss: 1.139560E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.636 | TFLOPs: 44.37 | 31: iteration 48230/ 52452 | consumed samples: 24693760 | consumed tokens: 50572820480 | elapsed time per iteration (s): 1.81 | learning rate: 2.292E-05 | global batch size: 512 | lm loss: 1.145943E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.615 | TFLOPs: 42.42 | 31: iteration 48240/ 52452 | consumed samples: 24698880 | consumed tokens: 50583306240 | elapsed time per iteration (s): 1.75 | learning rate: 2.291E-05 | global batch size: 512 | lm loss: 1.151233E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.417 | TFLOPs: 43.89 | 31: iteration 48250/ 52452 | consumed samples: 24704000 | consumed tokens: 50593792000 | elapsed time per iteration (s): 1.76 | learning rate: 2.289E-05 | global batch size: 512 | lm loss: 1.149663E+00 | grad norm: 0.179 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.893 | TFLOPs: 43.66 | 31: iteration 48260/ 52452 | consumed samples: 24709120 | consumed tokens: 50604277760 | elapsed time per iteration (s): 1.83 | learning rate: 2.288E-05 | global batch size: 512 | lm loss: 1.134612E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.705 | TFLOPs: 41.98 | 31: iteration 48270/ 52452 | consumed samples: 24714240 | consumed tokens: 50614763520 | elapsed time per iteration (s): 1.72 | learning rate: 2.287E-05 | global batch size: 512 | lm loss: 1.154469E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.090 | TFLOPs: 44.59 | 31: iteration 48280/ 52452 | consumed samples: 24719360 | consumed tokens: 50625249280 | elapsed time per iteration (s): 1.75 | learning rate: 2.285E-05 | global batch size: 512 | lm loss: 1.130796E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.241 | TFLOPs: 44.01 | 31: iteration 48290/ 52452 | consumed samples: 24724480 | consumed tokens: 50635735040 | elapsed time per iteration (s): 1.74 | learning rate: 2.284E-05 | global batch size: 512 | lm loss: 1.152243E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.567 | TFLOPs: 44.21 | 31: iteration 48300/ 52452 | consumed samples: 24729600 | consumed tokens: 50646220800 | elapsed time per iteration (s): 1.76 | learning rate: 2.282E-05 | global batch size: 512 | lm loss: 1.145228E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.400 | TFLOPs: 43.74 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48300 | lm loss value: 4.223367E+00 | lm loss PPL: 6.826292E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48300 | lm loss value: 4.077798E+00 | lm loss PPL: 5.901540E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48310/ 52452 | consumed samples: 24734720 | consumed tokens: 50656706560 | elapsed time per iteration (s): 9.23 | learning rate: 2.281E-05 | global batch size: 512 | lm loss: 1.143882E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.444 | TFLOPs: 8.32 | 31: iteration 48320/ 52452 | consumed samples: 24739840 | consumed tokens: 50667192320 | elapsed time per iteration (s): 1.79 | learning rate: 2.280E-05 | global batch size: 512 | lm loss: 1.146281E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.892 | TFLOPs: 42.91 | 31: iteration 48330/ 52452 | consumed samples: 24744960 | consumed tokens: 50677678080 | elapsed time per iteration (s): 1.77 | learning rate: 2.278E-05 | global batch size: 512 | lm loss: 1.134525E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.844 | TFLOPs: 43.35 | 31: iteration 48340/ 52452 | consumed samples: 24750080 | consumed tokens: 50688163840 | elapsed time per iteration (s): 1.78 | learning rate: 2.277E-05 | global batch size: 512 | lm loss: 1.154118E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.009 | TFLOPs: 43.23 | 31: iteration 48350/ 52452 | consumed samples: 24755200 | consumed tokens: 50698649600 | elapsed time per iteration (s): 1.75 | learning rate: 2.276E-05 | global batch size: 512 | lm loss: 1.161505E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.006 | TFLOPs: 43.83 | 31: iteration 48360/ 52452 | consumed samples: 24760320 | consumed tokens: 50709135360 | elapsed time per iteration (s): 1.78 | learning rate: 2.274E-05 | global batch size: 512 | lm loss: 1.151822E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.389 | TFLOPs: 43.29 | 31: iteration 48370/ 52452 | consumed samples: 24765440 | consumed tokens: 50719621120 | elapsed time per iteration (s): 1.73 | learning rate: 2.273E-05 | global batch size: 512 | lm loss: 1.144913E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.062 | TFLOPs: 44.44 | 31: iteration 48380/ 52452 | consumed samples: 24770560 | consumed tokens: 50730106880 | elapsed time per iteration (s): 1.75 | learning rate: 2.272E-05 | global batch size: 512 | lm loss: 1.140842E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.261 | TFLOPs: 43.87 | 31: iteration 48390/ 52452 | consumed samples: 24775680 | consumed tokens: 50740592640 | elapsed time per iteration (s): 1.79 | learning rate: 2.270E-05 | global batch size: 512 | lm loss: 1.137346E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.088 | TFLOPs: 42.94 | 31: iteration 48400/ 52452 | consumed samples: 24780800 | consumed tokens: 50751078400 | elapsed time per iteration (s): 1.79 | learning rate: 2.269E-05 | global batch size: 512 | lm loss: 1.146283E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.466 | TFLOPs: 42.85 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48400 | lm loss value: 4.202240E+00 | lm loss PPL: 6.683584E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48400 | lm loss value: 4.057399E+00 | lm loss PPL: 5.782370E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48410/ 52452 | consumed samples: 24785920 | consumed tokens: 50761564160 | elapsed time per iteration (s): 9.17 | learning rate: 2.268E-05 | global batch size: 512 | lm loss: 1.153425E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.825 | TFLOPs: 8.38 | 31: iteration 48420/ 52452 | consumed samples: 24791040 | consumed tokens: 50772049920 | elapsed time per iteration (s): 1.85 | learning rate: 2.266E-05 | global batch size: 512 | lm loss: 1.142321E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.233 | TFLOPs: 41.46 | 31: iteration 48430/ 52452 | consumed samples: 24796160 | consumed tokens: 50782535680 | elapsed time per iteration (s): 1.77 | learning rate: 2.265E-05 | global batch size: 512 | lm loss: 1.151074E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.824 | TFLOPs: 43.50 | 31: iteration 48440/ 52452 | consumed samples: 24801280 | consumed tokens: 50793021440 | elapsed time per iteration (s): 1.72 | learning rate: 2.264E-05 | global batch size: 512 | lm loss: 1.151827E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.962 | TFLOPs: 44.57 | 31: iteration 48450/ 52452 | consumed samples: 24806400 | consumed tokens: 50803507200 | elapsed time per iteration (s): 1.76 | learning rate: 2.263E-05 | global batch size: 512 | lm loss: 1.160013E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.301 | TFLOPs: 43.57 | 31: iteration 48460/ 52452 | consumed samples: 24811520 | consumed tokens: 50813992960 | elapsed time per iteration (s): 1.79 | learning rate: 2.261E-05 | global batch size: 512 | lm loss: 1.142750E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.550 | TFLOPs: 43.01 | 31: iteration 48470/ 52452 | consumed samples: 24816640 | consumed tokens: 50824478720 | elapsed time per iteration (s): 1.78 | learning rate: 2.260E-05 | global batch size: 512 | lm loss: 1.139095E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.964 | TFLOPs: 43.22 | 31: iteration 48480/ 52452 | consumed samples: 24821760 | consumed tokens: 50834964480 | elapsed time per iteration (s): 1.76 | learning rate: 2.259E-05 | global batch size: 512 | lm loss: 1.126858E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.437 | TFLOPs: 43.74 | 31: iteration 48490/ 52452 | consumed samples: 24826880 | consumed tokens: 50845450240 | elapsed time per iteration (s): 1.79 | learning rate: 2.257E-05 | global batch size: 512 | lm loss: 1.132777E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.390 | TFLOPs: 42.99 | 31: iteration 48500/ 52452 | consumed samples: 24832000 | consumed tokens: 50855936000 | elapsed time per iteration (s): 1.79 | learning rate: 2.256E-05 | global batch size: 512 | lm loss: 1.137963E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.362 | TFLOPs: 42.98 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48500 | lm loss value: 4.203738E+00 | lm loss PPL: 6.693605E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48500 | lm loss value: 4.055748E+00 | lm loss PPL: 5.772835E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48510/ 52452 | consumed samples: 24837120 | consumed tokens: 50866421760 | elapsed time per iteration (s): 9.11 | learning rate: 2.255E-05 | global batch size: 512 | lm loss: 1.143790E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.177 | TFLOPs: 8.43 | 31: iteration 48520/ 52452 | consumed samples: 24842240 | consumed tokens: 50876907520 | elapsed time per iteration (s): 1.73 | learning rate: 2.253E-05 | global batch size: 512 | lm loss: 1.146651E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.226 | TFLOPs: 44.31 | 31: iteration 48530/ 52452 | consumed samples: 24847360 | consumed tokens: 50887393280 | elapsed time per iteration (s): 1.77 | learning rate: 2.252E-05 | global batch size: 512 | lm loss: 1.156756E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.023 | TFLOPs: 43.38 | 31: iteration 48540/ 52452 | consumed samples: 24852480 | consumed tokens: 50897879040 | elapsed time per iteration (s): 1.73 | learning rate: 2.251E-05 | global batch size: 512 | lm loss: 1.143753E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.307 | TFLOPs: 44.32 | 31: iteration 48550/ 52452 | consumed samples: 24857600 | consumed tokens: 50908364800 | elapsed time per iteration (s): 1.78 | learning rate: 2.250E-05 | global batch size: 512 | lm loss: 1.146685E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.393 | TFLOPs: 43.29 | 31: iteration 48560/ 52452 | consumed samples: 24862720 | consumed tokens: 50918850560 | elapsed time per iteration (s): 1.77 | learning rate: 2.248E-05 | global batch size: 512 | lm loss: 1.147235E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.946 | TFLOPs: 43.52 | 31: iteration 48570/ 52452 | consumed samples: 24867840 | consumed tokens: 50929336320 | elapsed time per iteration (s): 1.75 | learning rate: 2.247E-05 | global batch size: 512 | lm loss: 1.154042E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.806 | TFLOPs: 43.95 | 31: iteration 48580/ 52452 | consumed samples: 24872960 | consumed tokens: 50939822080 | elapsed time per iteration (s): 1.83 | learning rate: 2.246E-05 | global batch size: 512 | lm loss: 1.158483E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.881 | TFLOPs: 42.01 | 31: iteration 48590/ 52452 | consumed samples: 24878080 | consumed tokens: 50950307840 | elapsed time per iteration (s): 1.78 | learning rate: 2.245E-05 | global batch size: 512 | lm loss: 1.143568E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.154 | TFLOPs: 43.10 | 31: iteration 48600/ 52452 | consumed samples: 24883200 | consumed tokens: 50960793600 | elapsed time per iteration (s): 1.72 | learning rate: 2.243E-05 | global batch size: 512 | lm loss: 1.151502E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.967 | TFLOPs: 44.57 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48600 | lm loss value: 4.183398E+00 | lm loss PPL: 6.558836E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48600 | lm loss value: 4.031563E+00 | lm loss PPL: 5.634893E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48610/ 52452 | consumed samples: 24888320 | consumed tokens: 50971279360 | elapsed time per iteration (s): 9.13 | learning rate: 2.242E-05 | global batch size: 512 | lm loss: 1.139623E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.103 | TFLOPs: 8.42 | 31: iteration 48620/ 52452 | consumed samples: 24893440 | consumed tokens: 50981765120 | elapsed time per iteration (s): 1.74 | learning rate: 2.241E-05 | global batch size: 512 | lm loss: 1.148144E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.353 | TFLOPs: 44.18 | 31: iteration 48630/ 52452 | consumed samples: 24898560 | consumed tokens: 50992250880 | elapsed time per iteration (s): 1.81 | learning rate: 2.240E-05 | global batch size: 512 | lm loss: 1.138419E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.600 | TFLOPs: 42.42 | 31: iteration 48640/ 52452 | consumed samples: 24903680 | consumed tokens: 51002736640 | elapsed time per iteration (s): 1.94 | learning rate: 2.238E-05 | global batch size: 512 | lm loss: 1.152038E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 264.144 | TFLOPs: 39.65 | 31: iteration 48650/ 52452 | consumed samples: 24908800 | consumed tokens: 51013222400 | elapsed time per iteration (s): 1.79 | learning rate: 2.237E-05 | global batch size: 512 | lm loss: 1.135034E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.303 | TFLOPs: 42.82 | 31: iteration 48660/ 52452 | consumed samples: 24913920 | consumed tokens: 51023708160 | elapsed time per iteration (s): 1.73 | learning rate: 2.236E-05 | global batch size: 512 | lm loss: 1.157780E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.547 | TFLOPs: 44.51 | 31: iteration 48670/ 52452 | consumed samples: 24919040 | consumed tokens: 51034193920 | elapsed time per iteration (s): 1.79 | learning rate: 2.235E-05 | global batch size: 512 | lm loss: 1.145377E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.773 | TFLOPs: 42.89 | 31: iteration 48680/ 52452 | consumed samples: 24924160 | consumed tokens: 51044679680 | elapsed time per iteration (s): 1.74 | learning rate: 2.233E-05 | global batch size: 512 | lm loss: 1.138624E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.014 | TFLOPs: 44.13 | 31: iteration 48690/ 52452 | consumed samples: 24929280 | consumed tokens: 51055165440 | elapsed time per iteration (s): 1.76 | learning rate: 2.232E-05 | global batch size: 512 | lm loss: 1.141483E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.012 | TFLOPs: 43.68 | 31: iteration 48700/ 52452 | consumed samples: 24934400 | consumed tokens: 51065651200 | elapsed time per iteration (s): 1.73 | learning rate: 2.231E-05 | global batch size: 512 | lm loss: 1.141901E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.734 | TFLOPs: 44.54 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48700 | lm loss value: 4.189104E+00 | lm loss PPL: 6.596363E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48700 | lm loss value: 4.053800E+00 | lm loss PPL: 5.761596E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48710/ 52452 | consumed samples: 24939520 | consumed tokens: 51076136960 | elapsed time per iteration (s): 9.15 | learning rate: 2.230E-05 | global batch size: 512 | lm loss: 1.143605E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.983 | TFLOPs: 8.40 | 31: iteration 48720/ 52452 | consumed samples: 24944640 | consumed tokens: 51086622720 | elapsed time per iteration (s): 1.77 | learning rate: 2.228E-05 | global batch size: 512 | lm loss: 1.145167E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.535 | TFLOPs: 43.46 | 31: iteration 48730/ 52452 | consumed samples: 24949760 | consumed tokens: 51097108480 | elapsed time per iteration (s): 1.79 | learning rate: 2.227E-05 | global batch size: 512 | lm loss: 1.144496E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.687 | TFLOPs: 43.03 | 31: iteration 48740/ 52452 | consumed samples: 24954880 | consumed tokens: 51107594240 | elapsed time per iteration (s): 1.76 | learning rate: 2.226E-05 | global batch size: 512 | lm loss: 1.147625E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.246 | TFLOPs: 43.56 | 31: iteration 48750/ 52452 | consumed samples: 24960000 | consumed tokens: 51118080000 | elapsed time per iteration (s): 1.75 | learning rate: 2.225E-05 | global batch size: 512 | lm loss: 1.145631E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.133 | TFLOPs: 44.00 | 31: iteration 48760/ 52452 | consumed samples: 24965120 | consumed tokens: 51128565760 | elapsed time per iteration (s): 1.77 | learning rate: 2.224E-05 | global batch size: 512 | lm loss: 1.146171E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.626 | TFLOPs: 43.47 | 31: iteration 48770/ 52452 | consumed samples: 24970240 | consumed tokens: 51139051520 | elapsed time per iteration (s): 1.75 | learning rate: 2.222E-05 | global batch size: 512 | lm loss: 1.141900E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.608 | TFLOPs: 43.92 | 31: iteration 48780/ 52452 | consumed samples: 24975360 | consumed tokens: 51149537280 | elapsed time per iteration (s): 1.80 | learning rate: 2.221E-05 | global batch size: 512 | lm loss: 1.151822E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.204 | TFLOPs: 42.66 | 31: iteration 48790/ 52452 | consumed samples: 24980480 | consumed tokens: 51160023040 | elapsed time per iteration (s): 1.76 | learning rate: 2.220E-05 | global batch size: 512 | lm loss: 1.147123E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.495 | TFLOPs: 43.60 | 31: iteration 48800/ 52452 | consumed samples: 24985600 | consumed tokens: 51170508800 | elapsed time per iteration (s): 1.80 | learning rate: 2.219E-05 | global batch size: 512 | lm loss: 1.151070E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.214 | TFLOPs: 42.66 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48800 | lm loss value: 4.259756E+00 | lm loss PPL: 7.079271E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48800 | lm loss value: 4.087939E+00 | lm loss PPL: 5.961691E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48810/ 52452 | consumed samples: 24990720 | consumed tokens: 51180994560 | elapsed time per iteration (s): 9.22 | learning rate: 2.218E-05 | global batch size: 512 | lm loss: 1.133034E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.546 | TFLOPs: 8.34 | 31: iteration 48820/ 52452 | consumed samples: 24995840 | consumed tokens: 51191480320 | elapsed time per iteration (s): 1.78 | learning rate: 2.216E-05 | global batch size: 512 | lm loss: 1.140795E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.946 | TFLOPs: 43.22 | 31: iteration 48830/ 52452 | consumed samples: 25000960 | consumed tokens: 51201966080 | elapsed time per iteration (s): 1.76 | learning rate: 2.215E-05 | global batch size: 512 | lm loss: 1.150124E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.639 | TFLOPs: 43.77 | 31: iteration 48840/ 52452 | consumed samples: 25006080 | consumed tokens: 51212451840 | elapsed time per iteration (s): 1.75 | learning rate: 2.214E-05 | global batch size: 512 | lm loss: 1.148612E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.252 | TFLOPs: 44.02 | 31: iteration 48850/ 52452 | consumed samples: 25011200 | consumed tokens: 51222937600 | elapsed time per iteration (s): 1.76 | learning rate: 2.213E-05 | global batch size: 512 | lm loss: 1.139179E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.263 | TFLOPs: 43.57 | 31: iteration 48860/ 52452 | consumed samples: 25016320 | consumed tokens: 51233423360 | elapsed time per iteration (s): 1.77 | learning rate: 2.212E-05 | global batch size: 512 | lm loss: 1.145142E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.993 | TFLOPs: 43.53 | 31: iteration 48870/ 52452 | consumed samples: 25021440 | consumed tokens: 51243909120 | elapsed time per iteration (s): 1.75 | learning rate: 2.211E-05 | global batch size: 512 | lm loss: 1.148593E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.895 | TFLOPs: 43.81 | 31: iteration 48880/ 52452 | consumed samples: 25026560 | consumed tokens: 51254394880 | elapsed time per iteration (s): 1.72 | learning rate: 2.209E-05 | global batch size: 512 | lm loss: 1.146017E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.889 | TFLOPs: 44.56 | 31: iteration 48890/ 52452 | consumed samples: 25031680 | consumed tokens: 51264880640 | elapsed time per iteration (s): 1.82 | learning rate: 2.208E-05 | global batch size: 512 | lm loss: 1.146381E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.642 | TFLOPs: 42.12 | 31: iteration 48900/ 52452 | consumed samples: 25036800 | consumed tokens: 51275366400 | elapsed time per iteration (s): 1.74 | learning rate: 2.207E-05 | global batch size: 512 | lm loss: 1.133848E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.852 | TFLOPs: 44.26 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 48900 | lm loss value: 4.251327E+00 | lm loss PPL: 7.019847E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 48900 | lm loss value: 4.080120E+00 | lm loss PPL: 5.915257E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 48910/ 52452 | consumed samples: 25041920 | consumed tokens: 51285852160 | elapsed time per iteration (s): 9.40 | learning rate: 2.206E-05 | global batch size: 512 | lm loss: 1.148993E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 54.452 | TFLOPs: 8.17 | 31: iteration 48920/ 52452 | consumed samples: 25047040 | consumed tokens: 51296337920 | elapsed time per iteration (s): 1.80 | learning rate: 2.205E-05 | global batch size: 512 | lm loss: 1.132555E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.998 | TFLOPs: 42.63 | 31: iteration 48930/ 52452 | consumed samples: 25052160 | consumed tokens: 51306823680 | elapsed time per iteration (s): 1.79 | learning rate: 2.204E-05 | global batch size: 512 | lm loss: 1.141189E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.441 | TFLOPs: 42.99 | 31: iteration 48940/ 52452 | consumed samples: 25057280 | consumed tokens: 51317309440 | elapsed time per iteration (s): 1.80 | learning rate: 2.202E-05 | global batch size: 512 | lm loss: 1.138809E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.553 | TFLOPs: 42.71 | 31: iteration 48950/ 52452 | consumed samples: 25062400 | consumed tokens: 51327795200 | elapsed time per iteration (s): 1.75 | learning rate: 2.201E-05 | global batch size: 512 | lm loss: 1.140877E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.044 | TFLOPs: 43.98 | 31: iteration 48960/ 52452 | consumed samples: 25067520 | consumed tokens: 51338280960 | elapsed time per iteration (s): 1.76 | learning rate: 2.200E-05 | global batch size: 512 | lm loss: 1.143129E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.132 | TFLOPs: 43.55 | 31: iteration 48970/ 52452 | consumed samples: 25072640 | consumed tokens: 51348766720 | elapsed time per iteration (s): 1.90 | learning rate: 2.199E-05 | global batch size: 512 | lm loss: 1.136114E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.133 | TFLOPs: 40.40 | 31: iteration 48980/ 52452 | consumed samples: 25077760 | consumed tokens: 51359252480 | elapsed time per iteration (s): 1.78 | learning rate: 2.198E-05 | global batch size: 512 | lm loss: 1.149071E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.869 | TFLOPs: 43.06 | 31: iteration 48990/ 52452 | consumed samples: 25082880 | consumed tokens: 51369738240 | elapsed time per iteration (s): 1.80 | learning rate: 2.197E-05 | global batch size: 512 | lm loss: 1.136361E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.818 | TFLOPs: 42.60 | 31: iteration 49000/ 52452 | consumed samples: 25088000 | consumed tokens: 51380224000 | elapsed time per iteration (s): 1.88 | learning rate: 2.196E-05 | global batch size: 512 | lm loss: 1.148247E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.783 | TFLOPs: 40.94 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49000 | lm loss value: 4.181247E+00 | lm loss PPL: 6.544743E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49000 | lm loss value: 4.036729E+00 | lm loss PPL: 5.664076E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49010/ 52452 | consumed samples: 25093120 | consumed tokens: 51390709760 | elapsed time per iteration (s): 9.13 | learning rate: 2.194E-05 | global batch size: 512 | lm loss: 1.135500E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.082 | TFLOPs: 8.42 | 31: iteration 49020/ 52452 | consumed samples: 25098240 | consumed tokens: 51401195520 | elapsed time per iteration (s): 1.78 | learning rate: 2.193E-05 | global batch size: 512 | lm loss: 1.147317E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.307 | TFLOPs: 43.27 | 31: iteration 49030/ 52452 | consumed samples: 25103360 | consumed tokens: 51411681280 | elapsed time per iteration (s): 1.79 | learning rate: 2.192E-05 | global batch size: 512 | lm loss: 1.129874E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.275 | TFLOPs: 42.97 | 31: iteration 49040/ 52452 | consumed samples: 25108480 | consumed tokens: 51422167040 | elapsed time per iteration (s): 1.77 | learning rate: 2.191E-05 | global batch size: 512 | lm loss: 1.124821E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.691 | TFLOPs: 43.33 | 31: iteration 49050/ 52452 | consumed samples: 25113600 | consumed tokens: 51432652800 | elapsed time per iteration (s): 1.82 | learning rate: 2.190E-05 | global batch size: 512 | lm loss: 1.146724E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.288 | TFLOPs: 42.22 | 31: iteration 49060/ 52452 | consumed samples: 25118720 | consumed tokens: 51443138560 | elapsed time per iteration (s): 1.76 | learning rate: 2.189E-05 | global batch size: 512 | lm loss: 1.136590E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.650 | TFLOPs: 43.78 | 31: iteration 49070/ 52452 | consumed samples: 25123840 | consumed tokens: 51453624320 | elapsed time per iteration (s): 1.77 | learning rate: 2.188E-05 | global batch size: 512 | lm loss: 1.149796E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.278 | TFLOPs: 43.42 | 31: iteration 49080/ 52452 | consumed samples: 25128960 | consumed tokens: 51464110080 | elapsed time per iteration (s): 1.75 | learning rate: 2.187E-05 | global batch size: 512 | lm loss: 1.149365E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.881 | TFLOPs: 43.96 | 31: iteration 49090/ 52452 | consumed samples: 25134080 | consumed tokens: 51474595840 | elapsed time per iteration (s): 1.76 | learning rate: 2.186E-05 | global batch size: 512 | lm loss: 1.151225E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.638 | TFLOPs: 43.62 | 31: iteration 49100/ 52452 | consumed samples: 25139200 | consumed tokens: 51485081600 | elapsed time per iteration (s): 1.78 | learning rate: 2.184E-05 | global batch size: 512 | lm loss: 1.150461E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.736 | TFLOPs: 43.19 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49100 | lm loss value: 4.267176E+00 | lm loss PPL: 7.131992E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49100 | lm loss value: 4.090341E+00 | lm loss PPL: 5.976024E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49110/ 52452 | consumed samples: 25144320 | consumed tokens: 51495567360 | elapsed time per iteration (s): 9.13 | learning rate: 2.183E-05 | global batch size: 512 | lm loss: 1.136841E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.095 | TFLOPs: 8.42 | 31: iteration 49120/ 52452 | consumed samples: 25149440 | consumed tokens: 51506053120 | elapsed time per iteration (s): 1.74 | learning rate: 2.182E-05 | global batch size: 512 | lm loss: 1.148350E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.541 | TFLOPs: 44.06 | 31: iteration 49130/ 52452 | consumed samples: 25154560 | consumed tokens: 51516538880 | elapsed time per iteration (s): 1.77 | learning rate: 2.181E-05 | global batch size: 512 | lm loss: 1.139046E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.712 | TFLOPs: 43.33 | 31: iteration 49140/ 52452 | consumed samples: 25159680 | consumed tokens: 51527024640 | elapsed time per iteration (s): 1.84 | learning rate: 2.180E-05 | global batch size: 512 | lm loss: 1.145776E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.507 | TFLOPs: 41.80 | 31: iteration 49150/ 52452 | consumed samples: 25164800 | consumed tokens: 51537510400 | elapsed time per iteration (s): 1.81 | learning rate: 2.179E-05 | global batch size: 512 | lm loss: 1.142655E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.284 | TFLOPs: 42.52 | 31: iteration 49160/ 52452 | consumed samples: 25169920 | consumed tokens: 51547996160 | elapsed time per iteration (s): 1.84 | learning rate: 2.178E-05 | global batch size: 512 | lm loss: 1.137461E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.881 | TFLOPs: 41.71 | 31: iteration 49170/ 52452 | consumed samples: 25175040 | consumed tokens: 51558481920 | elapsed time per iteration (s): 1.79 | learning rate: 2.177E-05 | global batch size: 512 | lm loss: 1.140757E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.707 | TFLOPs: 43.03 | 31: iteration 49180/ 52452 | consumed samples: 25180160 | consumed tokens: 51568967680 | elapsed time per iteration (s): 1.75 | learning rate: 2.176E-05 | global batch size: 512 | lm loss: 1.147761E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.778 | TFLOPs: 43.79 | 31: iteration 49190/ 52452 | consumed samples: 25185280 | consumed tokens: 51579453440 | elapsed time per iteration (s): 1.80 | learning rate: 2.175E-05 | global batch size: 512 | lm loss: 1.157199E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.049 | TFLOPs: 42.63 | 31: iteration 49200/ 52452 | consumed samples: 25190400 | consumed tokens: 51589939200 | elapsed time per iteration (s): 1.74 | learning rate: 2.174E-05 | global batch size: 512 | lm loss: 1.143480E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.530 | TFLOPs: 44.06 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49200 | lm loss value: 4.260701E+00 | lm loss PPL: 7.085962E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49200 | lm loss value: 4.085608E+00 | lm loss PPL: 5.947809E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49210/ 52452 | consumed samples: 25195520 | consumed tokens: 51600424960 | elapsed time per iteration (s): 9.13 | learning rate: 2.173E-05 | global batch size: 512 | lm loss: 1.142326E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.064 | TFLOPs: 8.41 | 31: iteration 49220/ 52452 | consumed samples: 25200640 | consumed tokens: 51610910720 | elapsed time per iteration (s): 1.83 | learning rate: 2.172E-05 | global batch size: 512 | lm loss: 1.145669E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.694 | TFLOPs: 41.98 | 31: iteration 49230/ 52452 | consumed samples: 25205760 | consumed tokens: 51621396480 | elapsed time per iteration (s): 1.76 | learning rate: 2.170E-05 | global batch size: 512 | lm loss: 1.153834E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.472 | TFLOPs: 43.75 | 31: iteration 49240/ 52452 | consumed samples: 25210880 | consumed tokens: 51631882240 | elapsed time per iteration (s): 1.74 | learning rate: 2.169E-05 | global batch size: 512 | lm loss: 1.139951E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.463 | TFLOPs: 44.05 | 31: iteration 49250/ 52452 | consumed samples: 25216000 | consumed tokens: 51642368000 | elapsed time per iteration (s): 1.76 | learning rate: 2.168E-05 | global batch size: 512 | lm loss: 1.151664E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.377 | TFLOPs: 43.58 | 31: iteration 49260/ 52452 | consumed samples: 25221120 | consumed tokens: 51652853760 | elapsed time per iteration (s): 1.77 | learning rate: 2.167E-05 | global batch size: 512 | lm loss: 1.144346E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.519 | TFLOPs: 43.31 | 31: iteration 49270/ 52452 | consumed samples: 25226240 | consumed tokens: 51663339520 | elapsed time per iteration (s): 1.77 | learning rate: 2.166E-05 | global batch size: 512 | lm loss: 1.148644E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.873 | TFLOPs: 43.36 | 31: iteration 49280/ 52452 | consumed samples: 25231360 | consumed tokens: 51673825280 | elapsed time per iteration (s): 1.76 | learning rate: 2.165E-05 | global batch size: 512 | lm loss: 1.142460E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.461 | TFLOPs: 43.60 | 31: iteration 49290/ 52452 | consumed samples: 25236480 | consumed tokens: 51684311040 | elapsed time per iteration (s): 1.74 | learning rate: 2.164E-05 | global batch size: 512 | lm loss: 1.141685E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.788 | TFLOPs: 44.25 | 31: iteration 49300/ 52452 | consumed samples: 25241600 | consumed tokens: 51694796800 | elapsed time per iteration (s): 1.78 | learning rate: 2.163E-05 | global batch size: 512 | lm loss: 1.152701E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.894 | TFLOPs: 43.21 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49300 | lm loss value: 4.203483E+00 | lm loss PPL: 6.691901E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49300 | lm loss value: 4.054070E+00 | lm loss PPL: 5.763154E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49310/ 52452 | consumed samples: 25246720 | consumed tokens: 51705282560 | elapsed time per iteration (s): 9.19 | learning rate: 2.162E-05 | global batch size: 512 | lm loss: 1.148013E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.711 | TFLOPs: 8.36 | 31: iteration 49320/ 52452 | consumed samples: 25251840 | consumed tokens: 51715768320 | elapsed time per iteration (s): 1.78 | learning rate: 2.161E-05 | global batch size: 512 | lm loss: 1.145998E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.498 | TFLOPs: 43.15 | 31: iteration 49330/ 52452 | consumed samples: 25256960 | consumed tokens: 51726254080 | elapsed time per iteration (s): 1.78 | learning rate: 2.160E-05 | global batch size: 512 | lm loss: 1.135981E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.946 | TFLOPs: 43.07 | 31: iteration 49340/ 52452 | consumed samples: 25262080 | consumed tokens: 51736739840 | elapsed time per iteration (s): 1.80 | learning rate: 2.159E-05 | global batch size: 512 | lm loss: 1.144839E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.883 | TFLOPs: 42.76 | 31: iteration 49350/ 52452 | consumed samples: 25267200 | consumed tokens: 51747225600 | elapsed time per iteration (s): 1.76 | learning rate: 2.158E-05 | global batch size: 512 | lm loss: 1.146306E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.342 | TFLOPs: 43.58 | 31: iteration 49360/ 52452 | consumed samples: 25272320 | consumed tokens: 51757711360 | elapsed time per iteration (s): 1.73 | learning rate: 2.157E-05 | global batch size: 512 | lm loss: 1.140157E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.921 | TFLOPs: 44.42 | 31: iteration 49370/ 52452 | consumed samples: 25277440 | consumed tokens: 51768197120 | elapsed time per iteration (s): 2.18 | learning rate: 2.156E-05 | global batch size: 512 | lm loss: 1.137087E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 234.515 | TFLOPs: 35.20 | 31: iteration 49380/ 52452 | consumed samples: 25282560 | consumed tokens: 51778682880 | elapsed time per iteration (s): 1.83 | learning rate: 2.155E-05 | global batch size: 512 | lm loss: 1.133048E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.706 | TFLOPs: 41.98 | 31: iteration 49390/ 52452 | consumed samples: 25287680 | consumed tokens: 51789168640 | elapsed time per iteration (s): 1.79 | learning rate: 2.154E-05 | global batch size: 512 | lm loss: 1.140654E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.603 | TFLOPs: 42.87 | 31: iteration 49400/ 52452 | consumed samples: 25292800 | consumed tokens: 51799654400 | elapsed time per iteration (s): 1.75 | learning rate: 2.153E-05 | global batch size: 512 | lm loss: 1.142218E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.438 | TFLOPs: 43.89 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49400 | lm loss value: 4.205071E+00 | lm loss PPL: 6.702539E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49400 | lm loss value: 4.055129E+00 | lm loss PPL: 5.769258E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49410/ 52452 | consumed samples: 25297920 | consumed tokens: 51810140160 | elapsed time per iteration (s): 9.21 | learning rate: 2.152E-05 | global batch size: 512 | lm loss: 1.143970E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.589 | TFLOPs: 8.34 | 31: iteration 49420/ 52452 | consumed samples: 25303040 | consumed tokens: 51820625920 | elapsed time per iteration (s): 1.78 | learning rate: 2.151E-05 | global batch size: 512 | lm loss: 1.144144E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.318 | TFLOPs: 43.27 | 31: iteration 49430/ 52452 | consumed samples: 25308160 | consumed tokens: 51831111680 | elapsed time per iteration (s): 1.82 | learning rate: 2.150E-05 | global batch size: 512 | lm loss: 1.129342E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.652 | TFLOPs: 42.27 | 31: iteration 49440/ 52452 | consumed samples: 25313280 | consumed tokens: 51841597440 | elapsed time per iteration (s): 1.83 | learning rate: 2.149E-05 | global batch size: 512 | lm loss: 1.139283E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.451 | TFLOPs: 42.09 | 31: iteration 49450/ 52452 | consumed samples: 25318400 | consumed tokens: 51852083200 | elapsed time per iteration (s): 1.74 | learning rate: 2.148E-05 | global batch size: 512 | lm loss: 1.140184E+00 | grad norm: 0.207 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.603 | TFLOPs: 44.07 | 31: iteration 49460/ 52452 | consumed samples: 25323520 | consumed tokens: 51862568960 | elapsed time per iteration (s): 1.76 | learning rate: 2.147E-05 | global batch size: 512 | lm loss: 1.140429E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.205 | TFLOPs: 43.56 | 31: iteration 49470/ 52452 | consumed samples: 25328640 | consumed tokens: 51873054720 | elapsed time per iteration (s): 1.78 | learning rate: 2.146E-05 | global batch size: 512 | lm loss: 1.145704E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.933 | TFLOPs: 43.22 | 31: iteration 49480/ 52452 | consumed samples: 25333760 | consumed tokens: 51883540480 | elapsed time per iteration (s): 1.77 | learning rate: 2.145E-05 | global batch size: 512 | lm loss: 1.150414E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.456 | TFLOPs: 43.45 | 31: iteration 49490/ 52452 | consumed samples: 25338880 | consumed tokens: 51894026240 | elapsed time per iteration (s): 1.77 | learning rate: 2.144E-05 | global batch size: 512 | lm loss: 1.148758E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.861 | TFLOPs: 43.36 | 31: iteration 49500/ 52452 | consumed samples: 25344000 | consumed tokens: 51904512000 | elapsed time per iteration (s): 1.73 | learning rate: 2.143E-05 | global batch size: 512 | lm loss: 1.137457E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.215 | TFLOPs: 44.31 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49500 | lm loss value: 4.272891E+00 | lm loss PPL: 7.172871E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49500 | lm loss value: 4.087196E+00 | lm loss PPL: 5.957261E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49510/ 52452 | consumed samples: 25349120 | consumed tokens: 51914997760 | elapsed time per iteration (s): 9.23 | learning rate: 2.142E-05 | global batch size: 512 | lm loss: 1.140195E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.458 | TFLOPs: 8.32 | 31: iteration 49520/ 52452 | consumed samples: 25354240 | consumed tokens: 51925483520 | elapsed time per iteration (s): 1.78 | learning rate: 2.141E-05 | global batch size: 512 | lm loss: 1.143859E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.866 | TFLOPs: 43.06 | 31: iteration 49530/ 52452 | consumed samples: 25359360 | consumed tokens: 51935969280 | elapsed time per iteration (s): 1.79 | learning rate: 2.140E-05 | global batch size: 512 | lm loss: 1.135603E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.196 | TFLOPs: 42.96 | 31: iteration 49540/ 52452 | consumed samples: 25364480 | consumed tokens: 51946455040 | elapsed time per iteration (s): 1.75 | learning rate: 2.139E-05 | global batch size: 512 | lm loss: 1.152354E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.651 | TFLOPs: 43.93 | 31: iteration 49550/ 52452 | consumed samples: 25369600 | consumed tokens: 51956940800 | elapsed time per iteration (s): 1.76 | learning rate: 2.138E-05 | global batch size: 512 | lm loss: 1.139093E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.260 | TFLOPs: 43.57 | 31: iteration 49560/ 52452 | consumed samples: 25374720 | consumed tokens: 51967426560 | elapsed time per iteration (s): 1.80 | learning rate: 2.137E-05 | global batch size: 512 | lm loss: 1.144472E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.408 | TFLOPs: 42.69 | 31: iteration 49570/ 52452 | consumed samples: 25379840 | consumed tokens: 51977912320 | elapsed time per iteration (s): 1.79 | learning rate: 2.136E-05 | global batch size: 512 | lm loss: 1.148420E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.139 | TFLOPs: 42.95 | 31: iteration 49580/ 52452 | consumed samples: 25384960 | consumed tokens: 51988398080 | elapsed time per iteration (s): 1.85 | learning rate: 2.136E-05 | global batch size: 512 | lm loss: 1.144512E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.467 | TFLOPs: 41.50 | 31: iteration 49590/ 52452 | consumed samples: 25390080 | consumed tokens: 51998883840 | elapsed time per iteration (s): 1.78 | learning rate: 2.135E-05 | global batch size: 512 | lm loss: 1.143903E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.907 | TFLOPs: 43.06 | 31: iteration 49600/ 52452 | consumed samples: 25395200 | consumed tokens: 52009369600 | elapsed time per iteration (s): 1.81 | learning rate: 2.134E-05 | global batch size: 512 | lm loss: 1.137211E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.405 | TFLOPs: 42.54 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49600 | lm loss value: 4.193376E+00 | lm loss PPL: 6.624603E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49600 | lm loss value: 4.048240E+00 | lm loss PPL: 5.729654E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49610/ 52452 | consumed samples: 25400320 | consumed tokens: 52019855360 | elapsed time per iteration (s): 9.19 | learning rate: 2.133E-05 | global batch size: 512 | lm loss: 1.130354E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.729 | TFLOPs: 8.36 | 31: iteration 49620/ 52452 | consumed samples: 25405440 | consumed tokens: 52030341120 | elapsed time per iteration (s): 1.81 | learning rate: 2.132E-05 | global batch size: 512 | lm loss: 1.148885E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.238 | TFLOPs: 42.51 | 31: iteration 49630/ 52452 | consumed samples: 25410560 | consumed tokens: 52040826880 | elapsed time per iteration (s): 1.78 | learning rate: 2.131E-05 | global batch size: 512 | lm loss: 1.144216E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.484 | TFLOPs: 43.15 | 31: iteration 49640/ 52452 | consumed samples: 25415680 | consumed tokens: 52051312640 | elapsed time per iteration (s): 1.80 | learning rate: 2.130E-05 | global batch size: 512 | lm loss: 1.135259E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.867 | TFLOPs: 42.61 | 31: iteration 49650/ 52452 | consumed samples: 25420800 | consumed tokens: 52061798400 | elapsed time per iteration (s): 1.75 | learning rate: 2.129E-05 | global batch size: 512 | lm loss: 1.138257E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.270 | TFLOPs: 44.02 | 31: iteration 49660/ 52452 | consumed samples: 25425920 | consumed tokens: 52072284160 | elapsed time per iteration (s): 1.82 | learning rate: 2.128E-05 | global batch size: 512 | lm loss: 1.132184E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.839 | TFLOPs: 42.30 | 31: iteration 49670/ 52452 | consumed samples: 25431040 | consumed tokens: 52082769920 | elapsed time per iteration (s): 1.86 | learning rate: 2.127E-05 | global batch size: 512 | lm loss: 1.143065E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.665 | TFLOPs: 41.23 | 31: iteration 49680/ 52452 | consumed samples: 25436160 | consumed tokens: 52093255680 | elapsed time per iteration (s): 1.78 | learning rate: 2.126E-05 | global batch size: 512 | lm loss: 1.140549E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.051 | TFLOPs: 43.23 | 31: iteration 49690/ 52452 | consumed samples: 25441280 | consumed tokens: 52103741440 | elapsed time per iteration (s): 1.75 | learning rate: 2.125E-05 | global batch size: 512 | lm loss: 1.141171E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.655 | TFLOPs: 43.93 | 31: iteration 49700/ 52452 | consumed samples: 25446400 | consumed tokens: 52114227200 | elapsed time per iteration (s): 1.77 | learning rate: 2.124E-05 | global batch size: 512 | lm loss: 1.148533E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.731 | TFLOPs: 43.34 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49700 | lm loss value: 4.258737E+00 | lm loss PPL: 7.072061E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49700 | lm loss value: 4.089554E+00 | lm loss PPL: 5.971324E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49710/ 52452 | consumed samples: 25451520 | consumed tokens: 52124712960 | elapsed time per iteration (s): 9.21 | learning rate: 2.124E-05 | global batch size: 512 | lm loss: 1.134352E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.569 | TFLOPs: 8.34 | 31: iteration 49720/ 52452 | consumed samples: 25456640 | consumed tokens: 52135198720 | elapsed time per iteration (s): 1.78 | learning rate: 2.123E-05 | global batch size: 512 | lm loss: 1.128177E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.386 | TFLOPs: 43.14 | 31: iteration 49730/ 52452 | consumed samples: 25461760 | consumed tokens: 52145684480 | elapsed time per iteration (s): 1.75 | learning rate: 2.122E-05 | global batch size: 512 | lm loss: 1.138535E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.939 | TFLOPs: 43.97 | 31: iteration 49740/ 52452 | consumed samples: 25466880 | consumed tokens: 52156170240 | elapsed time per iteration (s): 1.78 | learning rate: 2.121E-05 | global batch size: 512 | lm loss: 1.134347E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.175 | TFLOPs: 43.25 | 31: iteration 49750/ 52452 | consumed samples: 25472000 | consumed tokens: 52166656000 | elapsed time per iteration (s): 1.75 | learning rate: 2.120E-05 | global batch size: 512 | lm loss: 1.153697E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.729 | TFLOPs: 43.94 | 31: iteration 49760/ 52452 | consumed samples: 25477120 | consumed tokens: 52177141760 | elapsed time per iteration (s): 1.79 | learning rate: 2.119E-05 | global batch size: 512 | lm loss: 1.141231E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.248 | TFLOPs: 42.81 | 31: iteration 49770/ 52452 | consumed samples: 25482240 | consumed tokens: 52187627520 | elapsed time per iteration (s): 1.77 | learning rate: 2.118E-05 | global batch size: 512 | lm loss: 1.140168E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.557 | TFLOPs: 43.31 | 31: iteration 49780/ 52452 | consumed samples: 25487360 | consumed tokens: 52198113280 | elapsed time per iteration (s): 1.89 | learning rate: 2.117E-05 | global batch size: 512 | lm loss: 1.137598E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.893 | TFLOPs: 40.66 | 31: iteration 49790/ 52452 | consumed samples: 25492480 | consumed tokens: 52208599040 | elapsed time per iteration (s): 1.74 | learning rate: 2.116E-05 | global batch size: 512 | lm loss: 1.152462E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.860 | TFLOPs: 44.11 | 31: iteration 49800/ 52452 | consumed samples: 25497600 | consumed tokens: 52219084800 | elapsed time per iteration (s): 1.83 | learning rate: 2.116E-05 | global batch size: 512 | lm loss: 1.138747E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.072 | TFLOPs: 42.04 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49800 | lm loss value: 4.165111E+00 | lm loss PPL: 6.439983E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49800 | lm loss value: 4.031827E+00 | lm loss PPL: 5.636382E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49810/ 52452 | consumed samples: 25502720 | consumed tokens: 52229570560 | elapsed time per iteration (s): 9.18 | learning rate: 2.115E-05 | global batch size: 512 | lm loss: 1.140929E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.795 | TFLOPs: 8.37 | 31: iteration 49820/ 52452 | consumed samples: 25507840 | consumed tokens: 52240056320 | elapsed time per iteration (s): 1.74 | learning rate: 2.114E-05 | global batch size: 512 | lm loss: 1.141474E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.568 | TFLOPs: 44.06 | 31: iteration 49830/ 52452 | consumed samples: 25512960 | consumed tokens: 52250542080 | elapsed time per iteration (s): 1.78 | learning rate: 2.113E-05 | global batch size: 512 | lm loss: 1.141807E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.227 | TFLOPs: 43.11 | 31: iteration 49840/ 52452 | consumed samples: 25518080 | consumed tokens: 52261027840 | elapsed time per iteration (s): 1.73 | learning rate: 2.112E-05 | global batch size: 512 | lm loss: 1.142038E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.877 | TFLOPs: 44.41 | 31: iteration 49850/ 52452 | consumed samples: 25523200 | consumed tokens: 52271513600 | elapsed time per iteration (s): 1.78 | learning rate: 2.111E-05 | global batch size: 512 | lm loss: 1.127144E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.241 | TFLOPs: 43.26 | 31: iteration 49860/ 52452 | consumed samples: 25528320 | consumed tokens: 52281999360 | elapsed time per iteration (s): 1.75 | learning rate: 2.110E-05 | global batch size: 512 | lm loss: 1.134080E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.204 | TFLOPs: 44.01 | 31: iteration 49870/ 52452 | consumed samples: 25533440 | consumed tokens: 52292485120 | elapsed time per iteration (s): 1.80 | learning rate: 2.110E-05 | global batch size: 512 | lm loss: 1.136251E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.557 | TFLOPs: 42.71 | 31: iteration 49880/ 52452 | consumed samples: 25538560 | consumed tokens: 52302970880 | elapsed time per iteration (s): 1.73 | learning rate: 2.109E-05 | global batch size: 512 | lm loss: 1.131563E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.403 | TFLOPs: 44.34 | 31: iteration 49890/ 52452 | consumed samples: 25543680 | consumed tokens: 52313456640 | elapsed time per iteration (s): 1.74 | learning rate: 2.108E-05 | global batch size: 512 | lm loss: 1.145422E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.734 | TFLOPs: 44.09 | 31: iteration 49900/ 52452 | consumed samples: 25548800 | consumed tokens: 52323942400 | elapsed time per iteration (s): 1.81 | learning rate: 2.107E-05 | global batch size: 512 | lm loss: 1.142998E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.918 | TFLOPs: 42.46 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 49900 | lm loss value: 4.250565E+00 | lm loss PPL: 7.014504E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 49900 | lm loss value: 4.086362E+00 | lm loss PPL: 5.952295E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 49910/ 52452 | consumed samples: 25553920 | consumed tokens: 52334428160 | elapsed time per iteration (s): 9.13 | learning rate: 2.106E-05 | global batch size: 512 | lm loss: 1.149775E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.067 | TFLOPs: 8.42 | 31: iteration 49920/ 52452 | consumed samples: 25559040 | consumed tokens: 52344913920 | elapsed time per iteration (s): 1.77 | learning rate: 2.105E-05 | global batch size: 512 | lm loss: 1.149736E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.493 | TFLOPs: 43.30 | 31: iteration 49930/ 52452 | consumed samples: 25564160 | consumed tokens: 52355399680 | elapsed time per iteration (s): 1.87 | learning rate: 2.105E-05 | global batch size: 512 | lm loss: 1.138159E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.420 | TFLOPs: 41.19 | 31: iteration 49940/ 52452 | consumed samples: 25569280 | consumed tokens: 52365885440 | elapsed time per iteration (s): 1.76 | learning rate: 2.104E-05 | global batch size: 512 | lm loss: 1.136196E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.155 | TFLOPs: 43.55 | 31: iteration 49950/ 52452 | consumed samples: 25574400 | consumed tokens: 52376371200 | elapsed time per iteration (s): 1.80 | learning rate: 2.103E-05 | global batch size: 512 | lm loss: 1.123544E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.602 | TFLOPs: 42.72 | 31: iteration 49960/ 52452 | consumed samples: 25579520 | consumed tokens: 52386856960 | elapsed time per iteration (s): 1.74 | learning rate: 2.102E-05 | global batch size: 512 | lm loss: 1.137905E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.766 | TFLOPs: 44.09 | 31: iteration 49970/ 52452 | consumed samples: 25584640 | consumed tokens: 52397342720 | elapsed time per iteration (s): 1.77 | learning rate: 2.101E-05 | global batch size: 512 | lm loss: 1.138905E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.788 | TFLOPs: 43.50 | 31: iteration 49980/ 52452 | consumed samples: 25589760 | consumed tokens: 52407828480 | elapsed time per iteration (s): 1.78 | learning rate: 2.100E-05 | global batch size: 512 | lm loss: 1.141696E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.486 | TFLOPs: 43.15 | 31: iteration 49990/ 52452 | consumed samples: 25594880 | consumed tokens: 52418314240 | elapsed time per iteration (s): 1.79 | learning rate: 2.100E-05 | global batch size: 512 | lm loss: 1.151017E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.244 | TFLOPs: 42.96 | 0: [2022-12-26 20:08:15,613] [INFO] [logging.py:68:log_dist] [Rank 0] step=50000, skipped=0, lr=[2.0988534318797655e-05, 2.0988534318797655e-05, 2.0988534318797655e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 50000/ 52452 | consumed samples: 25600000 | consumed tokens: 52428800000 | elapsed time per iteration (s): 1.83 | learning rate: 2.099E-05 | global batch size: 512 | lm loss: 1.142111E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.414 | TFLOPs: 42.09 | 0: steps: 50000 loss: 1.1581 iter time (s): 1.781 samples/sec: 287.401 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50000 | lm loss value: 4.180789E+00 | lm loss PPL: 6.541748E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50000 | lm loss value: 4.044436E+00 | lm loss PPL: 5.707896E+01 | 31: ------------------------------------------------------------------------------------------------------ 0: saving checkpoint at iteration 50000 to checkpoints_2b855brealtasky 0: [2022-12-26 20:09:29,323] [INFO] [logging.py:68:log_dist] [Rank 0] [Torch] Checkpoint global_step50000 is begin to save! 0: [2022-12-26 20:09:29,332] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_01-model_00-model_states.pt... 0: [2022-12-26 20:09:29,728] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_01-model_00-model_states.pt. 0: [2022-12-26 20:09:29,728] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_03-model_00-model_states.pt... 0: [2022-12-26 20:09:29,915] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_03-model_00-model_states.pt. 0: [2022-12-26 20:09:29,916] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_04-model_00-model_states.pt... 0: [2022-12-26 20:09:30,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_04-model_00-model_states.pt. 0: [2022-12-26 20:09:30,096] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_05-model_00-model_states.pt... 0: [2022-12-26 20:09:30,278] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_05-model_00-model_states.pt. 0: [2022-12-26 20:09:30,279] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_06-model_00-model_states.pt... 0: [2022-12-26 20:09:30,455] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_06-model_00-model_states.pt. 0: [2022-12-26 20:09:30,456] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_07-model_00-model_states.pt... 0: [2022-12-26 20:09:30,635] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_07-model_00-model_states.pt. 0: [2022-12-26 20:09:30,636] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_08-model_00-model_states.pt... 0: [2022-12-26 20:09:30,813] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_08-model_00-model_states.pt. 0: [2022-12-26 20:09:30,813] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_09-model_00-model_states.pt... 0: [2022-12-26 20:09:30,987] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_09-model_00-model_states.pt. 0: [2022-12-26 20:09:30,988] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_10-model_00-model_states.pt... 0: [2022-12-26 20:09:31,167] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_10-model_00-model_states.pt. 0: [2022-12-26 20:09:31,168] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_11-model_00-model_states.pt... 0: [2022-12-26 20:09:31,342] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_11-model_00-model_states.pt. 0: [2022-12-26 20:09:31,342] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_12-model_00-model_states.pt... 0: [2022-12-26 20:09:31,526] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_12-model_00-model_states.pt. 0: [2022-12-26 20:09:31,527] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_13-model_00-model_states.pt... 0: [2022-12-26 20:09:31,704] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_13-model_00-model_states.pt. 0: [2022-12-26 20:09:31,704] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_14-model_00-model_states.pt... 0: [2022-12-26 20:09:31,885] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_14-model_00-model_states.pt. 0: [2022-12-26 20:09:31,885] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_15-model_00-model_states.pt... 0: [2022-12-26 20:09:32,058] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_15-model_00-model_states.pt. 0: [2022-12-26 20:09:32,058] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_16-model_00-model_states.pt... 0: [2022-12-26 20:09:32,233] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_16-model_00-model_states.pt. 0: [2022-12-26 20:09:32,234] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_17-model_00-model_states.pt... 0: [2022-12-26 20:09:32,417] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_17-model_00-model_states.pt. 0: [2022-12-26 20:09:32,418] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_18-model_00-model_states.pt... 0: [2022-12-26 20:09:32,589] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_18-model_00-model_states.pt. 0: [2022-12-26 20:09:32,589] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_19-model_00-model_states.pt... 0: [2022-12-26 20:09:32,761] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_19-model_00-model_states.pt. 0: [2022-12-26 20:09:32,762] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_20-model_00-model_states.pt... 0: [2022-12-26 20:09:32,936] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_20-model_00-model_states.pt. 0: [2022-12-26 20:09:32,937] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_21-model_00-model_states.pt... 0: [2022-12-26 20:09:33,110] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_21-model_00-model_states.pt. 0: [2022-12-26 20:09:33,110] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_22-model_00-model_states.pt... 0: [2022-12-26 20:09:33,284] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_22-model_00-model_states.pt. 0: [2022-12-26 20:09:33,285] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_23-model_00-model_states.pt... 0: [2022-12-26 20:09:33,460] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_23-model_00-model_states.pt. 0: [2022-12-26 20:09:33,461] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_24-model_00-model_states.pt... 0: [2022-12-26 20:09:33,634] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_24-model_00-model_states.pt. 0: [2022-12-26 20:09:33,635] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_25-model_00-model_states.pt... 0: [2022-12-26 20:09:33,806] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_25-model_00-model_states.pt. 0: [2022-12-26 20:09:33,806] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_26-model_00-model_states.pt... 0: [2022-12-26 20:09:33,978] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_26-model_00-model_states.pt. 0: [2022-12-26 20:09:33,978] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_27-model_00-model_states.pt... 0: [2022-12-26 20:09:34,144] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_27-model_00-model_states.pt. 0: [2022-12-26 20:09:34,145] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_28-model_00-model_states.pt... 0: [2022-12-26 20:09:34,320] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_28-model_00-model_states.pt. 0: [2022-12-26 20:09:34,320] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_29-model_00-model_states.pt... 0: [2022-12-26 20:09:34,493] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_29-model_00-model_states.pt. 0: [2022-12-26 20:09:34,494] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_30-model_00-model_states.pt... 0: [2022-12-26 20:09:34,660] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_30-model_00-model_states.pt. 0: [2022-12-26 20:09:34,661] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_31-model_00-model_states.pt... 0: [2022-12-26 20:09:34,838] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_31-model_00-model_states.pt. 0: [2022-12-26 20:09:34,838] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_32-model_00-model_states.pt... 0: [2022-12-26 20:09:35,005] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_32-model_00-model_states.pt. 0: [2022-12-26 20:09:35,006] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_33-model_00-model_states.pt... 0: [2022-12-26 20:09:35,178] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_33-model_00-model_states.pt. 0: [2022-12-26 20:09:35,178] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_34-model_00-model_states.pt... 0: [2022-12-26 20:09:35,342] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_34-model_00-model_states.pt. 0: [2022-12-26 20:09:35,343] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_35-model_00-model_states.pt... 0: [2022-12-26 20:09:35,511] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_35-model_00-model_states.pt. 0: [2022-12-26 20:09:35,511] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_36-model_00-model_states.pt... 0: [2022-12-26 20:09:35,684] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_36-model_00-model_states.pt. 0: [2022-12-26 20:09:35,684] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/layer_38-model_00-model_states.pt... 0: [2022-12-26 20:09:35,686] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/layer_38-model_00-model_states.pt. 0: [2022-12-26 20:09:35,687] [INFO] [logging.py:68:log_dist] [Rank 0] Saving model checkpoint: checkpoints_2b855brealtasky/global_step50000/mp_rank_00_model_states.pt 0: [2022-12-26 20:09:35,687] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/mp_rank_00_model_states.pt... 0: [2022-12-26 20:09:35,693] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/mp_rank_00_model_states.pt. 0: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... 0: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... 3: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt... 24: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt... 24: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt... 24: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt... 30: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt... 10: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt... 10: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt... 0: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... 0: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... 0: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... 16: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt... 14: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt... 14: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt... 14: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt... 14: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt... 14: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt... 12: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt... 12: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt... 12: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt... 12: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt... 12: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt... 3: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt... 3: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt... 3: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt... 3: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt... 3: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt... 18: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt... 18: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt... 8: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt... 8: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt... 8: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt... 8: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt... 8: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt... 5: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt... 5: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt... 4: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt... 4: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt... 4: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt... 4: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt... 4: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt... 7: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt... 9: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt... 13: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt... 24: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt... 24: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt... 17: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt... 17: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt... 17: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt... 17: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt... 17: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt... 28: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt... 28: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt... 28: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt... 28: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt... 28: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt... 21: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt... 21: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt... 21: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt... 21: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt... 30: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt... 30: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt... 30: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt... 30: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt... 26: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt... 26: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt... 26: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt... 19: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt... 19: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt... 25: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt... 25: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt... 25: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt... 25: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt... 25: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt... 31: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt... 31: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt... 31: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt... 29: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt... 29: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt... 29: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt... 29: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt... 29: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt... 11: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt... 10: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt... 10: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt... 10: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt... 10: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt... 23: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt... 23: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt... 23: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt... 23: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt... 20: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt... 20: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt... 20: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt... 20: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt... 20: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt... 2: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt... 2: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt... 2: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt... 2: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt... 2: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt... 27: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt... 27: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt... 27: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt... 27: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt... 27: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt... 22: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt... 22: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt... 22: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt... 22: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt... 22: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt... 1: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... 1: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... 1: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... 1: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... 1: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... 6: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt... 6: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt... 6: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt... 6: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt... 6: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt... 15: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt... 15: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt... 15: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt... 15: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt... 15: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt... 0: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... 16: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt... 16: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt... 16: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt... 14: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt... 12: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt... 12: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt... 12: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt... 3: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt... 3: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt... 18: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt... 18: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt... 18: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt... 18: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt... 8: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt... 5: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt... 5: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt... 5: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt... 4: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt... 7: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt... 7: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt... 9: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt... 9: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt... 13: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt... 13: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt... 13: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt... 13: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt... 13: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt... 24: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt... 24: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt... 24: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt... 17: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt... 28: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt... 28: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt... 21: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt... 21: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt... 21: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt... 30: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt... 26: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt... 26: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt... 26: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt... 19: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt... 19: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt... 19: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt... 19: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt... 25: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt... 31: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt... 31: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt... 31: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt... 29: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt... 11: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt... 10: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt... 10: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt... 23: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt... 23: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt... 23: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt... 23: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt... 20: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt... 20: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt... 20: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt... 2: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt... 2: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt... 2: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt... 27: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt... 22: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt... 1: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... 1: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... 1: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... 6: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt... 6: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt... 15: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt... 15: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt... 0: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... 16: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt... 16: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt... 16: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt... 14: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt... 18: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt... 8: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt... 8: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt... 5: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt... 5: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt... 4: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt... 4: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt... 7: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt... 9: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt... 9: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt... 13: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt... 17: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt... 17: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt... 28: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt... 21: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt... 30: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt... 26: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt... 26: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt... 19: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt... 25: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt... 31: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt... 29: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt... 29: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt... 27: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt... 27: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt... 22: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt... 6: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt... 15: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt... 0: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... 16: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt... 14: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt... 18: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt... 5: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt... 7: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt... 9: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt... 9: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt... 13: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt... 30: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt... 19: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt... 25: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt... 31: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt... 11: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt... 7: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt... 9: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt... 11: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt... 11: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt... 22: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt... 7: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt... 7: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt... 11: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt... 11: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt... 11: [2022-12-26 20:09:35,778] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt... 0: [2022-12-26 20:09:35,919] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. 0: [2022-12-26 20:09:35,940] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. 0: [2022-12-26 20:09:35,940] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt 0: [2022-12-26 20:09:35,941] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 0: [2022-12-26 20:09:35,941] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. 0: [2022-12-26 20:09:35,941] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt 0: [2022-12-26 20:09:35,941] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 0: [2022-12-26 20:09:35,941] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. 0: [2022-12-26 20:09:35,941] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. 0: [2022-12-26 20:09:35,941] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt 0: [2022-12-26 20:09:35,941] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt 0: [2022-12-26 20:09:35,941] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 0: [2022-12-26 20:09:35,941] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt. 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt. 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt. 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt. 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt. 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt. 17: [2022-12-26 20:09:35,971] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt 17: [2022-12-26 20:09:35,971] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt 17: [2022-12-26 20:09:35,971] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt 17: [2022-12-26 20:09:35,971] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt 17: [2022-12-26 20:09:35,971] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 17: [2022-12-26 20:09:35,971] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt 17: [2022-12-26 20:09:35,971] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 27: [2022-12-26 20:09:35,972] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt. 27: [2022-12-26 20:09:35,972] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt. 27: [2022-12-26 20:09:35,972] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt 27: [2022-12-26 20:09:35,972] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt. 27: [2022-12-26 20:09:35,972] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt 27: [2022-12-26 20:09:35,972] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 27: [2022-12-26 20:09:35,972] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt 27: [2022-12-26 20:09:35,972] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 27: [2022-12-26 20:09:35,972] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 27: [2022-12-26 20:09:35,973] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt. 27: [2022-12-26 20:09:35,973] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt 27: [2022-12-26 20:09:35,973] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 27: [2022-12-26 20:09:35,973] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt. 27: [2022-12-26 20:09:35,973] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt 27: [2022-12-26 20:09:35,973] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 0: [2022-12-26 20:09:35,983] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. 0: [2022-12-26 20:09:35,984] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt 0: [2022-12-26 20:09:35,984] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. 0: [2022-12-26 20:09:35,984] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 0: [2022-12-26 20:09:35,984] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt 0: [2022-12-26 20:09:35,984] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 0: [2022-12-26 20:09:35,984] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. 0: [2022-12-26 20:09:35,984] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt 0: [2022-12-26 20:09:35,984] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 22: [2022-12-26 20:09:35,986] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt. 22: [2022-12-26 20:09:35,986] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt. 22: [2022-12-26 20:09:35,986] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt. 22: [2022-12-26 20:09:35,986] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt 22: [2022-12-26 20:09:35,986] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt 22: [2022-12-26 20:09:35,986] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt 22: [2022-12-26 20:09:35,987] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 22: [2022-12-26 20:09:35,987] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 22: [2022-12-26 20:09:35,987] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 22: [2022-12-26 20:09:35,987] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt. 22: [2022-12-26 20:09:35,987] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt 22: [2022-12-26 20:09:35,988] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 4: [2022-12-26 20:09:35,989] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt. 4: [2022-12-26 20:09:35,989] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt. 4: [2022-12-26 20:09:35,989] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt. 4: [2022-12-26 20:09:35,989] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt 4: [2022-12-26 20:09:35,989] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt 4: [2022-12-26 20:09:35,989] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt 4: [2022-12-26 20:09:35,989] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 4: [2022-12-26 20:09:35,989] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 4: [2022-12-26 20:09:35,989] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 4: [2022-12-26 20:09:35,989] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt. 4: [2022-12-26 20:09:35,989] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt 4: [2022-12-26 20:09:35,989] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 4: [2022-12-26 20:09:35,989] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt. 4: [2022-12-26 20:09:35,990] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt 4: [2022-12-26 20:09:35,990] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 4: [2022-12-26 20:09:35,990] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt. 4: [2022-12-26 20:09:35,990] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt 4: [2022-12-26 20:09:35,990] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt. 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt. 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt. 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt. 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt. 24: [2022-12-26 20:09:35,994] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt 24: [2022-12-26 20:09:35,994] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt. 24: [2022-12-26 20:09:35,994] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 24: [2022-12-26 20:09:35,994] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 24: [2022-12-26 20:09:35,994] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt 24: [2022-12-26 20:09:35,994] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 24: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 22: [2022-12-26 20:09:35,994] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt. 22: [2022-12-26 20:09:35,994] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt 22: [2022-12-26 20:09:35,995] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 22: [2022-12-26 20:09:35,995] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt. 22: [2022-12-26 20:09:35,995] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt 22: [2022-12-26 20:09:35,995] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt. 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt. 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt. 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt. 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt. 25: [2022-12-26 20:09:36,002] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt 25: [2022-12-26 20:09:36,002] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt 25: [2022-12-26 20:09:36,002] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt. 25: [2022-12-26 20:09:36,002] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt 25: [2022-12-26 20:09:36,002] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 25: [2022-12-26 20:09:36,002] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt 25: [2022-12-26 20:09:36,002] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 27: [2022-12-26 20:09:36,007] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt. 27: [2022-12-26 20:09:36,007] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt 27: [2022-12-26 20:09:36,007] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt. 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt. 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt. 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt. 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt. 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt. 20: [2022-12-26 20:09:36,011] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt 20: [2022-12-26 20:09:36,011] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt 20: [2022-12-26 20:09:36,011] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt 20: [2022-12-26 20:09:36,011] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt 20: [2022-12-26 20:09:36,011] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt 20: [2022-12-26 20:09:36,011] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 20: [2022-12-26 20:09:36,011] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 6: [2022-12-26 20:09:36,013] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt. 6: [2022-12-26 20:09:36,013] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt. 6: [2022-12-26 20:09:36,013] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt. 6: [2022-12-26 20:09:36,013] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt. 6: [2022-12-26 20:09:36,013] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt. 6: [2022-12-26 20:09:36,013] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt. 6: [2022-12-26 20:09:36,014] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt 6: [2022-12-26 20:09:36,014] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt 6: [2022-12-26 20:09:36,014] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt 6: [2022-12-26 20:09:36,014] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt 6: [2022-12-26 20:09:36,014] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt 6: [2022-12-26 20:09:36,014] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt 6: [2022-12-26 20:09:36,014] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 6: [2022-12-26 20:09:36,014] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 6: [2022-12-26 20:09:36,014] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 6: [2022-12-26 20:09:36,014] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 6: [2022-12-26 20:09:36,014] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 6: [2022-12-26 20:09:36,014] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt. 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt. 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt. 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt. 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt. 18: [2022-12-26 20:09:36,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt. 18: [2022-12-26 20:09:36,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt. 18: [2022-12-26 20:09:36,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt. 18: [2022-12-26 20:09:36,021] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt. 18: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt. 18: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt. 18: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt 18: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt 18: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt 18: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 18: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 18: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt 18: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt 18: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 18: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt 18: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 18: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 18: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 26: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt. 26: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt. 26: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt. 26: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt. 26: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt. 26: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt 26: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt 26: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt 26: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt. 26: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt 26: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 26: [2022-12-26 20:09:36,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt 26: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 26: [2022-12-26 20:09:36,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 26: [2022-12-26 20:09:36,023] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt 26: [2022-12-26 20:09:36,023] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 26: [2022-12-26 20:09:36,023] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 26: [2022-12-26 20:09:36,023] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 23: [2022-12-26 20:09:36,041] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt. 23: [2022-12-26 20:09:36,041] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt. 23: [2022-12-26 20:09:36,041] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt. 23: [2022-12-26 20:09:36,041] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt. 23: [2022-12-26 20:09:36,041] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt 23: [2022-12-26 20:09:36,041] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt 23: [2022-12-26 20:09:36,041] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt 23: [2022-12-26 20:09:36,041] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 23: [2022-12-26 20:09:36,041] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt 23: [2022-12-26 20:09:36,041] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 23: [2022-12-26 20:09:36,041] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 23: [2022-12-26 20:09:36,041] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt. 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt. 30: [2022-12-26 20:09:36,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt 30: [2022-12-26 20:09:36,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt 30: [2022-12-26 20:09:36,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt 30: [2022-12-26 20:09:36,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt 30: [2022-12-26 20:09:36,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 30: [2022-12-26 20:09:36,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt 30: [2022-12-26 20:09:36,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 30: [2022-12-26 20:09:36,004] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt. 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt. 14: [2022-12-26 20:09:36,055] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt. 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt. 28: [2022-12-26 20:09:36,057] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt. 28: [2022-12-26 20:09:36,057] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt. 28: [2022-12-26 20:09:36,057] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt. 28: [2022-12-26 20:09:36,057] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt. 28: [2022-12-26 20:09:36,057] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt. 28: [2022-12-26 20:09:36,057] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt. 28: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt 28: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt 28: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt 28: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 28: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 28: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 28: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt 28: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt 28: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt 28: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 28: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 28: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt. 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt. 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt. 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt. 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt. 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt. 7: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt 7: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt 7: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt 7: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 7: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 7: [2022-12-26 20:09:36,058] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 7: [2022-12-26 20:09:36,058] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 7: [2022-12-26 20:09:36,059] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 7: [2022-12-26 20:09:36,059] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 23: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt. 23: [2022-12-26 20:09:36,060] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt 23: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt. 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt. 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt. 11: [2022-12-26 20:09:36,060] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt 11: [2022-12-26 20:09:36,060] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt 11: [2022-12-26 20:09:36,060] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt. 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt. 11: [2022-12-26 20:09:36,060] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt. 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 11: [2022-12-26 20:09:36,060] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt 11: [2022-12-26 20:09:36,060] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 11: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 23: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt. 23: [2022-12-26 20:09:36,060] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt 23: [2022-12-26 20:09:36,060] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt. 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt. 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt. 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt. 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt. 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt. 2: [2022-12-26 20:09:36,061] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt 2: [2022-12-26 20:09:36,061] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt 2: [2022-12-26 20:09:36,061] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt 2: [2022-12-26 20:09:36,061] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt 2: [2022-12-26 20:09:36,061] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt 2: [2022-12-26 20:09:36,061] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 2: [2022-12-26 20:09:36,061] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt. 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt. 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt. 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt. 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt. 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt. 15: [2022-12-26 20:09:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt 15: [2022-12-26 20:09:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt 15: [2022-12-26 20:09:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt 15: [2022-12-26 20:09:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt 15: [2022-12-26 20:09:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 15: [2022-12-26 20:09:36,067] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 15: [2022-12-26 20:09:36,067] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt. 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt. 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt. 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt. 13: [2022-12-26 20:09:36,069] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt 13: [2022-12-26 20:09:36,069] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt 13: [2022-12-26 20:09:36,069] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt 13: [2022-12-26 20:09:36,069] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt. 13: [2022-12-26 20:09:36,070] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt 13: [2022-12-26 20:09:36,069] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt. 13: [2022-12-26 20:09:36,070] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 13: [2022-12-26 20:09:36,070] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt 13: [2022-12-26 20:09:36,070] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt. 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt. 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt. 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt. 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt. 29: [2022-12-26 20:09:36,065] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt. 29: [2022-12-26 20:09:36,065] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt 29: [2022-12-26 20:09:36,065] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt 29: [2022-12-26 20:09:36,065] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 29: [2022-12-26 20:09:36,065] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt 29: [2022-12-26 20:09:36,065] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 29: [2022-12-26 20:09:36,065] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt. 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt. 14: [2022-12-26 20:09:36,056] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt 14: [2022-12-26 20:09:36,056] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt 14: [2022-12-26 20:09:36,056] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt 14: [2022-12-26 20:09:36,056] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt 14: [2022-12-26 20:09:36,056] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 14: [2022-12-26 20:09:36,056] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 14: [2022-12-26 20:09:36,056] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 31: [2022-12-26 20:09:36,072] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt. 31: [2022-12-26 20:09:36,072] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,074] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt 9: [2022-12-26 20:09:36,074] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,074] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt 9: [2022-12-26 20:09:36,074] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 9: [2022-12-26 20:09:36,074] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt 9: [2022-12-26 20:09:36,074] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 9: [2022-12-26 20:09:36,074] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 31: [2022-12-26 20:09:36,072] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt. 31: [2022-12-26 20:09:36,072] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt 31: [2022-12-26 20:09:36,072] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt 31: [2022-12-26 20:09:36,072] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 31: [2022-12-26 20:09:36,072] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt 31: [2022-12-26 20:09:36,072] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 31: [2022-12-26 20:09:36,072] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 31: [2022-12-26 20:09:36,072] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt. 31: [2022-12-26 20:09:36,072] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt 31: [2022-12-26 20:09:36,072] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 31: [2022-12-26 20:09:36,073] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt. 31: [2022-12-26 20:09:36,073] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt 31: [2022-12-26 20:09:36,073] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 31: [2022-12-26 20:09:36,073] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt. 31: [2022-12-26 20:09:36,073] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt 31: [2022-12-26 20:09:36,073] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt. 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt. 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt. 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt. 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt. 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt. 21: [2022-12-26 20:09:36,076] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt 21: [2022-12-26 20:09:36,076] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt 21: [2022-12-26 20:09:36,076] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt 21: [2022-12-26 20:09:36,076] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt 21: [2022-12-26 20:09:36,076] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 21: [2022-12-26 20:09:36,076] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt 21: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 23: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt. 23: [2022-12-26 20:09:36,076] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt 23: [2022-12-26 20:09:36,076] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt. 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt. 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt. 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt. 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt. 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt. 3: [2022-12-26 20:09:36,075] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt 3: [2022-12-26 20:09:36,075] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt 3: [2022-12-26 20:09:36,075] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt 3: [2022-12-26 20:09:36,075] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt 3: [2022-12-26 20:09:36,075] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt 3: [2022-12-26 20:09:36,075] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 3: [2022-12-26 20:09:36,075] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt. 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt. 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt. 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt. 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt. 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt. 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt. 16: [2022-12-26 20:09:36,095] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt 16: [2022-12-26 20:09:36,095] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt 16: [2022-12-26 20:09:36,095] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt 16: [2022-12-26 20:09:36,095] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt 16: [2022-12-26 20:09:36,095] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt 16: [2022-12-26 20:09:36,095] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt 16: [2022-12-26 20:09:36,095] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 16: [2022-12-26 20:09:36,095] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 1: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. 1: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. 1: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. 1: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. 1: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. 1: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. 1: [2022-12-26 20:09:36,096] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt 1: [2022-12-26 20:09:36,096] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt 1: [2022-12-26 20:09:36,097] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt 1: [2022-12-26 20:09:36,097] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt 1: [2022-12-26 20:09:36,097] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 1: [2022-12-26 20:09:36,097] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt 1: [2022-12-26 20:09:36,097] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 1: [2022-12-26 20:09:36,097] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt 1: [2022-12-26 20:09:36,097] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 1: [2022-12-26 20:09:36,097] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 1: [2022-12-26 20:09:36,097] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 1: [2022-12-26 20:09:36,097] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt. 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt. 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt. 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt. 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt. 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt. 5: [2022-12-26 20:09:36,096] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt 5: [2022-12-26 20:09:36,096] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt 5: [2022-12-26 20:09:36,096] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 5: [2022-12-26 20:09:36,096] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 5: [2022-12-26 20:09:36,096] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt 5: [2022-12-26 20:09:36,096] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 5: [2022-12-26 20:09:36,096] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 19: [2022-12-26 20:09:36,110] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt. 19: [2022-12-26 20:09:36,110] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt. 19: [2022-12-26 20:09:36,110] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt. 19: [2022-12-26 20:09:36,110] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt. 19: [2022-12-26 20:09:36,110] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt. 19: [2022-12-26 20:09:36,110] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt. 19: [2022-12-26 20:09:36,110] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt 19: [2022-12-26 20:09:36,110] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt. 19: [2022-12-26 20:09:36,110] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt 19: [2022-12-26 20:09:36,110] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt 19: [2022-12-26 20:09:36,110] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt 19: [2022-12-26 20:09:36,110] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt 19: [2022-12-26 20:09:36,110] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt 19: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 19: [2022-12-26 20:09:36,111] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt 19: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 19: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 19: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 19: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 19: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 19: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt. 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt. 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt. 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt. 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt. 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt. 8: [2022-12-26 20:09:36,111] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt 8: [2022-12-26 20:09:36,111] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt 8: [2022-12-26 20:09:36,111] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt 8: [2022-12-26 20:09:36,111] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt 8: [2022-12-26 20:09:36,111] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt 8: [2022-12-26 20:09:36,111] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 8: [2022-12-26 20:09:36,111] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt. 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt. 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt. 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt. 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt. 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt. 10: [2022-12-26 20:09:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt 10: [2022-12-26 20:09:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt 21: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt. 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 10: [2022-12-26 20:09:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt 10: [2022-12-26 20:09:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt 10: [2022-12-26 20:09:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt 21: [2022-12-26 20:09:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 21: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 10: [2022-12-26 20:09:36,115] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 10: [2022-12-26 20:09:36,115] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt 10: [2022-12-26 20:09:36,116] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt. 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt. 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt. 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt. 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt. 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt. 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt. 12: [2022-12-26 20:09:36,124] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt 12: [2022-12-26 20:09:36,124] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 12: [2022-12-26 20:09:36,124] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt 12: [2022-12-26 20:09:36,124] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt 12: [2022-12-26 20:09:36,124] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt 12: [2022-12-26 20:09:36,124] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt 12: [2022-12-26 20:09:36,124] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 12: [2022-12-26 20:09:36,124] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 0: [2022-12-26 20:09:36,133] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt 0: [2022-12-26 20:09:36,133] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 21: [2022-12-26 20:09:36,133] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt. 21: [2022-12-26 20:09:36,133] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt 21: [2022-12-26 20:09:36,133] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 10: [2022-12-26 20:09:36,183] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt. 10: [2022-12-26 20:09:36,183] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt 10: [2022-12-26 20:09:36,183] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 10: [2022-12-26 20:09:36,205] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt. 10: [2022-12-26 20:09:36,205] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt 10: [2022-12-26 20:09:36,205] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 25: [2022-12-26 20:09:36,217] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt. 25: [2022-12-26 20:09:36,217] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt 25: [2022-12-26 20:09:36,217] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 13: [2022-12-26 20:09:36,220] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt. 13: [2022-12-26 20:09:36,220] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt 13: [2022-12-26 20:09:36,220] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 11: [2022-12-26 20:09:36,222] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt. 11: [2022-12-26 20:09:36,222] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt 11: [2022-12-26 20:09:36,222] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 26: [2022-12-26 20:09:36,224] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt. 26: [2022-12-26 20:09:36,224] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt 26: [2022-12-26 20:09:36,224] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 20: [2022-12-26 20:09:36,224] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt. 20: [2022-12-26 20:09:36,225] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt 20: [2022-12-26 20:09:36,225] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 3: [2022-12-26 20:09:36,227] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt. 3: [2022-12-26 20:09:36,227] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt 3: [2022-12-26 20:09:36,227] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 4: [2022-12-26 20:09:36,229] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt. 4: [2022-12-26 20:09:36,230] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt 4: [2022-12-26 20:09:36,230] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 28: [2022-12-26 20:09:36,229] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt. 28: [2022-12-26 20:09:36,230] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt 28: [2022-12-26 20:09:36,230] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 31: [2022-12-26 20:09:36,231] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt. 31: [2022-12-26 20:09:36,231] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt 31: [2022-12-26 20:09:36,231] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 30: [2022-12-26 20:09:36,230] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt. 30: [2022-12-26 20:09:36,230] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt 30: [2022-12-26 20:09:36,230] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 23: [2022-12-26 20:09:36,235] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt. 23: [2022-12-26 20:09:36,235] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt 23: [2022-12-26 20:09:36,235] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 29: [2022-12-26 20:09:36,236] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt. 29: [2022-12-26 20:09:36,236] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt 29: [2022-12-26 20:09:36,236] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 17: [2022-12-26 20:09:36,236] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt. 17: [2022-12-26 20:09:36,236] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt 17: [2022-12-26 20:09:36,236] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 26: [2022-12-26 20:09:36,237] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt. 26: [2022-12-26 20:09:36,237] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt 26: [2022-12-26 20:09:36,237] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 14: [2022-12-26 20:09:36,239] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt. 6: [2022-12-26 20:09:36,239] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt. 14: [2022-12-26 20:09:36,239] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt 14: [2022-12-26 20:09:36,239] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 6: [2022-12-26 20:09:36,239] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt 6: [2022-12-26 20:09:36,239] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 9: [2022-12-26 20:09:36,239] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,239] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt 9: [2022-12-26 20:09:36,239] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 18: [2022-12-26 20:09:36,239] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt. 18: [2022-12-26 20:09:36,239] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt 18: [2022-12-26 20:09:36,239] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 5: [2022-12-26 20:09:36,240] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt. 5: [2022-12-26 20:09:36,240] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt 5: [2022-12-26 20:09:36,240] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 20: [2022-12-26 20:09:36,242] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt. 20: [2022-12-26 20:09:36,242] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt 20: [2022-12-26 20:09:36,242] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 25: [2022-12-26 20:09:36,242] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt. 24: [2022-12-26 20:09:36,242] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt. 25: [2022-12-26 20:09:36,242] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt 25: [2022-12-26 20:09:36,242] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 24: [2022-12-26 20:09:36,242] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt 24: [2022-12-26 20:09:36,242] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 6: [2022-12-26 20:09:36,243] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt. 6: [2022-12-26 20:09:36,243] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt 6: [2022-12-26 20:09:36,243] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 2: [2022-12-26 20:09:36,243] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt. 22: [2022-12-26 20:09:36,243] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt. 22: [2022-12-26 20:09:36,243] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt 22: [2022-12-26 20:09:36,243] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 2: [2022-12-26 20:09:36,243] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt 2: [2022-12-26 20:09:36,243] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 2: [2022-12-26 20:09:36,245] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt. 2: [2022-12-26 20:09:36,246] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt 2: [2022-12-26 20:09:36,246] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 9: [2022-12-26 20:09:36,246] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,246] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt 13: [2022-12-26 20:09:36,246] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt. 9: [2022-12-26 20:09:36,246] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 13: [2022-12-26 20:09:36,246] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt 13: [2022-12-26 20:09:36,246] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 17: [2022-12-26 20:09:36,246] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt. 17: [2022-12-26 20:09:36,246] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt 17: [2022-12-26 20:09:36,246] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 31: [2022-12-26 20:09:36,246] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt. 31: [2022-12-26 20:09:36,246] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt 31: [2022-12-26 20:09:36,247] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 7: [2022-12-26 20:09:36,247] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt. 7: [2022-12-26 20:09:36,248] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt 7: [2022-12-26 20:09:36,248] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 15: [2022-12-26 20:09:36,248] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt. 15: [2022-12-26 20:09:36,248] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt 15: [2022-12-26 20:09:36,248] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 24: [2022-12-26 20:09:36,249] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt. 24: [2022-12-26 20:09:36,249] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt 24: [2022-12-26 20:09:36,249] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 27: [2022-12-26 20:09:36,249] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt. 27: [2022-12-26 20:09:36,249] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt 27: [2022-12-26 20:09:36,249] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 19: [2022-12-26 20:09:36,249] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt. 19: [2022-12-26 20:09:36,249] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt 19: [2022-12-26 20:09:36,249] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 11: [2022-12-26 20:09:36,249] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt. 11: [2022-12-26 20:09:36,250] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt 11: [2022-12-26 20:09:36,250] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 28: [2022-12-26 20:09:36,250] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt. 28: [2022-12-26 20:09:36,250] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt 28: [2022-12-26 20:09:36,250] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 29: [2022-12-26 20:09:36,251] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt. 29: [2022-12-26 20:09:36,251] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt 29: [2022-12-26 20:09:36,251] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 7: [2022-12-26 20:09:36,251] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt. 7: [2022-12-26 20:09:36,252] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt 7: [2022-12-26 20:09:36,252] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 8: [2022-12-26 20:09:36,252] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt. 4: [2022-12-26 20:09:36,252] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt. 8: [2022-12-26 20:09:36,252] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt 4: [2022-12-26 20:09:36,252] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt 8: [2022-12-26 20:09:36,252] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 22: [2022-12-26 20:09:36,252] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt. 4: [2022-12-26 20:09:36,252] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 22: [2022-12-26 20:09:36,252] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt 22: [2022-12-26 20:09:36,252] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 15: [2022-12-26 20:09:36,252] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt. 15: [2022-12-26 20:09:36,252] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt 15: [2022-12-26 20:09:36,252] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 12: [2022-12-26 20:09:36,253] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt. 12: [2022-12-26 20:09:36,253] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt 12: [2022-12-26 20:09:36,253] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 3: [2022-12-26 20:09:36,254] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt. 3: [2022-12-26 20:09:36,254] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt 3: [2022-12-26 20:09:36,254] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 5: [2022-12-26 20:09:36,257] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt. 5: [2022-12-26 20:09:36,257] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt 5: [2022-12-26 20:09:36,257] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 14: [2022-12-26 20:09:36,258] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt. 14: [2022-12-26 20:09:36,258] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt 14: [2022-12-26 20:09:36,258] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 1: [2022-12-26 20:09:36,259] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. 1: [2022-12-26 20:09:36,259] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt 1: [2022-12-26 20:09:36,259] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 8: [2022-12-26 20:09:36,261] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt. 8: [2022-12-26 20:09:36,261] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt 18: [2022-12-26 20:09:36,261] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt. 8: [2022-12-26 20:09:36,261] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 18: [2022-12-26 20:09:36,261] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt 18: [2022-12-26 20:09:36,261] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 27: [2022-12-26 20:09:36,262] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt. 27: [2022-12-26 20:09:36,262] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt 27: [2022-12-26 20:09:36,262] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 1: [2022-12-26 20:09:36,263] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. 1: [2022-12-26 20:09:36,263] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt 1: [2022-12-26 20:09:36,263] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 16: [2022-12-26 20:09:36,264] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt. 16: [2022-12-26 20:09:36,264] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step50000/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt 16: [2022-12-26 20:09:36,264] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step50000 is ready now! 0: successfully saved checkpoint at iteration 50000 to checkpoints_2b855brealtasky 31: time (ms) | save-checkpoint: 6997.45 31: iteration 50010/ 52452 | consumed samples: 25605120 | consumed tokens: 52439285760 | elapsed time per iteration (s): 9.81 | learning rate: 2.098E-05 | global batch size: 512 | lm loss: 1.129361E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 52.180 | TFLOPs: 7.83 | 31: iteration 50020/ 52452 | consumed samples: 25610240 | consumed tokens: 52449771520 | elapsed time per iteration (s): 1.74 | learning rate: 2.097E-05 | global batch size: 512 | lm loss: 1.146127E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.119 | TFLOPs: 44.15 | 31: iteration 50030/ 52452 | consumed samples: 25615360 | consumed tokens: 52460257280 | elapsed time per iteration (s): 1.85 | learning rate: 2.096E-05 | global batch size: 512 | lm loss: 1.153123E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.796 | TFLOPs: 41.55 | 31: iteration 50040/ 52452 | consumed samples: 25620480 | consumed tokens: 52470743040 | elapsed time per iteration (s): 1.81 | learning rate: 2.096E-05 | global batch size: 512 | lm loss: 1.137841E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.303 | TFLOPs: 42.37 | 31: iteration 50050/ 52452 | consumed samples: 25625600 | consumed tokens: 52481228800 | elapsed time per iteration (s): 1.75 | learning rate: 2.095E-05 | global batch size: 512 | lm loss: 1.137846E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.827 | TFLOPs: 43.80 | 31: iteration 50060/ 52452 | consumed samples: 25630720 | consumed tokens: 52491714560 | elapsed time per iteration (s): 1.76 | learning rate: 2.094E-05 | global batch size: 512 | lm loss: 1.136229E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.488 | TFLOPs: 43.75 | 31: iteration 50070/ 52452 | consumed samples: 25635840 | consumed tokens: 52502200320 | elapsed time per iteration (s): 1.80 | learning rate: 2.093E-05 | global batch size: 512 | lm loss: 1.142213E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.154 | TFLOPs: 42.65 | 31: iteration 50080/ 52452 | consumed samples: 25640960 | consumed tokens: 52512686080 | elapsed time per iteration (s): 1.90 | learning rate: 2.093E-05 | global batch size: 512 | lm loss: 1.145212E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.321 | TFLOPs: 40.42 | 31: iteration 50090/ 52452 | consumed samples: 25646080 | consumed tokens: 52523171840 | elapsed time per iteration (s): 1.75 | learning rate: 2.092E-05 | global batch size: 512 | lm loss: 1.147561E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.206 | TFLOPs: 43.86 | 31: iteration 50100/ 52452 | consumed samples: 25651200 | consumed tokens: 52533657600 | elapsed time per iteration (s): 1.84 | learning rate: 2.091E-05 | global batch size: 512 | lm loss: 1.148583E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.677 | TFLOPs: 41.83 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50100 | lm loss value: 4.261245E+00 | lm loss PPL: 7.089821E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50100 | lm loss value: 4.091883E+00 | lm loss PPL: 5.985247E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 50110/ 52452 | consumed samples: 25656320 | consumed tokens: 52544143360 | elapsed time per iteration (s): 9.16 | learning rate: 2.090E-05 | global batch size: 512 | lm loss: 1.140686E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.886 | TFLOPs: 8.39 | 31: iteration 50120/ 52452 | consumed samples: 25661440 | consumed tokens: 52554629120 | elapsed time per iteration (s): 1.76 | learning rate: 2.089E-05 | global batch size: 512 | lm loss: 1.142406E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.259 | TFLOPs: 43.57 | 31: iteration 50130/ 52452 | consumed samples: 25666560 | consumed tokens: 52565114880 | elapsed time per iteration (s): 1.81 | learning rate: 2.089E-05 | global batch size: 512 | lm loss: 1.143497E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.392 | TFLOPs: 42.39 | 31: iteration 50140/ 52452 | consumed samples: 25671680 | consumed tokens: 52575600640 | elapsed time per iteration (s): 1.73 | learning rate: 2.088E-05 | global batch size: 512 | lm loss: 1.128485E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.292 | TFLOPs: 44.32 | 31: iteration 50150/ 52452 | consumed samples: 25676800 | consumed tokens: 52586086400 | elapsed time per iteration (s): 1.81 | learning rate: 2.087E-05 | global batch size: 512 | lm loss: 1.149070E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.829 | TFLOPs: 42.45 | 31: iteration 50160/ 52452 | consumed samples: 25681920 | consumed tokens: 52596572160 | elapsed time per iteration (s): 1.83 | learning rate: 2.086E-05 | global batch size: 512 | lm loss: 1.148568E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.101 | TFLOPs: 41.89 | 31: iteration 50170/ 52452 | consumed samples: 25687040 | consumed tokens: 52607057920 | elapsed time per iteration (s): 1.78 | learning rate: 2.086E-05 | global batch size: 512 | lm loss: 1.140287E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.520 | TFLOPs: 43.16 | 31: iteration 50180/ 52452 | consumed samples: 25692160 | consumed tokens: 52617543680 | elapsed time per iteration (s): 1.84 | learning rate: 2.085E-05 | global batch size: 512 | lm loss: 1.149812E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.239 | TFLOPs: 41.76 | 31: iteration 50190/ 52452 | consumed samples: 25697280 | consumed tokens: 52628029440 | elapsed time per iteration (s): 1.74 | learning rate: 2.084E-05 | global batch size: 512 | lm loss: 1.138475E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.076 | TFLOPs: 44.29 | 31: iteration 50200/ 52452 | consumed samples: 25702400 | consumed tokens: 52638515200 | elapsed time per iteration (s): 1.74 | learning rate: 2.083E-05 | global batch size: 512 | lm loss: 1.149949E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.893 | TFLOPs: 44.26 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50200 | lm loss value: 4.275260E+00 | lm loss PPL: 7.189883E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50200 | lm loss value: 4.103772E+00 | lm loss PPL: 6.056830E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 50210/ 52452 | consumed samples: 25707520 | consumed tokens: 52649000960 | elapsed time per iteration (s): 9.17 | learning rate: 2.083E-05 | global batch size: 512 | lm loss: 1.139518E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.859 | TFLOPs: 8.38 | 31: iteration 50220/ 52452 | consumed samples: 25712640 | consumed tokens: 52659486720 | elapsed time per iteration (s): 1.79 | learning rate: 2.082E-05 | global batch size: 512 | lm loss: 1.138945E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.732 | TFLOPs: 42.89 | 31: iteration 50230/ 52452 | consumed samples: 25717760 | consumed tokens: 52669972480 | elapsed time per iteration (s): 1.76 | learning rate: 2.081E-05 | global batch size: 512 | lm loss: 1.140875E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.099 | TFLOPs: 43.69 | 31: iteration 50240/ 52452 | consumed samples: 25722880 | consumed tokens: 52680458240 | elapsed time per iteration (s): 1.87 | learning rate: 2.080E-05 | global batch size: 512 | lm loss: 1.155674E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.075 | TFLOPs: 41.14 | 31: iteration 50250/ 52452 | consumed samples: 25728000 | consumed tokens: 52690944000 | elapsed time per iteration (s): 1.76 | learning rate: 2.080E-05 | global batch size: 512 | lm loss: 1.145588E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.506 | TFLOPs: 43.60 | 31: iteration 50260/ 52452 | consumed samples: 25733120 | consumed tokens: 52701429760 | elapsed time per iteration (s): 1.93 | learning rate: 2.079E-05 | global batch size: 512 | lm loss: 1.143064E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 264.981 | TFLOPs: 39.77 | 31: iteration 50270/ 52452 | consumed samples: 25738240 | consumed tokens: 52711915520 | elapsed time per iteration (s): 1.75 | learning rate: 2.078E-05 | global batch size: 512 | lm loss: 1.143187E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.754 | TFLOPs: 43.79 | 31: iteration 50280/ 52452 | consumed samples: 25743360 | consumed tokens: 52722401280 | elapsed time per iteration (s): 1.76 | learning rate: 2.078E-05 | global batch size: 512 | lm loss: 1.147467E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.437 | TFLOPs: 43.74 | 31: iteration 50290/ 52452 | consumed samples: 25748480 | consumed tokens: 52732887040 | elapsed time per iteration (s): 1.77 | learning rate: 2.077E-05 | global batch size: 512 | lm loss: 1.136435E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.503 | TFLOPs: 43.30 | 31: iteration 50300/ 52452 | consumed samples: 25753600 | consumed tokens: 52743372800 | elapsed time per iteration (s): 1.80 | learning rate: 2.076E-05 | global batch size: 512 | lm loss: 1.140265E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.573 | TFLOPs: 42.71 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50300 | lm loss value: 4.188574E+00 | lm loss PPL: 6.592873E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50300 | lm loss value: 4.049486E+00 | lm loss PPL: 5.736794E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 50310/ 52452 | consumed samples: 25758720 | consumed tokens: 52753858560 | elapsed time per iteration (s): 9.13 | learning rate: 2.075E-05 | global batch size: 512 | lm loss: 1.149817E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.073 | TFLOPs: 8.42 | 31: iteration 50320/ 52452 | consumed samples: 25763840 | consumed tokens: 52764344320 | elapsed time per iteration (s): 1.81 | learning rate: 2.075E-05 | global batch size: 512 | lm loss: 1.153846E+00 | grad norm: 0.178 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.184 | TFLOPs: 42.50 | 31: iteration 50330/ 52452 | consumed samples: 25768960 | consumed tokens: 52774830080 | elapsed time per iteration (s): 1.77 | learning rate: 2.074E-05 | global batch size: 512 | lm loss: 1.152351E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.277 | TFLOPs: 43.42 | 31: iteration 50340/ 52452 | consumed samples: 25774080 | consumed tokens: 52785315840 | elapsed time per iteration (s): 1.86 | learning rate: 2.073E-05 | global batch size: 512 | lm loss: 1.145495E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 275.323 | TFLOPs: 41.32 | 31: iteration 50350/ 52452 | consumed samples: 25779200 | consumed tokens: 52795801600 | elapsed time per iteration (s): 1.72 | learning rate: 2.073E-05 | global batch size: 512 | lm loss: 1.146318E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.042 | TFLOPs: 44.58 | 31: iteration 50360/ 52452 | consumed samples: 25784320 | consumed tokens: 52806287360 | elapsed time per iteration (s): 1.78 | learning rate: 2.072E-05 | global batch size: 512 | lm loss: 1.133848E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.447 | TFLOPs: 43.14 | 31: iteration 50370/ 52452 | consumed samples: 25789440 | consumed tokens: 52816773120 | elapsed time per iteration (s): 1.79 | learning rate: 2.071E-05 | global batch size: 512 | lm loss: 1.143955E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.026 | TFLOPs: 42.93 | 31: iteration 50380/ 52452 | consumed samples: 25794560 | consumed tokens: 52827258880 | elapsed time per iteration (s): 1.76 | learning rate: 2.071E-05 | global batch size: 512 | lm loss: 1.144919E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.439 | TFLOPs: 43.74 | 31: iteration 50390/ 52452 | consumed samples: 25799680 | consumed tokens: 52837744640 | elapsed time per iteration (s): 1.81 | learning rate: 2.070E-05 | global batch size: 512 | lm loss: 1.147949E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.085 | TFLOPs: 42.49 | 31: iteration 50400/ 52452 | consumed samples: 25804800 | consumed tokens: 52848230400 | elapsed time per iteration (s): 1.74 | learning rate: 2.069E-05 | global batch size: 512 | lm loss: 1.131304E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.024 | TFLOPs: 44.28 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50400 | lm loss value: 4.257028E+00 | lm loss PPL: 7.059986E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50400 | lm loss value: 4.086476E+00 | lm loss PPL: 5.952973E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 50410/ 52452 | consumed samples: 25809920 | consumed tokens: 52858716160 | elapsed time per iteration (s): 9.22 | learning rate: 2.069E-05 | global batch size: 512 | lm loss: 1.143072E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.561 | TFLOPs: 8.34 | 31: iteration 50420/ 52452 | consumed samples: 25815040 | consumed tokens: 52869201920 | elapsed time per iteration (s): 1.81 | learning rate: 2.068E-05 | global batch size: 512 | lm loss: 1.144663E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.218 | TFLOPs: 42.36 | 31: iteration 50430/ 52452 | consumed samples: 25820160 | consumed tokens: 52879687680 | elapsed time per iteration (s): 1.81 | learning rate: 2.067E-05 | global batch size: 512 | lm loss: 1.141016E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.653 | TFLOPs: 42.57 | 31: iteration 50440/ 52452 | consumed samples: 25825280 | consumed tokens: 52890173440 | elapsed time per iteration (s): 1.76 | learning rate: 2.067E-05 | global batch size: 512 | lm loss: 1.144108E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.602 | TFLOPs: 43.77 | 31: iteration 50450/ 52452 | consumed samples: 25830400 | consumed tokens: 52900659200 | elapsed time per iteration (s): 2.34 | learning rate: 2.066E-05 | global batch size: 512 | lm loss: 1.147158E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 218.425 | TFLOPs: 32.78 | 31: iteration 50460/ 52452 | consumed samples: 25835520 | consumed tokens: 52911144960 | elapsed time per iteration (s): 1.75 | learning rate: 2.065E-05 | global batch size: 512 | lm loss: 1.148853E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.772 | TFLOPs: 43.79 | 31: iteration 50470/ 52452 | consumed samples: 25840640 | consumed tokens: 52921630720 | elapsed time per iteration (s): 2.24 | learning rate: 2.065E-05 | global batch size: 512 | lm loss: 1.136036E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 228.112 | TFLOPs: 34.24 | 31: iteration 50480/ 52452 | consumed samples: 25845760 | consumed tokens: 52932116480 | elapsed time per iteration (s): 1.75 | learning rate: 2.064E-05 | global batch size: 512 | lm loss: 1.128716E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.987 | TFLOPs: 43.83 | 31: iteration 50490/ 52452 | consumed samples: 25850880 | consumed tokens: 52942602240 | elapsed time per iteration (s): 1.74 | learning rate: 2.063E-05 | global batch size: 512 | lm loss: 1.144862E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.874 | TFLOPs: 44.26 | 31: iteration 50500/ 52452 | consumed samples: 25856000 | consumed tokens: 52953088000 | elapsed time per iteration (s): 1.74 | learning rate: 2.063E-05 | global batch size: 512 | lm loss: 1.134187E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.712 | TFLOPs: 44.08 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50500 | lm loss value: 4.241793E+00 | lm loss PPL: 6.953239E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50500 | lm loss value: 4.064242E+00 | lm loss PPL: 5.822078E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 50510/ 52452 | consumed samples: 25861120 | consumed tokens: 52963573760 | elapsed time per iteration (s): 9.13 | learning rate: 2.062E-05 | global batch size: 512 | lm loss: 1.138294E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.050 | TFLOPs: 8.41 | 31: iteration 50520/ 52452 | consumed samples: 25866240 | consumed tokens: 52974059520 | elapsed time per iteration (s): 1.73 | learning rate: 2.061E-05 | global batch size: 512 | lm loss: 1.149380E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.897 | TFLOPs: 44.41 | 31: iteration 50530/ 52452 | consumed samples: 25871360 | consumed tokens: 52984545280 | elapsed time per iteration (s): 1.73 | learning rate: 2.061E-05 | global batch size: 512 | lm loss: 1.129843E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.327 | TFLOPs: 44.33 | 31: iteration 50540/ 52452 | consumed samples: 25876480 | consumed tokens: 52995031040 | elapsed time per iteration (s): 1.74 | learning rate: 2.060E-05 | global batch size: 512 | lm loss: 1.140357E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.919 | TFLOPs: 44.27 | 31: iteration 50550/ 52452 | consumed samples: 25881600 | consumed tokens: 53005516800 | elapsed time per iteration (s): 1.72 | learning rate: 2.060E-05 | global batch size: 512 | lm loss: 1.130640E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.261 | TFLOPs: 44.62 | 31: iteration 50560/ 52452 | consumed samples: 25886720 | consumed tokens: 53016002560 | elapsed time per iteration (s): 1.73 | learning rate: 2.059E-05 | global batch size: 512 | lm loss: 1.143941E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.390 | TFLOPs: 44.49 | 31: iteration 50570/ 52452 | consumed samples: 25891840 | consumed tokens: 53026488320 | elapsed time per iteration (s): 1.74 | learning rate: 2.058E-05 | global batch size: 512 | lm loss: 1.152367E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.728 | TFLOPs: 44.09 | 31: iteration 50580/ 52452 | consumed samples: 25896960 | consumed tokens: 53036974080 | elapsed time per iteration (s): 1.78 | learning rate: 2.058E-05 | global batch size: 512 | lm loss: 1.134451E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.367 | TFLOPs: 43.28 | 31: iteration 50590/ 52452 | consumed samples: 25902080 | consumed tokens: 53047459840 | elapsed time per iteration (s): 1.73 | learning rate: 2.057E-05 | global batch size: 512 | lm loss: 1.135572E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.190 | TFLOPs: 44.31 | 31: iteration 50600/ 52452 | consumed samples: 25907200 | consumed tokens: 53057945600 | elapsed time per iteration (s): 1.76 | learning rate: 2.056E-05 | global batch size: 512 | lm loss: 1.128542E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.438 | TFLOPs: 43.74 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50600 | lm loss value: 4.198772E+00 | lm loss PPL: 6.660449E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50600 | lm loss value: 4.060925E+00 | lm loss PPL: 5.802796E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 50610/ 52452 | consumed samples: 25912320 | consumed tokens: 53068431360 | elapsed time per iteration (s): 9.11 | learning rate: 2.056E-05 | global batch size: 512 | lm loss: 1.152775E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.223 | TFLOPs: 8.44 | 31: iteration 50620/ 52452 | consumed samples: 25917440 | consumed tokens: 53078917120 | elapsed time per iteration (s): 1.73 | learning rate: 2.055E-05 | global batch size: 512 | lm loss: 1.133977E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.212 | TFLOPs: 44.46 | 31: iteration 50630/ 52452 | consumed samples: 25922560 | consumed tokens: 53089402880 | elapsed time per iteration (s): 1.75 | learning rate: 2.055E-05 | global batch size: 512 | lm loss: 1.129483E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.937 | TFLOPs: 43.97 | 31: iteration 50640/ 52452 | consumed samples: 25927680 | consumed tokens: 53099888640 | elapsed time per iteration (s): 1.76 | learning rate: 2.054E-05 | global batch size: 512 | lm loss: 1.135981E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.769 | TFLOPs: 43.64 | 31: iteration 50650/ 52452 | consumed samples: 25932800 | consumed tokens: 53110374400 | elapsed time per iteration (s): 1.76 | learning rate: 2.053E-05 | global batch size: 512 | lm loss: 1.127783E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.212 | TFLOPs: 43.71 | 31: iteration 50660/ 52452 | consumed samples: 25937920 | consumed tokens: 53120860160 | elapsed time per iteration (s): 1.74 | learning rate: 2.053E-05 | global batch size: 512 | lm loss: 1.146911E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.941 | TFLOPs: 44.27 | 31: iteration 50670/ 52452 | consumed samples: 25943040 | consumed tokens: 53131345920 | elapsed time per iteration (s): 1.72 | learning rate: 2.052E-05 | global batch size: 512 | lm loss: 1.126536E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 297.456 | TFLOPs: 44.65 | 31: iteration 50680/ 52452 | consumed samples: 25948160 | consumed tokens: 53141831680 | elapsed time per iteration (s): 1.73 | learning rate: 2.052E-05 | global batch size: 512 | lm loss: 1.139807E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.259 | TFLOPs: 44.32 | 31: iteration 50690/ 52452 | consumed samples: 25953280 | consumed tokens: 53152317440 | elapsed time per iteration (s): 1.73 | learning rate: 2.051E-05 | global batch size: 512 | lm loss: 1.149255E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.261 | TFLOPs: 44.32 | 31: iteration 50700/ 52452 | consumed samples: 25958400 | consumed tokens: 53162803200 | elapsed time per iteration (s): 1.74 | learning rate: 2.051E-05 | global batch size: 512 | lm loss: 1.140402E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.904 | TFLOPs: 44.26 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50700 | lm loss value: 4.213977E+00 | lm loss PPL: 6.762497E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50700 | lm loss value: 4.056311E+00 | lm loss PPL: 5.776082E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 50710/ 52452 | consumed samples: 25963520 | consumed tokens: 53173288960 | elapsed time per iteration (s): 9.15 | learning rate: 2.050E-05 | global batch size: 512 | lm loss: 1.144182E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.977 | TFLOPs: 8.40 | 31: iteration 50720/ 52452 | consumed samples: 25968640 | consumed tokens: 53183774720 | elapsed time per iteration (s): 1.81 | learning rate: 2.049E-05 | global batch size: 512 | lm loss: 1.143274E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.457 | TFLOPs: 42.55 | 31: iteration 50730/ 52452 | consumed samples: 25973760 | consumed tokens: 53194260480 | elapsed time per iteration (s): 1.76 | learning rate: 2.049E-05 | global batch size: 512 | lm loss: 1.144227E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.853 | TFLOPs: 43.66 | 31: iteration 50740/ 52452 | consumed samples: 25978880 | consumed tokens: 53204746240 | elapsed time per iteration (s): 1.76 | learning rate: 2.048E-05 | global batch size: 512 | lm loss: 1.155182E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.271 | TFLOPs: 43.57 | 31: iteration 50750/ 52452 | consumed samples: 25984000 | consumed tokens: 53215232000 | elapsed time per iteration (s): 1.85 | learning rate: 2.048E-05 | global batch size: 512 | lm loss: 1.138566E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.479 | TFLOPs: 41.50 | 31: iteration 50760/ 52452 | consumed samples: 25989120 | consumed tokens: 53225717760 | elapsed time per iteration (s): 1.77 | learning rate: 2.047E-05 | global batch size: 512 | lm loss: 1.128673E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.007 | TFLOPs: 43.38 | 31: iteration 50770/ 52452 | consumed samples: 25994240 | consumed tokens: 53236203520 | elapsed time per iteration (s): 1.75 | learning rate: 2.047E-05 | global batch size: 512 | lm loss: 1.124257E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.436 | TFLOPs: 43.89 | 31: iteration 50780/ 52452 | consumed samples: 25999360 | consumed tokens: 53246689280 | elapsed time per iteration (s): 1.85 | learning rate: 2.046E-05 | global batch size: 512 | lm loss: 1.134160E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.134 | TFLOPs: 41.60 | 31: iteration 50790/ 52452 | consumed samples: 26004480 | consumed tokens: 53257175040 | elapsed time per iteration (s): 1.94 | learning rate: 2.045E-05 | global batch size: 512 | lm loss: 1.152352E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 263.594 | TFLOPs: 39.56 | 31: iteration 50800/ 52452 | consumed samples: 26009600 | consumed tokens: 53267660800 | elapsed time per iteration (s): 1.82 | learning rate: 2.045E-05 | global batch size: 512 | lm loss: 1.144741E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.906 | TFLOPs: 42.16 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50800 | lm loss value: 4.230461E+00 | lm loss PPL: 6.874889E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50800 | lm loss value: 4.082810E+00 | lm loss PPL: 5.931193E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 50810/ 52452 | consumed samples: 26014720 | consumed tokens: 53278146560 | elapsed time per iteration (s): 9.17 | learning rate: 2.044E-05 | global batch size: 512 | lm loss: 1.132676E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.849 | TFLOPs: 8.38 | 31: iteration 50820/ 52452 | consumed samples: 26019840 | consumed tokens: 53288632320 | elapsed time per iteration (s): 1.83 | learning rate: 2.044E-05 | global batch size: 512 | lm loss: 1.138549E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.482 | TFLOPs: 41.95 | 31: iteration 50830/ 52452 | consumed samples: 26024960 | consumed tokens: 53299118080 | elapsed time per iteration (s): 1.77 | learning rate: 2.043E-05 | global batch size: 512 | lm loss: 1.138674E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.715 | TFLOPs: 43.33 | 31: iteration 50840/ 52452 | consumed samples: 26030080 | consumed tokens: 53309603840 | elapsed time per iteration (s): 1.85 | learning rate: 2.043E-05 | global batch size: 512 | lm loss: 1.152627E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.632 | TFLOPs: 41.52 | 31: iteration 50850/ 52452 | consumed samples: 26035200 | consumed tokens: 53320089600 | elapsed time per iteration (s): 1.83 | learning rate: 2.042E-05 | global batch size: 512 | lm loss: 1.141095E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.465 | TFLOPs: 41.95 | 31: iteration 50860/ 52452 | consumed samples: 26040320 | consumed tokens: 53330575360 | elapsed time per iteration (s): 1.77 | learning rate: 2.042E-05 | global batch size: 512 | lm loss: 1.129795E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.215 | TFLOPs: 43.41 | 31: iteration 50870/ 52452 | consumed samples: 26045440 | consumed tokens: 53341061120 | elapsed time per iteration (s): 1.74 | learning rate: 2.041E-05 | global batch size: 512 | lm loss: 1.136860E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.546 | TFLOPs: 44.21 | 31: iteration 50880/ 52452 | consumed samples: 26050560 | consumed tokens: 53351546880 | elapsed time per iteration (s): 1.76 | learning rate: 2.041E-05 | global batch size: 512 | lm loss: 1.142352E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.925 | TFLOPs: 43.67 | 31: iteration 50890/ 52452 | consumed samples: 26055680 | consumed tokens: 53362032640 | elapsed time per iteration (s): 1.77 | learning rate: 2.040E-05 | global batch size: 512 | lm loss: 1.140114E+00 | grad norm: 0.168 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.451 | TFLOPs: 43.29 | 31: iteration 50900/ 52452 | consumed samples: 26060800 | consumed tokens: 53372518400 | elapsed time per iteration (s): 1.84 | learning rate: 2.040E-05 | global batch size: 512 | lm loss: 1.136164E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.654 | TFLOPs: 41.82 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 50900 | lm loss value: 4.196297E+00 | lm loss PPL: 6.643983E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 50900 | lm loss value: 4.058677E+00 | lm loss PPL: 5.789767E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 50910/ 52452 | consumed samples: 26065920 | consumed tokens: 53383004160 | elapsed time per iteration (s): 9.12 | learning rate: 2.039E-05 | global batch size: 512 | lm loss: 1.134374E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.158 | TFLOPs: 8.43 | 31: iteration 50920/ 52452 | consumed samples: 26071040 | consumed tokens: 53393489920 | elapsed time per iteration (s): 1.73 | learning rate: 2.039E-05 | global batch size: 512 | lm loss: 1.139166E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.988 | TFLOPs: 44.43 | 31: iteration 50930/ 52452 | consumed samples: 26076160 | consumed tokens: 53403975680 | elapsed time per iteration (s): 1.84 | learning rate: 2.038E-05 | global batch size: 512 | lm loss: 1.137054E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.297 | TFLOPs: 41.77 | 31: iteration 50940/ 52452 | consumed samples: 26081280 | consumed tokens: 53414461440 | elapsed time per iteration (s): 1.90 | learning rate: 2.038E-05 | global batch size: 512 | lm loss: 1.144360E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.229 | TFLOPs: 40.41 | 31: iteration 50950/ 52452 | consumed samples: 26086400 | consumed tokens: 53424947200 | elapsed time per iteration (s): 1.87 | learning rate: 2.037E-05 | global batch size: 512 | lm loss: 1.146570E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.338 | TFLOPs: 41.18 | 31: iteration 50960/ 52452 | consumed samples: 26091520 | consumed tokens: 53435432960 | elapsed time per iteration (s): 1.80 | learning rate: 2.037E-05 | global batch size: 512 | lm loss: 1.141474E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.113 | TFLOPs: 42.64 | 31: iteration 50970/ 52452 | consumed samples: 26096640 | consumed tokens: 53445918720 | elapsed time per iteration (s): 1.87 | learning rate: 2.036E-05 | global batch size: 512 | lm loss: 1.131424E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.531 | TFLOPs: 41.06 | 31: iteration 50980/ 52452 | consumed samples: 26101760 | consumed tokens: 53456404480 | elapsed time per iteration (s): 1.74 | learning rate: 2.036E-05 | global batch size: 512 | lm loss: 1.143583E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.218 | TFLOPs: 44.16 | 31: iteration 50990/ 52452 | consumed samples: 26106880 | consumed tokens: 53466890240 | elapsed time per iteration (s): 1.80 | learning rate: 2.035E-05 | global batch size: 512 | lm loss: 1.135036E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.349 | TFLOPs: 42.68 | 31: iteration 51000/ 52452 | consumed samples: 26112000 | consumed tokens: 53477376000 | elapsed time per iteration (s): 1.82 | learning rate: 2.035E-05 | global batch size: 512 | lm loss: 1.141542E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.951 | TFLOPs: 42.17 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51000 | lm loss value: 4.212446E+00 | lm loss PPL: 6.752148E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51000 | lm loss value: 4.065274E+00 | lm loss PPL: 5.828086E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51010/ 52452 | consumed samples: 26117120 | consumed tokens: 53487861760 | elapsed time per iteration (s): 9.17 | learning rate: 2.034E-05 | global batch size: 512 | lm loss: 1.132893E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.815 | TFLOPs: 8.38 | 31: iteration 51020/ 52452 | consumed samples: 26122240 | consumed tokens: 53498347520 | elapsed time per iteration (s): 1.84 | learning rate: 2.034E-05 | global batch size: 512 | lm loss: 1.150548E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 278.427 | TFLOPs: 41.79 | 31: iteration 51030/ 52452 | consumed samples: 26127360 | consumed tokens: 53508833280 | elapsed time per iteration (s): 1.81 | learning rate: 2.033E-05 | global batch size: 512 | lm loss: 1.141532E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.505 | TFLOPs: 42.55 | 31: iteration 51040/ 52452 | consumed samples: 26132480 | consumed tokens: 53519319040 | elapsed time per iteration (s): 1.80 | learning rate: 2.033E-05 | global batch size: 512 | lm loss: 1.142364E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.593 | TFLOPs: 42.72 | 31: iteration 51050/ 52452 | consumed samples: 26137600 | consumed tokens: 53529804800 | elapsed time per iteration (s): 1.77 | learning rate: 2.032E-05 | global batch size: 512 | lm loss: 1.141643E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.930 | TFLOPs: 43.52 | 31: iteration 51060/ 52452 | consumed samples: 26142720 | consumed tokens: 53540290560 | elapsed time per iteration (s): 1.83 | learning rate: 2.032E-05 | global batch size: 512 | lm loss: 1.125175E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.375 | TFLOPs: 42.08 | 31: iteration 51070/ 52452 | consumed samples: 26147840 | consumed tokens: 53550776320 | elapsed time per iteration (s): 1.81 | learning rate: 2.031E-05 | global batch size: 512 | lm loss: 1.157956E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.844 | TFLOPs: 42.45 | 31: iteration 51080/ 52452 | consumed samples: 26152960 | consumed tokens: 53561262080 | elapsed time per iteration (s): 1.73 | learning rate: 2.031E-05 | global batch size: 512 | lm loss: 1.152290E+00 | grad norm: 0.172 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.865 | TFLOPs: 44.41 | 31: iteration 51090/ 52452 | consumed samples: 26158080 | consumed tokens: 53571747840 | elapsed time per iteration (s): 1.81 | learning rate: 2.031E-05 | global batch size: 512 | lm loss: 1.154225E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.569 | TFLOPs: 42.41 | 31: iteration 51100/ 52452 | consumed samples: 26163200 | consumed tokens: 53582233600 | elapsed time per iteration (s): 1.76 | learning rate: 2.030E-05 | global batch size: 512 | lm loss: 1.152071E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.694 | TFLOPs: 43.78 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51100 | lm loss value: 4.118897E+00 | lm loss PPL: 6.149141E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51100 | lm loss value: 4.004580E+00 | lm loss PPL: 5.484876E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51110/ 52452 | consumed samples: 26168320 | consumed tokens: 53592719360 | elapsed time per iteration (s): 9.28 | learning rate: 2.030E-05 | global batch size: 512 | lm loss: 1.140406E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.165 | TFLOPs: 8.28 | 31: iteration 51120/ 52452 | consumed samples: 26173440 | consumed tokens: 53603205120 | elapsed time per iteration (s): 1.87 | learning rate: 2.029E-05 | global batch size: 512 | lm loss: 1.140876E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.926 | TFLOPs: 41.11 | 31: iteration 51130/ 52452 | consumed samples: 26178560 | consumed tokens: 53613690880 | elapsed time per iteration (s): 1.81 | learning rate: 2.029E-05 | global batch size: 512 | lm loss: 1.144592E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.888 | TFLOPs: 42.46 | 31: iteration 51140/ 52452 | consumed samples: 26183680 | consumed tokens: 53624176640 | elapsed time per iteration (s): 1.77 | learning rate: 2.028E-05 | global batch size: 512 | lm loss: 1.142812E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.039 | TFLOPs: 43.38 | 31: iteration 51150/ 52452 | consumed samples: 26188800 | consumed tokens: 53634662400 | elapsed time per iteration (s): 1.77 | learning rate: 2.028E-05 | global batch size: 512 | lm loss: 1.138258E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.945 | TFLOPs: 43.37 | 31: iteration 51160/ 52452 | consumed samples: 26193920 | consumed tokens: 53645148160 | elapsed time per iteration (s): 1.81 | learning rate: 2.027E-05 | global batch size: 512 | lm loss: 1.145418E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.809 | TFLOPs: 42.45 | 31: iteration 51170/ 52452 | consumed samples: 26199040 | consumed tokens: 53655633920 | elapsed time per iteration (s): 1.73 | learning rate: 2.027E-05 | global batch size: 512 | lm loss: 1.154812E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.252 | TFLOPs: 44.47 | 31: iteration 51180/ 52452 | consumed samples: 26204160 | consumed tokens: 53666119680 | elapsed time per iteration (s): 1.75 | learning rate: 2.027E-05 | global batch size: 512 | lm loss: 1.138122E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.040 | TFLOPs: 43.98 | 31: iteration 51190/ 52452 | consumed samples: 26209280 | consumed tokens: 53676605440 | elapsed time per iteration (s): 1.77 | learning rate: 2.026E-05 | global batch size: 512 | lm loss: 1.140710E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.606 | TFLOPs: 43.47 | 31: iteration 51200/ 52452 | consumed samples: 26214400 | consumed tokens: 53687091200 | elapsed time per iteration (s): 1.74 | learning rate: 2.026E-05 | global batch size: 512 | lm loss: 1.140596E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.751 | TFLOPs: 44.24 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51200 | lm loss value: 4.148318E+00 | lm loss PPL: 6.332738E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51200 | lm loss value: 4.027248E+00 | lm loss PPL: 5.610629E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51210/ 52452 | consumed samples: 26219520 | consumed tokens: 53697576960 | elapsed time per iteration (s): 9.17 | learning rate: 2.025E-05 | global batch size: 512 | lm loss: 1.133734E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.847 | TFLOPs: 8.38 | 31: iteration 51220/ 52452 | consumed samples: 26224640 | consumed tokens: 53708062720 | elapsed time per iteration (s): 1.83 | learning rate: 2.025E-05 | global batch size: 512 | lm loss: 1.144934E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.456 | TFLOPs: 42.09 | 31: iteration 51230/ 52452 | consumed samples: 26229760 | consumed tokens: 53718548480 | elapsed time per iteration (s): 1.73 | learning rate: 2.025E-05 | global batch size: 512 | lm loss: 1.124870E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.469 | TFLOPs: 44.50 | 31: iteration 51240/ 52452 | consumed samples: 26234880 | consumed tokens: 53729034240 | elapsed time per iteration (s): 1.74 | learning rate: 2.024E-05 | global batch size: 512 | lm loss: 1.135110E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.963 | TFLOPs: 44.27 | 31: iteration 51250/ 52452 | consumed samples: 26240000 | consumed tokens: 53739520000 | elapsed time per iteration (s): 1.73 | learning rate: 2.024E-05 | global batch size: 512 | lm loss: 1.129524E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.439 | TFLOPs: 44.49 | 31: iteration 51260/ 52452 | consumed samples: 26245120 | consumed tokens: 53750005760 | elapsed time per iteration (s): 1.80 | learning rate: 2.023E-05 | global batch size: 512 | lm loss: 1.151138E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.881 | TFLOPs: 42.61 | 31: iteration 51270/ 52452 | consumed samples: 26250240 | consumed tokens: 53760491520 | elapsed time per iteration (s): 1.73 | learning rate: 2.023E-05 | global batch size: 512 | lm loss: 1.129169E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.818 | TFLOPs: 44.40 | 31: iteration 51280/ 52452 | consumed samples: 26255360 | consumed tokens: 53770977280 | elapsed time per iteration (s): 1.74 | learning rate: 2.023E-05 | global batch size: 512 | lm loss: 1.146608E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.949 | TFLOPs: 44.12 | 31: iteration 51290/ 52452 | consumed samples: 26260480 | consumed tokens: 53781463040 | elapsed time per iteration (s): 1.75 | learning rate: 2.022E-05 | global batch size: 512 | lm loss: 1.139923E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.011 | TFLOPs: 43.83 | 31: iteration 51300/ 52452 | consumed samples: 26265600 | consumed tokens: 53791948800 | elapsed time per iteration (s): 1.74 | learning rate: 2.022E-05 | global batch size: 512 | lm loss: 1.142536E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.083 | TFLOPs: 44.29 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51300 | lm loss value: 4.135766E+00 | lm loss PPL: 6.253748E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51300 | lm loss value: 4.014435E+00 | lm loss PPL: 5.539198E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51310/ 52452 | consumed samples: 26270720 | consumed tokens: 53802434560 | elapsed time per iteration (s): 9.14 | learning rate: 2.021E-05 | global batch size: 512 | lm loss: 1.142068E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.037 | TFLOPs: 8.41 | 31: iteration 51320/ 52452 | consumed samples: 26275840 | consumed tokens: 53812920320 | elapsed time per iteration (s): 1.76 | learning rate: 2.021E-05 | global batch size: 512 | lm loss: 1.132897E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.400 | TFLOPs: 43.74 | 31: iteration 51330/ 52452 | consumed samples: 26280960 | consumed tokens: 53823406080 | elapsed time per iteration (s): 1.76 | learning rate: 2.021E-05 | global batch size: 512 | lm loss: 1.128549E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.493 | TFLOPs: 43.60 | 31: iteration 51340/ 52452 | consumed samples: 26286080 | consumed tokens: 53833891840 | elapsed time per iteration (s): 1.80 | learning rate: 2.020E-05 | global batch size: 512 | lm loss: 1.130424E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.288 | TFLOPs: 42.67 | 31: iteration 51350/ 52452 | consumed samples: 26291200 | consumed tokens: 53844377600 | elapsed time per iteration (s): 1.74 | learning rate: 2.020E-05 | global batch size: 512 | lm loss: 1.136218E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.124 | TFLOPs: 44.15 | 31: iteration 51360/ 52452 | consumed samples: 26296320 | consumed tokens: 53854863360 | elapsed time per iteration (s): 1.78 | learning rate: 2.020E-05 | global batch size: 512 | lm loss: 1.150361E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.047 | TFLOPs: 43.08 | 31: iteration 51370/ 52452 | consumed samples: 26301440 | consumed tokens: 53865349120 | elapsed time per iteration (s): 1.78 | learning rate: 2.019E-05 | global batch size: 512 | lm loss: 1.138001E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.326 | TFLOPs: 43.13 | 31: iteration 51380/ 52452 | consumed samples: 26306560 | consumed tokens: 53875834880 | elapsed time per iteration (s): 1.77 | learning rate: 2.019E-05 | global batch size: 512 | lm loss: 1.137901E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.531 | TFLOPs: 43.46 | 31: iteration 51390/ 52452 | consumed samples: 26311680 | consumed tokens: 53886320640 | elapsed time per iteration (s): 1.75 | learning rate: 2.019E-05 | global batch size: 512 | lm loss: 1.133992E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.395 | TFLOPs: 44.04 | 31: iteration 51400/ 52452 | consumed samples: 26316800 | consumed tokens: 53896806400 | elapsed time per iteration (s): 1.80 | learning rate: 2.018E-05 | global batch size: 512 | lm loss: 1.132440E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.055 | TFLOPs: 42.79 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51400 | lm loss value: 4.213967E+00 | lm loss PPL: 6.762430E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51400 | lm loss value: 4.056787E+00 | lm loss PPL: 5.778834E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51410/ 52452 | consumed samples: 26321920 | consumed tokens: 53907292160 | elapsed time per iteration (s): 9.29 | learning rate: 2.018E-05 | global batch size: 512 | lm loss: 1.138224E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.109 | TFLOPs: 8.27 | 31: iteration 51420/ 52452 | consumed samples: 26327040 | consumed tokens: 53917777920 | elapsed time per iteration (s): 1.76 | learning rate: 2.018E-05 | global batch size: 512 | lm loss: 1.142845E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.552 | TFLOPs: 43.76 | 31: iteration 51430/ 52452 | consumed samples: 26332160 | consumed tokens: 53928263680 | elapsed time per iteration (s): 1.83 | learning rate: 2.017E-05 | global batch size: 512 | lm loss: 1.128554E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.284 | TFLOPs: 41.92 | 31: iteration 51440/ 52452 | consumed samples: 26337280 | consumed tokens: 53938749440 | elapsed time per iteration (s): 1.75 | learning rate: 2.017E-05 | global batch size: 512 | lm loss: 1.125139E+00 | grad norm: 0.154 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.015 | TFLOPs: 43.83 | 31: iteration 51450/ 52452 | consumed samples: 26342400 | consumed tokens: 53949235200 | elapsed time per iteration (s): 1.81 | learning rate: 2.017E-05 | global batch size: 512 | lm loss: 1.132698E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.734 | TFLOPs: 42.44 | 31: iteration 51460/ 52452 | consumed samples: 26347520 | consumed tokens: 53959720960 | elapsed time per iteration (s): 1.78 | learning rate: 2.016E-05 | global batch size: 512 | lm loss: 1.142873E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.790 | TFLOPs: 43.20 | 31: iteration 51470/ 52452 | consumed samples: 26352640 | consumed tokens: 53970206720 | elapsed time per iteration (s): 1.80 | learning rate: 2.016E-05 | global batch size: 512 | lm loss: 1.130088E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 283.726 | TFLOPs: 42.59 | 31: iteration 51480/ 52452 | consumed samples: 26357760 | consumed tokens: 53980692480 | elapsed time per iteration (s): 1.73 | learning rate: 2.016E-05 | global batch size: 512 | lm loss: 1.148184E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.951 | TFLOPs: 44.42 | 31: iteration 51490/ 52452 | consumed samples: 26362880 | consumed tokens: 53991178240 | elapsed time per iteration (s): 1.73 | learning rate: 2.015E-05 | global batch size: 512 | lm loss: 1.145422E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 296.507 | TFLOPs: 44.50 | 31: iteration 51500/ 52452 | consumed samples: 26368000 | consumed tokens: 54001664000 | elapsed time per iteration (s): 1.78 | learning rate: 2.015E-05 | global batch size: 512 | lm loss: 1.139374E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.927 | TFLOPs: 43.22 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51500 | lm loss value: 4.196299E+00 | lm loss PPL: 6.643995E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51500 | lm loss value: 4.057647E+00 | lm loss PPL: 5.783804E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51510/ 52452 | consumed samples: 26373120 | consumed tokens: 54012149760 | elapsed time per iteration (s): 9.13 | learning rate: 2.015E-05 | global batch size: 512 | lm loss: 1.137030E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 56.085 | TFLOPs: 8.42 | 31: iteration 51520/ 52452 | consumed samples: 26378240 | consumed tokens: 54022635520 | elapsed time per iteration (s): 1.77 | learning rate: 2.014E-05 | global batch size: 512 | lm loss: 1.137393E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.678 | TFLOPs: 43.48 | 31: iteration 51530/ 52452 | consumed samples: 26383360 | consumed tokens: 54033121280 | elapsed time per iteration (s): 1.79 | learning rate: 2.014E-05 | global batch size: 512 | lm loss: 1.143840E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.988 | TFLOPs: 42.93 | 31: iteration 51540/ 52452 | consumed samples: 26388480 | consumed tokens: 54043607040 | elapsed time per iteration (s): 1.75 | learning rate: 2.014E-05 | global batch size: 512 | lm loss: 1.135242E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.994 | TFLOPs: 43.83 | 31: iteration 51550/ 52452 | consumed samples: 26393600 | consumed tokens: 54054092800 | elapsed time per iteration (s): 1.81 | learning rate: 2.013E-05 | global batch size: 512 | lm loss: 1.127912E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.183 | TFLOPs: 42.35 | 31: iteration 51560/ 52452 | consumed samples: 26398720 | consumed tokens: 54064578560 | elapsed time per iteration (s): 1.75 | learning rate: 2.013E-05 | global batch size: 512 | lm loss: 1.139214E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.662 | TFLOPs: 43.93 | 31: iteration 51570/ 52452 | consumed samples: 26403840 | consumed tokens: 54075064320 | elapsed time per iteration (s): 1.89 | learning rate: 2.013E-05 | global batch size: 512 | lm loss: 1.150357E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.308 | TFLOPs: 40.57 | 31: iteration 51580/ 52452 | consumed samples: 26408960 | consumed tokens: 54085550080 | elapsed time per iteration (s): 1.75 | learning rate: 2.013E-05 | global batch size: 512 | lm loss: 1.133281E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.902 | TFLOPs: 43.81 | 31: iteration 51590/ 52452 | consumed samples: 26414080 | consumed tokens: 54096035840 | elapsed time per iteration (s): 1.77 | learning rate: 2.012E-05 | global batch size: 512 | lm loss: 1.131085E+00 | grad norm: 0.170 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.706 | TFLOPs: 43.33 | 31: iteration 51600/ 52452 | consumed samples: 26419200 | consumed tokens: 54106521600 | elapsed time per iteration (s): 1.82 | learning rate: 2.012E-05 | global batch size: 512 | lm loss: 1.144907E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.081 | TFLOPs: 42.34 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51600 | lm loss value: 4.180141E+00 | lm loss PPL: 6.537510E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51600 | lm loss value: 4.037355E+00 | lm loss PPL: 5.667626E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51610/ 52452 | consumed samples: 26424320 | consumed tokens: 54117007360 | elapsed time per iteration (s): 9.17 | learning rate: 2.012E-05 | global batch size: 512 | lm loss: 1.140241E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.807 | TFLOPs: 8.38 | 31: iteration 51620/ 52452 | consumed samples: 26429440 | consumed tokens: 54127493120 | elapsed time per iteration (s): 1.77 | learning rate: 2.011E-05 | global batch size: 512 | lm loss: 1.136907E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.696 | TFLOPs: 43.48 | 31: iteration 51630/ 52452 | consumed samples: 26434560 | consumed tokens: 54137978880 | elapsed time per iteration (s): 1.83 | learning rate: 2.011E-05 | global batch size: 512 | lm loss: 1.132879E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.512 | TFLOPs: 42.10 | 31: iteration 51640/ 52452 | consumed samples: 26439680 | consumed tokens: 54148464640 | elapsed time per iteration (s): 1.82 | learning rate: 2.011E-05 | global batch size: 512 | lm loss: 1.143541E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.982 | TFLOPs: 42.17 | 31: iteration 51650/ 52452 | consumed samples: 26444800 | consumed tokens: 54158950400 | elapsed time per iteration (s): 1.82 | learning rate: 2.011E-05 | global batch size: 512 | lm loss: 1.132939E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.519 | TFLOPs: 42.25 | 31: iteration 51660/ 52452 | consumed samples: 26449920 | consumed tokens: 54169436160 | elapsed time per iteration (s): 1.76 | learning rate: 2.010E-05 | global batch size: 512 | lm loss: 1.132359E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.346 | TFLOPs: 43.58 | 31: iteration 51670/ 52452 | consumed samples: 26455040 | consumed tokens: 54179921920 | elapsed time per iteration (s): 1.83 | learning rate: 2.010E-05 | global batch size: 512 | lm loss: 1.135820E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.209 | TFLOPs: 42.06 | 31: iteration 51680/ 52452 | consumed samples: 26460160 | consumed tokens: 54190407680 | elapsed time per iteration (s): 1.76 | learning rate: 2.010E-05 | global batch size: 512 | lm loss: 1.130955E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.361 | TFLOPs: 43.58 | 31: iteration 51690/ 52452 | consumed samples: 26465280 | consumed tokens: 54200893440 | elapsed time per iteration (s): 1.82 | learning rate: 2.010E-05 | global batch size: 512 | lm loss: 1.129881E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.155 | TFLOPs: 42.20 | 31: iteration 51700/ 52452 | consumed samples: 26470400 | consumed tokens: 54211379200 | elapsed time per iteration (s): 1.87 | learning rate: 2.009E-05 | global batch size: 512 | lm loss: 1.138429E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 273.097 | TFLOPs: 40.99 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51700 | lm loss value: 4.161290E+00 | lm loss PPL: 6.415424E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51700 | lm loss value: 4.024209E+00 | lm loss PPL: 5.593607E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51710/ 52452 | consumed samples: 26475520 | consumed tokens: 54221864960 | elapsed time per iteration (s): 9.31 | learning rate: 2.009E-05 | global batch size: 512 | lm loss: 1.132002E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.019 | TFLOPs: 8.26 | 31: iteration 51720/ 52452 | consumed samples: 26480640 | consumed tokens: 54232350720 | elapsed time per iteration (s): 1.77 | learning rate: 2.009E-05 | global batch size: 512 | lm loss: 1.145855E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.074 | TFLOPs: 43.54 | 31: iteration 51730/ 52452 | consumed samples: 26485760 | consumed tokens: 54242836480 | elapsed time per iteration (s): 1.78 | learning rate: 2.009E-05 | global batch size: 512 | lm loss: 1.120024E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.973 | TFLOPs: 43.22 | 31: iteration 51740/ 52452 | consumed samples: 26490880 | consumed tokens: 54253322240 | elapsed time per iteration (s): 1.76 | learning rate: 2.008E-05 | global batch size: 512 | lm loss: 1.139867E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.102 | TFLOPs: 43.69 | 31: iteration 51750/ 52452 | consumed samples: 26496000 | consumed tokens: 54263808000 | elapsed time per iteration (s): 1.78 | learning rate: 2.008E-05 | global batch size: 512 | lm loss: 1.136902E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.293 | TFLOPs: 43.27 | 31: iteration 51760/ 52452 | consumed samples: 26501120 | consumed tokens: 54274293760 | elapsed time per iteration (s): 1.76 | learning rate: 2.008E-05 | global batch size: 512 | lm loss: 1.129107E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.927 | TFLOPs: 43.67 | 31: iteration 51770/ 52452 | consumed samples: 26506240 | consumed tokens: 54284779520 | elapsed time per iteration (s): 1.76 | learning rate: 2.008E-05 | global batch size: 512 | lm loss: 1.155630E+00 | grad norm: 0.174 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.417 | TFLOPs: 43.74 | 31: iteration 51780/ 52452 | consumed samples: 26511360 | consumed tokens: 54295265280 | elapsed time per iteration (s): 1.83 | learning rate: 2.007E-05 | global batch size: 512 | lm loss: 1.123743E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.519 | TFLOPs: 41.95 | 31: iteration 51790/ 52452 | consumed samples: 26516480 | consumed tokens: 54305751040 | elapsed time per iteration (s): 1.80 | learning rate: 2.007E-05 | global batch size: 512 | lm loss: 1.139484E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.658 | TFLOPs: 42.73 | 31: iteration 51800/ 52452 | consumed samples: 26521600 | consumed tokens: 54316236800 | elapsed time per iteration (s): 1.79 | learning rate: 2.007E-05 | global batch size: 512 | lm loss: 1.125237E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.686 | TFLOPs: 42.88 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51800 | lm loss value: 4.211836E+00 | lm loss PPL: 6.748034E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51800 | lm loss value: 4.059824E+00 | lm loss PPL: 5.796414E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51810/ 52452 | consumed samples: 26526720 | consumed tokens: 54326722560 | elapsed time per iteration (s): 9.22 | learning rate: 2.007E-05 | global batch size: 512 | lm loss: 1.128177E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.506 | TFLOPs: 8.33 | 31: iteration 51820/ 52452 | consumed samples: 26531840 | consumed tokens: 54337208320 | elapsed time per iteration (s): 1.95 | learning rate: 2.007E-05 | global batch size: 512 | lm loss: 1.135586E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 261.999 | TFLOPs: 39.32 | 31: iteration 51830/ 52452 | consumed samples: 26536960 | consumed tokens: 54347694080 | elapsed time per iteration (s): 1.82 | learning rate: 2.006E-05 | global batch size: 512 | lm loss: 1.144974E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.950 | TFLOPs: 42.17 | 31: iteration 51840/ 52452 | consumed samples: 26542080 | consumed tokens: 54358179840 | elapsed time per iteration (s): 1.80 | learning rate: 2.006E-05 | global batch size: 512 | lm loss: 1.145911E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.676 | TFLOPs: 42.73 | 31: iteration 51850/ 52452 | consumed samples: 26547200 | consumed tokens: 54368665600 | elapsed time per iteration (s): 1.88 | learning rate: 2.006E-05 | global batch size: 512 | lm loss: 1.130218E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 272.005 | TFLOPs: 40.83 | 31: iteration 51860/ 52452 | consumed samples: 26552320 | consumed tokens: 54379151360 | elapsed time per iteration (s): 1.86 | learning rate: 2.006E-05 | global batch size: 512 | lm loss: 1.141030E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.644 | TFLOPs: 41.22 | 31: iteration 51870/ 52452 | consumed samples: 26557440 | consumed tokens: 54389637120 | elapsed time per iteration (s): 1.80 | learning rate: 2.006E-05 | global batch size: 512 | lm loss: 1.144135E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.190 | TFLOPs: 42.66 | 31: iteration 51880/ 52452 | consumed samples: 26562560 | consumed tokens: 54400122880 | elapsed time per iteration (s): 1.83 | learning rate: 2.005E-05 | global batch size: 512 | lm loss: 1.144479E+00 | grad norm: 0.153 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.098 | TFLOPs: 42.04 | 31: iteration 51890/ 52452 | consumed samples: 26567680 | consumed tokens: 54410608640 | elapsed time per iteration (s): 1.83 | learning rate: 2.005E-05 | global batch size: 512 | lm loss: 1.137825E+00 | grad norm: 0.169 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 279.023 | TFLOPs: 41.88 | 31: iteration 51900/ 52452 | consumed samples: 26572800 | consumed tokens: 54421094400 | elapsed time per iteration (s): 1.81 | learning rate: 2.005E-05 | global batch size: 512 | lm loss: 1.137709E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.637 | TFLOPs: 42.42 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 51900 | lm loss value: 4.205350E+00 | lm loss PPL: 6.704409E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 51900 | lm loss value: 4.052992E+00 | lm loss PPL: 5.756947E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 51910/ 52452 | consumed samples: 26577920 | consumed tokens: 54431580160 | elapsed time per iteration (s): 9.17 | learning rate: 2.005E-05 | global batch size: 512 | lm loss: 1.140521E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.832 | TFLOPs: 8.38 | 31: iteration 51920/ 52452 | consumed samples: 26583040 | consumed tokens: 54442065920 | elapsed time per iteration (s): 1.90 | learning rate: 2.005E-05 | global batch size: 512 | lm loss: 1.139995E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.582 | TFLOPs: 40.46 | 31: iteration 51930/ 52452 | consumed samples: 26588160 | consumed tokens: 54452551680 | elapsed time per iteration (s): 1.87 | learning rate: 2.004E-05 | global batch size: 512 | lm loss: 1.139026E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 274.253 | TFLOPs: 41.16 | 31: iteration 51940/ 52452 | consumed samples: 26593280 | consumed tokens: 54463037440 | elapsed time per iteration (s): 1.79 | learning rate: 2.004E-05 | global batch size: 512 | lm loss: 1.145724E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.270 | TFLOPs: 42.97 | 31: iteration 51950/ 52452 | consumed samples: 26598400 | consumed tokens: 54473523200 | elapsed time per iteration (s): 1.77 | learning rate: 2.004E-05 | global batch size: 512 | lm loss: 1.123709E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.925 | TFLOPs: 43.52 | 31: iteration 51960/ 52452 | consumed samples: 26603520 | consumed tokens: 54484008960 | elapsed time per iteration (s): 1.78 | learning rate: 2.004E-05 | global batch size: 512 | lm loss: 1.144918E+00 | grad norm: 0.156 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.736 | TFLOPs: 43.19 | 31: iteration 51970/ 52452 | consumed samples: 26608640 | consumed tokens: 54494494720 | elapsed time per iteration (s): 1.83 | learning rate: 2.004E-05 | global batch size: 512 | lm loss: 1.141024E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.431 | TFLOPs: 42.09 | 31: iteration 51980/ 52452 | consumed samples: 26613760 | consumed tokens: 54504980480 | elapsed time per iteration (s): 1.78 | learning rate: 2.004E-05 | global batch size: 512 | lm loss: 1.142258E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.783 | TFLOPs: 43.19 | 31: iteration 51990/ 52452 | consumed samples: 26618880 | consumed tokens: 54515466240 | elapsed time per iteration (s): 1.81 | learning rate: 2.004E-05 | global batch size: 512 | lm loss: 1.145832E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.119 | TFLOPs: 42.34 | 0: [2022-12-26 21:32:45,815] [INFO] [logging.py:68:log_dist] [Rank 0] step=52000, skipped=0, lr=[2.003366161854507e-05, 2.003366161854507e-05, 2.003366161854507e-05], mom=[(0.9, 0.95), (0.9, 0.95), (0.9, 0.95)] 31: iteration 52000/ 52452 | consumed samples: 26624000 | consumed tokens: 54525952000 | elapsed time per iteration (s): 1.82 | learning rate: 2.003E-05 | global batch size: 512 | lm loss: 1.144160E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.082 | TFLOPs: 42.34 | 0: steps: 52000 loss: 1.1321 iter time (s): 1.793 samples/sec: 285.497 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 52000 | lm loss value: 4.171833E+00 | lm loss PPL: 6.483416E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 52000 | lm loss value: 4.035876E+00 | lm loss PPL: 5.659249E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 52010/ 52452 | consumed samples: 26629120 | consumed tokens: 54536437760 | elapsed time per iteration (s): 9.19 | learning rate: 2.003E-05 | global batch size: 512 | lm loss: 1.140793E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.701 | TFLOPs: 8.36 | 31: iteration 52020/ 52452 | consumed samples: 26634240 | consumed tokens: 54546923520 | elapsed time per iteration (s): 1.78 | learning rate: 2.003E-05 | global batch size: 512 | lm loss: 1.133647E+00 | grad norm: 0.180 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.370 | TFLOPs: 43.13 | 31: iteration 52030/ 52452 | consumed samples: 26639360 | consumed tokens: 54557409280 | elapsed time per iteration (s): 1.77 | learning rate: 2.003E-05 | global batch size: 512 | lm loss: 1.141250E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.953 | TFLOPs: 43.37 | 31: iteration 52040/ 52452 | consumed samples: 26644480 | consumed tokens: 54567895040 | elapsed time per iteration (s): 1.78 | learning rate: 2.003E-05 | global batch size: 512 | lm loss: 1.132879E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.099 | TFLOPs: 43.09 | 31: iteration 52050/ 52452 | consumed samples: 26649600 | consumed tokens: 54578380800 | elapsed time per iteration (s): 1.82 | learning rate: 2.003E-05 | global batch size: 512 | lm loss: 1.135658E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 280.781 | TFLOPs: 42.14 | 31: iteration 52060/ 52452 | consumed samples: 26654720 | consumed tokens: 54588866560 | elapsed time per iteration (s): 1.77 | learning rate: 2.003E-05 | global batch size: 512 | lm loss: 1.155677E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.213 | TFLOPs: 43.41 | 31: iteration 52070/ 52452 | consumed samples: 26659840 | consumed tokens: 54599352320 | elapsed time per iteration (s): 1.79 | learning rate: 2.002E-05 | global batch size: 512 | lm loss: 1.130906E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.479 | TFLOPs: 42.85 | 31: iteration 52080/ 52452 | consumed samples: 26664960 | consumed tokens: 54609838080 | elapsed time per iteration (s): 1.75 | learning rate: 2.002E-05 | global batch size: 512 | lm loss: 1.132655E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 291.858 | TFLOPs: 43.81 | 31: iteration 52090/ 52452 | consumed samples: 26670080 | consumed tokens: 54620323840 | elapsed time per iteration (s): 1.77 | learning rate: 2.002E-05 | global batch size: 512 | lm loss: 1.137654E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.925 | TFLOPs: 43.37 | 31: iteration 52100/ 52452 | consumed samples: 26675200 | consumed tokens: 54630809600 | elapsed time per iteration (s): 1.79 | learning rate: 2.002E-05 | global batch size: 512 | lm loss: 1.143689E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.381 | TFLOPs: 42.98 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 52100 | lm loss value: 4.196022E+00 | lm loss PPL: 6.642155E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 52100 | lm loss value: 4.058759E+00 | lm loss PPL: 5.790239E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 52110/ 52452 | consumed samples: 26680320 | consumed tokens: 54641295360 | elapsed time per iteration (s): 9.14 | learning rate: 2.002E-05 | global batch size: 512 | lm loss: 1.133698E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.993 | TFLOPs: 8.40 | 31: iteration 52120/ 52452 | consumed samples: 26685440 | consumed tokens: 54651781120 | elapsed time per iteration (s): 1.78 | learning rate: 2.002E-05 | global batch size: 512 | lm loss: 1.126380E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.424 | TFLOPs: 43.29 | 31: iteration 52130/ 52452 | consumed samples: 26690560 | consumed tokens: 54662266880 | elapsed time per iteration (s): 1.77 | learning rate: 2.002E-05 | global batch size: 512 | lm loss: 1.127417E+00 | grad norm: 0.155 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.702 | TFLOPs: 43.48 | 31: iteration 52140/ 52452 | consumed samples: 26695680 | consumed tokens: 54672752640 | elapsed time per iteration (s): 1.73 | learning rate: 2.002E-05 | global batch size: 512 | lm loss: 1.136921E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 295.798 | TFLOPs: 44.40 | 31: iteration 52150/ 52452 | consumed samples: 26700800 | consumed tokens: 54683238400 | elapsed time per iteration (s): 1.89 | learning rate: 2.002E-05 | global batch size: 512 | lm loss: 1.122688E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 270.489 | TFLOPs: 40.60 | 31: iteration 52160/ 52452 | consumed samples: 26705920 | consumed tokens: 54693724160 | elapsed time per iteration (s): 1.74 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.142862E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.858 | TFLOPs: 44.26 | 31: iteration 52170/ 52452 | consumed samples: 26711040 | consumed tokens: 54704209920 | elapsed time per iteration (s): 1.75 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.126798E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.475 | TFLOPs: 43.90 | 31: iteration 52180/ 52452 | consumed samples: 26716160 | consumed tokens: 54714695680 | elapsed time per iteration (s): 1.77 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.141028E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.600 | TFLOPs: 43.47 | 31: iteration 52190/ 52452 | consumed samples: 26721280 | consumed tokens: 54725181440 | elapsed time per iteration (s): 1.76 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.147588E+00 | grad norm: 0.164 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 290.944 | TFLOPs: 43.67 | 31: iteration 52200/ 52452 | consumed samples: 26726400 | consumed tokens: 54735667200 | elapsed time per iteration (s): 1.75 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.137489E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.249 | TFLOPs: 44.02 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 52200 | lm loss value: 4.211977E+00 | lm loss PPL: 6.748987E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 52200 | lm loss value: 4.068907E+00 | lm loss PPL: 5.849301E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 52210/ 52452 | consumed samples: 26731520 | consumed tokens: 54746152960 | elapsed time per iteration (s): 9.14 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.147188E+00 | grad norm: 0.157 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.992 | TFLOPs: 8.40 | 31: iteration 52220/ 52452 | consumed samples: 26736640 | consumed tokens: 54756638720 | elapsed time per iteration (s): 1.79 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.143551E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.260 | TFLOPs: 42.82 | 31: iteration 52230/ 52452 | consumed samples: 26741760 | consumed tokens: 54767124480 | elapsed time per iteration (s): 1.75 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.132021E+00 | grad norm: 0.171 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.854 | TFLOPs: 43.96 | 31: iteration 52240/ 52452 | consumed samples: 26746880 | consumed tokens: 54777610240 | elapsed time per iteration (s): 1.80 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.123040E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 284.461 | TFLOPs: 42.70 | 31: iteration 52250/ 52452 | consumed samples: 26752000 | consumed tokens: 54788096000 | elapsed time per iteration (s): 1.74 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.121988E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.424 | TFLOPs: 44.19 | 31: iteration 52260/ 52452 | consumed samples: 26757120 | consumed tokens: 54798581760 | elapsed time per iteration (s): 1.79 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.134656E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.306 | TFLOPs: 42.82 | 31: iteration 52270/ 52452 | consumed samples: 26762240 | consumed tokens: 54809067520 | elapsed time per iteration (s): 1.82 | learning rate: 2.001E-05 | global batch size: 512 | lm loss: 1.144443E+00 | grad norm: 0.167 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 281.519 | TFLOPs: 42.25 | 31: iteration 52280/ 52452 | consumed samples: 26767360 | consumed tokens: 54819553280 | elapsed time per iteration (s): 1.74 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.131931E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.912 | TFLOPs: 44.11 | 31: iteration 52290/ 52452 | consumed samples: 26772480 | consumed tokens: 54830039040 | elapsed time per iteration (s): 1.96 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.139994E+00 | grad norm: 0.181 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 260.801 | TFLOPs: 39.14 | 31: iteration 52300/ 52452 | consumed samples: 26777600 | consumed tokens: 54840524800 | elapsed time per iteration (s): 1.81 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.125085E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 282.261 | TFLOPs: 42.37 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 52300 | lm loss value: 4.185086E+00 | lm loss PPL: 6.569914E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 52300 | lm loss value: 4.043207E+00 | lm loss PPL: 5.700889E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 52310/ 52452 | consumed samples: 26782720 | consumed tokens: 54851010560 | elapsed time per iteration (s): 9.15 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.134579E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.976 | TFLOPs: 8.40 | 31: iteration 52320/ 52452 | consumed samples: 26787840 | consumed tokens: 54861496320 | elapsed time per iteration (s): 1.90 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.136114E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 269.776 | TFLOPs: 40.49 | 31: iteration 52330/ 52452 | consumed samples: 26792960 | consumed tokens: 54871982080 | elapsed time per iteration (s): 1.74 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.133290E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.098 | TFLOPs: 44.14 | 31: iteration 52340/ 52452 | consumed samples: 26798080 | consumed tokens: 54882467840 | elapsed time per iteration (s): 1.78 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.138459E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.079 | TFLOPs: 43.24 | 31: iteration 52350/ 52452 | consumed samples: 26803200 | consumed tokens: 54892953600 | elapsed time per iteration (s): 1.77 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.146521E+00 | grad norm: 0.161 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 288.753 | TFLOPs: 43.34 | 31: iteration 52360/ 52452 | consumed samples: 26808320 | consumed tokens: 54903439360 | elapsed time per iteration (s): 1.74 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.136060E+00 | grad norm: 0.152 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 294.244 | TFLOPs: 44.16 | 31: iteration 52370/ 52452 | consumed samples: 26813440 | consumed tokens: 54913925120 | elapsed time per iteration (s): 1.75 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.137077E+00 | grad norm: 0.160 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 292.883 | TFLOPs: 43.96 | 31: iteration 52380/ 52452 | consumed samples: 26818560 | consumed tokens: 54924410880 | elapsed time per iteration (s): 1.84 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.130986E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 277.745 | TFLOPs: 41.69 | 31: iteration 52390/ 52452 | consumed samples: 26823680 | consumed tokens: 54934896640 | elapsed time per iteration (s): 1.80 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.130587E+00 | grad norm: 0.166 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 285.103 | TFLOPs: 42.79 | 31: iteration 52400/ 52452 | consumed samples: 26828800 | consumed tokens: 54945382400 | elapsed time per iteration (s): 1.85 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.128655E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 276.670 | TFLOPs: 41.53 | 31: --------------------------------------------------------------------------------------------------- 31: validation_c4 loss at iteration 52400 | lm loss value: 4.267966E+00 | lm loss PPL: 7.137629E+01 | 31: --------------------------------------------------------------------------------------------------- 31: ------------------------------------------------------------------------------------------------------ 31: validation_oscar loss at iteration 52400 | lm loss value: 4.103270E+00 | lm loss PPL: 6.053793E+01 | 31: ------------------------------------------------------------------------------------------------------ 31: iteration 52410/ 52452 | consumed samples: 26833920 | consumed tokens: 54955868160 | elapsed time per iteration (s): 9.15 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.130429E+00 | grad norm: 0.162 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 55.968 | TFLOPs: 8.40 | 31: iteration 52420/ 52452 | consumed samples: 26839040 | consumed tokens: 54966353920 | elapsed time per iteration (s): 1.78 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.142886E+00 | grad norm: 0.165 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 287.492 | TFLOPs: 43.15 | 31: iteration 52430/ 52452 | consumed samples: 26844160 | consumed tokens: 54976839680 | elapsed time per iteration (s): 1.75 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.124325E+00 | grad norm: 0.159 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 293.046 | TFLOPs: 43.98 | 31: iteration 52440/ 52452 | consumed samples: 26849280 | consumed tokens: 54987325440 | elapsed time per iteration (s): 1.79 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.136813E+00 | grad norm: 0.163 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 286.346 | TFLOPs: 42.98 | 31: iteration 52450/ 52452 | consumed samples: 26854400 | consumed tokens: 54997811200 | elapsed time per iteration (s): 1.77 | learning rate: 2.000E-05 | global batch size: 512 | lm loss: 1.130271E+00 | grad norm: 0.158 | num zeros: 0.0 | number of skipped iterations: 0 | number of nan iterations: 0 | samples per second: 289.959 | TFLOPs: 43.52 | 0: [after training is done] datetime: 2022-12-26 21:52:21 31: -------------------------------------------------------------------------------------------------------------------- 31: validation_c4 loss at the end of training for val data | lm loss value: 4.253663E+00 | lm loss PPL: 7.036268E+01 | 31: -------------------------------------------------------------------------------------------------------------------- 0: saving checkpoint at iteration 52452 to checkpoints_2b855brealtasky 31: ----------------------------------------------------------------------------------------------------------------------- 31: validation_oscar loss at the end of training for val data | lm loss value: 4.092075E+00 | lm loss PPL: 5.986397E+01 | 31: ----------------------------------------------------------------------------------------------------------------------- 0: [2022-12-26 21:53:35,048] [INFO] [logging.py:68:log_dist] [Rank 0] [Torch] Checkpoint global_step52452 is begin to save! 0: [2022-12-26 21:53:35,059] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_01-model_00-model_states.pt... 0: [2022-12-26 21:53:35,400] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_01-model_00-model_states.pt. 0: [2022-12-26 21:53:35,400] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_03-model_00-model_states.pt... 0: [2022-12-26 21:53:35,581] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_03-model_00-model_states.pt. 0: [2022-12-26 21:53:35,581] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_04-model_00-model_states.pt... 0: [2022-12-26 21:53:35,769] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_04-model_00-model_states.pt. 0: [2022-12-26 21:53:35,770] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_05-model_00-model_states.pt... 0: [2022-12-26 21:53:35,950] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_05-model_00-model_states.pt. 0: [2022-12-26 21:53:35,950] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_06-model_00-model_states.pt... 0: [2022-12-26 21:53:36,130] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_06-model_00-model_states.pt. 0: [2022-12-26 21:53:36,130] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_07-model_00-model_states.pt... 0: [2022-12-26 21:53:36,318] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_07-model_00-model_states.pt. 0: [2022-12-26 21:53:36,318] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_08-model_00-model_states.pt... 0: [2022-12-26 21:53:36,503] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_08-model_00-model_states.pt. 0: [2022-12-26 21:53:36,503] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_09-model_00-model_states.pt... 0: [2022-12-26 21:53:36,689] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_09-model_00-model_states.pt. 0: [2022-12-26 21:53:36,690] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_10-model_00-model_states.pt... 0: [2022-12-26 21:53:36,866] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_10-model_00-model_states.pt. 0: [2022-12-26 21:53:36,866] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_11-model_00-model_states.pt... 0: [2022-12-26 21:53:37,049] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_11-model_00-model_states.pt. 0: [2022-12-26 21:53:37,049] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_12-model_00-model_states.pt... 0: [2022-12-26 21:53:37,231] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_12-model_00-model_states.pt. 0: [2022-12-26 21:53:37,231] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_13-model_00-model_states.pt... 0: [2022-12-26 21:53:37,409] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_13-model_00-model_states.pt. 0: [2022-12-26 21:53:37,410] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_14-model_00-model_states.pt... 0: [2022-12-26 21:53:37,588] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_14-model_00-model_states.pt. 0: [2022-12-26 21:53:37,589] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_15-model_00-model_states.pt... 0: [2022-12-26 21:53:37,770] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_15-model_00-model_states.pt. 0: [2022-12-26 21:53:37,771] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_16-model_00-model_states.pt... 0: [2022-12-26 21:53:37,947] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_16-model_00-model_states.pt. 0: [2022-12-26 21:53:37,947] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_17-model_00-model_states.pt... 0: [2022-12-26 21:53:38,121] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_17-model_00-model_states.pt. 0: [2022-12-26 21:53:38,121] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_18-model_00-model_states.pt... 0: [2022-12-26 21:53:38,301] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_18-model_00-model_states.pt. 0: [2022-12-26 21:53:38,301] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_19-model_00-model_states.pt... 0: [2022-12-26 21:53:38,477] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_19-model_00-model_states.pt. 0: [2022-12-26 21:53:38,477] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_20-model_00-model_states.pt... 0: [2022-12-26 21:53:38,654] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_20-model_00-model_states.pt. 0: [2022-12-26 21:53:38,655] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_21-model_00-model_states.pt... 0: [2022-12-26 21:53:38,831] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_21-model_00-model_states.pt. 0: [2022-12-26 21:53:38,831] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_22-model_00-model_states.pt... 0: [2022-12-26 21:53:39,006] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_22-model_00-model_states.pt. 0: [2022-12-26 21:53:39,007] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_23-model_00-model_states.pt... 0: [2022-12-26 21:53:39,188] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_23-model_00-model_states.pt. 0: [2022-12-26 21:53:39,189] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_24-model_00-model_states.pt... 0: [2022-12-26 21:53:39,361] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_24-model_00-model_states.pt. 0: [2022-12-26 21:53:39,361] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_25-model_00-model_states.pt... 0: [2022-12-26 21:53:39,537] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_25-model_00-model_states.pt. 0: [2022-12-26 21:53:39,538] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_26-model_00-model_states.pt... 0: [2022-12-26 21:53:39,715] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_26-model_00-model_states.pt. 0: [2022-12-26 21:53:39,715] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_27-model_00-model_states.pt... 0: [2022-12-26 21:53:39,887] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_27-model_00-model_states.pt. 0: [2022-12-26 21:53:39,887] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_28-model_00-model_states.pt... 0: [2022-12-26 21:53:40,061] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_28-model_00-model_states.pt. 0: [2022-12-26 21:53:40,061] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_29-model_00-model_states.pt... 0: [2022-12-26 21:53:40,230] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_29-model_00-model_states.pt. 0: [2022-12-26 21:53:40,231] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_30-model_00-model_states.pt... 0: [2022-12-26 21:53:40,405] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_30-model_00-model_states.pt. 0: [2022-12-26 21:53:40,406] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_31-model_00-model_states.pt... 0: [2022-12-26 21:53:40,573] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_31-model_00-model_states.pt. 0: [2022-12-26 21:53:40,574] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_32-model_00-model_states.pt... 0: [2022-12-26 21:53:40,747] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_32-model_00-model_states.pt. 0: [2022-12-26 21:53:40,748] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_33-model_00-model_states.pt... 0: [2022-12-26 21:53:40,924] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_33-model_00-model_states.pt. 0: [2022-12-26 21:53:40,925] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_34-model_00-model_states.pt... 0: [2022-12-26 21:53:41,093] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_34-model_00-model_states.pt. 0: [2022-12-26 21:53:41,093] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_35-model_00-model_states.pt... 0: [2022-12-26 21:53:41,274] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_35-model_00-model_states.pt. 0: [2022-12-26 21:53:41,274] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_36-model_00-model_states.pt... 0: [2022-12-26 21:53:41,452] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_36-model_00-model_states.pt. 0: [2022-12-26 21:53:41,452] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/layer_38-model_00-model_states.pt... 0: [2022-12-26 21:53:41,454] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/layer_38-model_00-model_states.pt. 0: [2022-12-26 21:53:41,455] [INFO] [logging.py:68:log_dist] [Rank 0] Saving model checkpoint: checkpoints_2b855brealtasky/global_step52452/mp_rank_00_model_states.pt 0: [2022-12-26 21:53:41,455] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/mp_rank_00_model_states.pt... 0: [2022-12-26 21:53:41,460] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/mp_rank_00_model_states.pt. 0: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... 16: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt... 16: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt... 24: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt... 28: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt... 28: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt... 28: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt... 28: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt... 0: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... 0: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... 16: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt... 14: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt... 14: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt... 12: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt... 12: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt... 12: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt... 12: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt... 3: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt... 3: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt... 3: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt... 3: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt... 3: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt... 18: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt... 18: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt... 18: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt... 18: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt... 18: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt... 8: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt... 8: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt... 5: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt... 5: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt... 5: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt... 5: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt... 5: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt... 4: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt... 4: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt... 4: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt... 4: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt... 4: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt... 7: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt... 7: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt... 7: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt... 7: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt... 7: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt... 9: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt... 9: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt... 9: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt... 9: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt... 13: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt... 13: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt... 13: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt... 17: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt... 17: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt... 17: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt... 17: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt... 17: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt... 21: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt... 21: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt... 21: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt... 21: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt... 21: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt... 30: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt... 30: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt... 30: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt... 30: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt... 30: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt... 26: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt... 26: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt... 26: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt... 26: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt... 26: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt... 19: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt... 19: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt... 19: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt... 19: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt... 25: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt... 25: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt... 25: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt... 25: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt... 25: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt... 31: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt... 31: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt... 31: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt... 31: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt... 29: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt... 11: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt... 11: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt... 11: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt... 11: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt... 10: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt... 10: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt... 10: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt... 10: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt... 10: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt... 23: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt... 23: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt... 23: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt... 23: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt... 23: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt... 20: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt... 2: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt... 27: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt... 27: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt... 27: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt... 27: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt... 27: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt... 22: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt... 22: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt... 22: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt... 22: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt... 22: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt... 1: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... 1: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... 6: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt... 15: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt... 15: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt... 15: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt... 0: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... 0: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... 0: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... 16: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt... 14: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt... 14: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt... 14: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt... 12: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt... 12: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt... 12: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt... 3: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt... 3: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt... 18: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt... 18: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt... 18: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt... 8: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt... 8: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt... 8: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt... 8: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt... 8: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt... 5: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt... 5: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt... 4: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt... 7: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt... 9: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt... 13: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt... 13: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt... 13: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt... 13: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt... 13: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt... 24: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt... 24: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt... 17: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt... 28: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt... 28: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt... 21: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt... 21: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt... 30: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt... 26: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt... 19: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt... 19: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt... 19: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt... 25: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt... 31: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt... 31: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt... 29: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt... 29: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt... 29: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt... 29: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt... 29: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt... 11: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt... 11: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt... 11: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt... 10: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt... 23: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt... 20: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt... 20: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt... 20: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt... 20: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt... 20: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt... 2: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt... 2: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt... 27: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt... 27: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt... 22: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt... 1: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... 1: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... 1: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... 1: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... 1: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... 6: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt... 15: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt... 15: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt... 15: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt... 0: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... 16: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt... 16: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt... 14: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt... 14: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt... 14: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt... 12: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt... 3: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt... 8: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt... 5: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt... 4: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt... 7: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt... 9: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt... 24: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt... 24: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt... 24: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt... 24: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt... 17: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt... 17: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt... 28: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt... 21: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt... 30: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt... 30: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt... 26: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt... 19: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt... 25: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt... 31: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt... 29: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt... 11: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt... 10: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt... 23: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt... 20: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt... 20: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt... 2: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt... 2: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt... 2: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt... 27: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt... 22: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt... 22: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt... 1: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... 6: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt... 15: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt... 15: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt... 0: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... 16: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt... 4: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt... 7: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt... 9: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt... 9: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt... 24: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt... 28: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt... 25: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt... 31: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt... 29: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt... 10: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt... 23: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt... 2: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt... 2: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt... 6: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt... 16: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt... 26: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt... 6: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt... 6: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt... 6: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt... 6: [2022-12-26 21:53:41,547] [INFO] [torch_checkpoint_engine.py:15:save] [Torch] Saving checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt... 17: [2022-12-26 21:53:41,717] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt. 17: [2022-12-26 21:53:41,717] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_143_mp_rank_00_optim_states.pt 17: [2022-12-26 21:53:41,717] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 7: [2022-12-26 21:53:41,719] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt. 7: [2022-12-26 21:53:41,719] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt. 7: [2022-12-26 21:53:41,719] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt. 7: [2022-12-26 21:53:41,719] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt. 7: [2022-12-26 21:53:41,719] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_59_mp_rank_00_optim_states.pt 7: [2022-12-26 21:53:41,719] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_60_mp_rank_00_optim_states.pt 7: [2022-12-26 21:53:41,719] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_57_mp_rank_00_optim_states.pt 7: [2022-12-26 21:53:41,719] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_62_mp_rank_00_optim_states.pt 7: [2022-12-26 21:53:41,719] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 7: [2022-12-26 21:53:41,719] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 7: [2022-12-26 21:53:41,719] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 7: [2022-12-26 21:53:41,719] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 18: [2022-12-26 21:53:41,720] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt. 18: [2022-12-26 21:53:41,720] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt. 18: [2022-12-26 21:53:41,720] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_144_mp_rank_00_optim_states.pt 18: [2022-12-26 21:53:41,720] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 18: [2022-12-26 21:53:41,720] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_149_mp_rank_00_optim_states.pt 18: [2022-12-26 21:53:41,720] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 17: [2022-12-26 21:53:41,721] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt. 17: [2022-12-26 21:53:41,721] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_139_mp_rank_00_optim_states.pt 17: [2022-12-26 21:53:41,721] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 17: [2022-12-26 21:53:41,723] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt. 17: [2022-12-26 21:53:41,723] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_140_mp_rank_00_optim_states.pt 17: [2022-12-26 21:53:41,723] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 4: [2022-12-26 21:53:41,725] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt. 4: [2022-12-26 21:53:41,725] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt. 4: [2022-12-26 21:53:41,725] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_33_mp_rank_00_optim_states.pt 4: [2022-12-26 21:53:41,725] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_39_mp_rank_00_optim_states.pt 4: [2022-12-26 21:53:41,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 4: [2022-12-26 21:53:41,726] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 7: [2022-12-26 21:53:41,727] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt. 7: [2022-12-26 21:53:41,727] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_61_mp_rank_00_optim_states.pt 7: [2022-12-26 21:53:41,727] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 18: [2022-12-26 21:53:41,727] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt. 18: [2022-12-26 21:53:41,727] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt. 18: [2022-12-26 21:53:41,727] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_148_mp_rank_00_optim_states.pt 18: [2022-12-26 21:53:41,727] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_151_mp_rank_00_optim_states.pt 18: [2022-12-26 21:53:41,727] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 18: [2022-12-26 21:53:41,727] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 18: [2022-12-26 21:53:41,728] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt. 18: [2022-12-26 21:53:41,729] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_145_mp_rank_00_optim_states.pt 18: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 15: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt. 15: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt. 15: [2022-12-26 21:53:41,729] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_120_mp_rank_00_optim_states.pt 15: [2022-12-26 21:53:41,729] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_126_mp_rank_00_optim_states.pt 15: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 15: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 21: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt. 21: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt. 21: [2022-12-26 21:53:41,729] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_172_mp_rank_00_optim_states.pt 18: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt. 21: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 21: [2022-12-26 21:53:41,729] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_171_mp_rank_00_optim_states.pt 18: [2022-12-26 21:53:41,729] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_150_mp_rank_00_optim_states.pt 21: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 18: [2022-12-26 21:53:41,729] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 21: [2022-12-26 21:53:41,730] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt. 21: [2022-12-26 21:53:41,730] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt. 21: [2022-12-26 21:53:41,730] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_175_mp_rank_00_optim_states.pt 21: [2022-12-26 21:53:41,730] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 21: [2022-12-26 21:53:41,730] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_173_mp_rank_00_optim_states.pt 21: [2022-12-26 21:53:41,730] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,731] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt. 2: [2022-12-26 21:53:41,732] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt 2: [2022-12-26 21:53:41,732] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt. 2: [2022-12-26 21:53:41,732] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt. 2: [2022-12-26 21:53:41,732] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,732] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt 2: [2022-12-26 21:53:41,732] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt 2: [2022-12-26 21:53:41,732] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,732] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 17: [2022-12-26 21:53:41,732] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt. 17: [2022-12-26 21:53:41,732] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_138_mp_rank_00_optim_states.pt 17: [2022-12-26 21:53:41,732] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,732] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt. 22: [2022-12-26 21:53:41,734] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt. 22: [2022-12-26 21:53:41,734] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt. 22: [2022-12-26 21:53:41,734] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt. 22: [2022-12-26 21:53:41,734] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_176_mp_rank_00_optim_states.pt 22: [2022-12-26 21:53:41,734] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 22: [2022-12-26 21:53:41,734] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_181_mp_rank_00_optim_states.pt 22: [2022-12-26 21:53:41,734] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_183_mp_rank_00_optim_states.pt 22: [2022-12-26 21:53:41,734] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 22: [2022-12-26 21:53:41,734] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 4: [2022-12-26 21:53:41,734] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt. 4: [2022-12-26 21:53:41,734] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_38_mp_rank_00_optim_states.pt 4: [2022-12-26 21:53:41,734] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 4: [2022-12-26 21:53:41,735] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt. 4: [2022-12-26 21:53:41,735] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_37_mp_rank_00_optim_states.pt 4: [2022-12-26 21:53:41,735] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 4: [2022-12-26 21:53:41,736] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt. 4: [2022-12-26 21:53:41,736] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_35_mp_rank_00_optim_states.pt 4: [2022-12-26 21:53:41,736] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 4: [2022-12-26 21:53:41,736] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt. 4: [2022-12-26 21:53:41,736] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_36_mp_rank_00_optim_states.pt 4: [2022-12-26 21:53:41,737] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 22: [2022-12-26 21:53:41,738] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt. 22: [2022-12-26 21:53:41,738] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_177_mp_rank_00_optim_states.pt 22: [2022-12-26 21:53:41,738] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,732] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt 2: [2022-12-26 21:53:41,733] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,733] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt. 2: [2022-12-26 21:53:41,733] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt 2: [2022-12-26 21:53:41,733] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,733] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt. 2: [2022-12-26 21:53:41,733] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt 2: [2022-12-26 21:53:41,733] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,733] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt. 2: [2022-12-26 21:53:41,733] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt 2: [2022-12-26 21:53:41,733] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 10: [2022-12-26 21:53:41,738] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt. 10: [2022-12-26 21:53:41,738] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt. 10: [2022-12-26 21:53:41,738] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt. 10: [2022-12-26 21:53:41,738] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_82_mp_rank_00_optim_states.pt 10: [2022-12-26 21:53:41,738] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_85_mp_rank_00_optim_states.pt 10: [2022-12-26 21:53:41,738] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_83_mp_rank_00_optim_states.pt 10: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 10: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 10: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt. 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt. 10: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt. 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt. 19: [2022-12-26 21:53:41,739] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_153_mp_rank_00_optim_states.pt 19: [2022-12-26 21:53:41,739] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_158_mp_rank_00_optim_states.pt 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 10: [2022-12-26 21:53:41,739] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_80_mp_rank_00_optim_states.pt 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 19: [2022-12-26 21:53:41,739] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_157_mp_rank_00_optim_states.pt 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt. 10: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 19: [2022-12-26 21:53:41,739] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_156_mp_rank_00_optim_states.pt 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt. 19: [2022-12-26 21:53:41,739] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_152_mp_rank_00_optim_states.pt 19: [2022-12-26 21:53:41,739] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 15: [2022-12-26 21:53:41,730] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt. 15: [2022-12-26 21:53:41,730] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_123_mp_rank_00_optim_states.pt 15: [2022-12-26 21:53:41,730] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 15: [2022-12-26 21:53:41,730] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt. 15: [2022-12-26 21:53:41,730] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_127_mp_rank_00_optim_states.pt 15: [2022-12-26 21:53:41,730] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 15: [2022-12-26 21:53:41,741] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt. 15: [2022-12-26 21:53:41,741] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_122_mp_rank_00_optim_states.pt 15: [2022-12-26 21:53:41,741] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 20: [2022-12-26 21:53:41,742] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt. 20: [2022-12-26 21:53:41,742] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_163_mp_rank_00_optim_states.pt 20: [2022-12-26 21:53:41,742] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 6: [2022-12-26 21:53:41,745] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt. 6: [2022-12-26 21:53:41,745] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt. 6: [2022-12-26 21:53:41,745] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt. 6: [2022-12-26 21:53:41,745] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_53_mp_rank_00_optim_states.pt 6: [2022-12-26 21:53:41,745] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_48_mp_rank_00_optim_states.pt 6: [2022-12-26 21:53:41,745] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_52_mp_rank_00_optim_states.pt 6: [2022-12-26 21:53:41,745] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 6: [2022-12-26 21:53:41,745] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 6: [2022-12-26 21:53:41,745] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 13: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt. 13: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt. 13: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt. 13: [2022-12-26 21:53:41,749] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_108_mp_rank_00_optim_states.pt 13: [2022-12-26 21:53:41,749] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_111_mp_rank_00_optim_states.pt 13: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 13: [2022-12-26 21:53:41,749] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_106_mp_rank_00_optim_states.pt 13: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 13: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 31: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt. 31: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt. 31: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt. 13: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt. 13: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt. 13: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt. 13: [2022-12-26 21:53:41,750] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_107_mp_rank_00_optim_states.pt 13: [2022-12-26 21:53:41,750] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_105_mp_rank_00_optim_states.pt 13: [2022-12-26 21:53:41,750] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_110_mp_rank_00_optim_states.pt 13: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 13: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 13: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 6: [2022-12-26 21:53:41,751] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt. 6: [2022-12-26 21:53:41,751] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_49_mp_rank_00_optim_states.pt 6: [2022-12-26 21:53:41,751] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 17: [2022-12-26 21:53:41,751] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt. 6: [2022-12-26 21:53:41,751] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt. 17: [2022-12-26 21:53:41,751] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_142_mp_rank_00_optim_states.pt 6: [2022-12-26 21:53:41,751] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_50_mp_rank_00_optim_states.pt 17: [2022-12-26 21:53:41,751] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 6: [2022-12-26 21:53:41,751] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 31: [2022-12-26 21:53:41,749] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_254_mp_rank_00_optim_states.pt 31: [2022-12-26 21:53:41,749] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_251_mp_rank_00_optim_states.pt 31: [2022-12-26 21:53:41,749] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_255_mp_rank_00_optim_states.pt 31: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 31: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 31: [2022-12-26 21:53:41,749] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 31: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt. 31: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt. 31: [2022-12-26 21:53:41,750] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_248_mp_rank_00_optim_states.pt 31: [2022-12-26 21:53:41,750] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_253_mp_rank_00_optim_states.pt 31: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 31: [2022-12-26 21:53:41,750] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 22: [2022-12-26 21:53:41,758] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt. 22: [2022-12-26 21:53:41,758] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_178_mp_rank_00_optim_states.pt 22: [2022-12-26 21:53:41,758] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 22: [2022-12-26 21:53:41,758] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt. 22: [2022-12-26 21:53:41,758] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_179_mp_rank_00_optim_states.pt 22: [2022-12-26 21:53:41,758] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 20: [2022-12-26 21:53:41,742] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt. 20: [2022-12-26 21:53:41,742] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_165_mp_rank_00_optim_states.pt 20: [2022-12-26 21:53:41,742] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 20: [2022-12-26 21:53:41,742] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt. 20: [2022-12-26 21:53:41,742] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_162_mp_rank_00_optim_states.pt 20: [2022-12-26 21:53:41,742] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 20: [2022-12-26 21:53:41,761] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt. 20: [2022-12-26 21:53:41,761] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt. 20: [2022-12-26 21:53:41,761] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_161_mp_rank_00_optim_states.pt 20: [2022-12-26 21:53:41,761] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 20: [2022-12-26 21:53:41,761] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_167_mp_rank_00_optim_states.pt 20: [2022-12-26 21:53:41,761] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 10: [2022-12-26 21:53:41,767] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt. 10: [2022-12-26 21:53:41,767] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_84_mp_rank_00_optim_states.pt 10: [2022-12-26 21:53:41,767] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 11: [2022-12-26 21:53:41,772] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt. 11: [2022-12-26 21:53:41,772] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt. 11: [2022-12-26 21:53:41,773] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_92_mp_rank_00_optim_states.pt 11: [2022-12-26 21:53:41,772] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt. 11: [2022-12-26 21:53:41,772] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt. 11: [2022-12-26 21:53:41,772] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt. 11: [2022-12-26 21:53:41,773] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_89_mp_rank_00_optim_states.pt 11: [2022-12-26 21:53:41,773] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 11: [2022-12-26 21:53:41,773] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_90_mp_rank_00_optim_states.pt 11: [2022-12-26 21:53:41,773] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 11: [2022-12-26 21:53:41,773] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_88_mp_rank_00_optim_states.pt 11: [2022-12-26 21:53:41,773] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_95_mp_rank_00_optim_states.pt 11: [2022-12-26 21:53:41,773] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 11: [2022-12-26 21:53:41,773] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 11: [2022-12-26 21:53:41,773] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 29: [2022-12-26 21:53:41,761] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt. 29: [2022-12-26 21:53:41,761] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt. 29: [2022-12-26 21:53:41,761] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt. 29: [2022-12-26 21:53:41,761] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt. 29: [2022-12-26 21:53:41,761] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt. 29: [2022-12-26 21:53:41,762] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_234_mp_rank_00_optim_states.pt 29: [2022-12-26 21:53:41,762] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_239_mp_rank_00_optim_states.pt 29: [2022-12-26 21:53:41,762] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_238_mp_rank_00_optim_states.pt 29: [2022-12-26 21:53:41,762] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 29: [2022-12-26 21:53:41,762] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_236_mp_rank_00_optim_states.pt 29: [2022-12-26 21:53:41,762] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 29: [2022-12-26 21:53:41,762] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 29: [2022-12-26 21:53:41,762] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 29: [2022-12-26 21:53:41,762] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_237_mp_rank_00_optim_states.pt 29: [2022-12-26 21:53:41,762] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: [2022-12-26 21:53:41,775] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. 0: [2022-12-26 21:53:41,775] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. 0: [2022-12-26 21:53:41,775] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. 0: [2022-12-26 21:53:41,775] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt 0: [2022-12-26 21:53:41,776] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt 0: [2022-12-26 21:53:41,776] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt 0: [2022-12-26 21:53:41,776] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: [2022-12-26 21:53:41,776] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: [2022-12-26 21:53:41,776] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: [2022-12-26 21:53:41,783] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. 0: [2022-12-26 21:53:41,783] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt 0: [2022-12-26 21:53:41,783] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: [2022-12-26 21:53:41,783] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. 0: [2022-12-26 21:53:41,783] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt 0: [2022-12-26 21:53:41,783] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: [2022-12-26 21:53:41,783] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. 0: [2022-12-26 21:53:41,784] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt 0: [2022-12-26 21:53:41,784] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. 0: [2022-12-26 21:53:41,784] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: [2022-12-26 21:53:41,784] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt 0: [2022-12-26 21:53:41,784] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: [2022-12-26 21:53:41,784] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. 1: [2022-12-26 21:53:41,786] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. 1: [2022-12-26 21:53:41,786] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. 1: [2022-12-26 21:53:41,786] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. 1: [2022-12-26 21:53:41,786] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. 1: [2022-12-26 21:53:41,786] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. 1: [2022-12-26 21:53:41,787] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt 1: [2022-12-26 21:53:41,787] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt 1: [2022-12-26 21:53:41,787] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt 1: [2022-12-26 21:53:41,787] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt 1: [2022-12-26 21:53:41,787] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 1: [2022-12-26 21:53:41,787] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt 1: [2022-12-26 21:53:41,787] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 1: [2022-12-26 21:53:41,787] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 1: [2022-12-26 21:53:41,787] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 1: [2022-12-26 21:53:41,787] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 4: [2022-12-26 21:53:41,788] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt. 4: [2022-12-26 21:53:41,788] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_32_mp_rank_00_optim_states.pt 4: [2022-12-26 21:53:41,788] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,788] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:41,790] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_200_mp_rank_00_optim_states.pt 25: [2022-12-26 21:53:41,790] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_201_mp_rank_00_optim_states.pt 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 25: [2022-12-26 21:53:41,790] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_204_mp_rank_00_optim_states.pt 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 25: [2022-12-26 21:53:41,790] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_205_mp_rank_00_optim_states.pt 25: [2022-12-26 21:53:41,790] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_203_mp_rank_00_optim_states.pt 28: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt. 28: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt. 28: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt. 28: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 28: [2022-12-26 21:53:41,790] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_225_mp_rank_00_optim_states.pt 28: [2022-12-26 21:53:41,790] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_224_mp_rank_00_optim_states.pt 28: [2022-12-26 21:53:41,790] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_228_mp_rank_00_optim_states.pt 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 25: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 28: [2022-12-26 21:53:41,790] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_227_mp_rank_00_optim_states.pt 28: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 28: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 28: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 28: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 28: [2022-12-26 21:53:41,790] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt. 28: [2022-12-26 21:53:41,791] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_231_mp_rank_00_optim_states.pt 28: [2022-12-26 21:53:41,791] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt. 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt. 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt. 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt. 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt. 27: [2022-12-26 21:53:41,793] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_218_mp_rank_00_optim_states.pt 27: [2022-12-26 21:53:41,793] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_220_mp_rank_00_optim_states.pt 27: [2022-12-26 21:53:41,793] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_216_mp_rank_00_optim_states.pt 27: [2022-12-26 21:53:41,793] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_219_mp_rank_00_optim_states.pt 27: [2022-12-26 21:53:41,793] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_221_mp_rank_00_optim_states.pt 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 27: [2022-12-26 21:53:41,793] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 2: [2022-12-26 21:53:41,788] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt 2: [2022-12-26 21:53:41,788] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt. 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt. 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt. 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt. 12: [2022-12-26 21:53:41,800] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_98_mp_rank_00_optim_states.pt 12: [2022-12-26 21:53:41,800] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_103_mp_rank_00_optim_states.pt 12: [2022-12-26 21:53:41,800] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_101_mp_rank_00_optim_states.pt 12: [2022-12-26 21:53:41,800] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_100_mp_rank_00_optim_states.pt 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt. 12: [2022-12-26 21:53:41,800] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_99_mp_rank_00_optim_states.pt 12: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 9: [2022-12-26 21:53:41,795] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt. 9: [2022-12-26 21:53:41,795] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt. 9: [2022-12-26 21:53:41,795] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt. 9: [2022-12-26 21:53:41,795] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt. 9: [2022-12-26 21:53:41,795] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt. 9: [2022-12-26 21:53:41,796] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_75_mp_rank_00_optim_states.pt 9: [2022-12-26 21:53:41,796] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_74_mp_rank_00_optim_states.pt 9: [2022-12-26 21:53:41,796] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_72_mp_rank_00_optim_states.pt 9: [2022-12-26 21:53:41,796] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_73_mp_rank_00_optim_states.pt 9: [2022-12-26 21:53:41,796] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 9: [2022-12-26 21:53:41,796] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 9: [2022-12-26 21:53:41,796] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_76_mp_rank_00_optim_states.pt 9: [2022-12-26 21:53:41,796] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 9: [2022-12-26 21:53:41,796] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 9: [2022-12-26 21:53:41,796] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 8: [2022-12-26 21:53:41,813] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt. 8: [2022-12-26 21:53:41,814] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt. 8: [2022-12-26 21:53:41,814] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt. 8: [2022-12-26 21:53:41,814] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_69_mp_rank_00_optim_states.pt 8: [2022-12-26 21:53:41,814] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_71_mp_rank_00_optim_states.pt 8: [2022-12-26 21:53:41,814] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 8: [2022-12-26 21:53:41,814] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 8: [2022-12-26 21:53:41,814] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_67_mp_rank_00_optim_states.pt 8: [2022-12-26 21:53:41,814] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 8: [2022-12-26 21:53:41,815] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt. 8: [2022-12-26 21:53:41,815] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt. 8: [2022-12-26 21:53:41,815] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_66_mp_rank_00_optim_states.pt 8: [2022-12-26 21:53:41,815] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_64_mp_rank_00_optim_states.pt 8: [2022-12-26 21:53:41,815] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 8: [2022-12-26 21:53:41,815] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 21: [2022-12-26 21:53:41,829] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt. 21: [2022-12-26 21:53:41,829] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_169_mp_rank_00_optim_states.pt 21: [2022-12-26 21:53:41,829] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 3: [2022-12-26 21:53:41,794] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt. 3: [2022-12-26 21:53:41,794] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt. 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt. 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt. 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt. 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt. 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt. 3: [2022-12-26 21:53:41,794] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt. 3: [2022-12-26 21:53:41,794] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt. 3: [2022-12-26 21:53:41,794] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt 3: [2022-12-26 21:53:41,794] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt 3: [2022-12-26 21:53:41,794] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt 3: [2022-12-26 21:53:41,794] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt 3: [2022-12-26 21:53:41,795] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 3: [2022-12-26 21:53:41,795] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 3: [2022-12-26 21:53:41,795] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 3: [2022-12-26 21:53:41,795] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 3: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt. 3: [2022-12-26 21:53:41,800] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt 3: [2022-12-26 21:53:41,800] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 5: [2022-12-26 21:53:41,798] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_46_mp_rank_00_optim_states.pt 5: [2022-12-26 21:53:41,798] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_44_mp_rank_00_optim_states.pt 5: [2022-12-26 21:53:41,798] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_42_mp_rank_00_optim_states.pt 5: [2022-12-26 21:53:41,798] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_41_mp_rank_00_optim_states.pt 5: [2022-12-26 21:53:41,798] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_47_mp_rank_00_optim_states.pt 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 5: [2022-12-26 21:53:41,798] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt. 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt. 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt. 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt. 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt. 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt. 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt. 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt. 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt. 14: [2022-12-26 21:53:41,858] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_112_mp_rank_00_optim_states.pt 14: [2022-12-26 21:53:41,858] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_116_mp_rank_00_optim_states.pt 14: [2022-12-26 21:53:41,858] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_118_mp_rank_00_optim_states.pt 14: [2022-12-26 21:53:41,858] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_119_mp_rank_00_optim_states.pt 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt. 14: [2022-12-26 21:53:41,858] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_117_mp_rank_00_optim_states.pt 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 14: [2022-12-26 21:53:41,858] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_115_mp_rank_00_optim_states.pt 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 14: [2022-12-26 21:53:41,858] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt. 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt. 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt. 30: [2022-12-26 21:53:41,839] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_244_mp_rank_00_optim_states.pt 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 30: [2022-12-26 21:53:41,839] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_242_mp_rank_00_optim_states.pt 30: [2022-12-26 21:53:41,839] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_241_mp_rank_00_optim_states.pt 30: [2022-12-26 21:53:41,839] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_243_mp_rank_00_optim_states.pt 30: [2022-12-26 21:53:41,839] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_246_mp_rank_00_optim_states.pt 30: [2022-12-26 21:53:41,839] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_240_mp_rank_00_optim_states.pt 30: [2022-12-26 21:53:41,839] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_247_mp_rank_00_optim_states.pt 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 30: [2022-12-26 21:53:41,839] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 20: [2022-12-26 21:53:41,874] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt. 20: [2022-12-26 21:53:41,874] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_166_mp_rank_00_optim_states.pt 20: [2022-12-26 21:53:41,874] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt. 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt. 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt. 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt. 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt. 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt. 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt. 23: [2022-12-26 21:53:41,880] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_191_mp_rank_00_optim_states.pt 23: [2022-12-26 21:53:41,880] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_190_mp_rank_00_optim_states.pt 23: [2022-12-26 21:53:41,880] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_188_mp_rank_00_optim_states.pt 23: [2022-12-26 21:53:41,880] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_186_mp_rank_00_optim_states.pt 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 23: [2022-12-26 21:53:41,880] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_189_mp_rank_00_optim_states.pt 23: [2022-12-26 21:53:41,880] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_187_mp_rank_00_optim_states.pt 23: [2022-12-26 21:53:41,880] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_184_mp_rank_00_optim_states.pt 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 23: [2022-12-26 21:53:41,880] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt. 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt. 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt. 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt. 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt. 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt. 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt. 16: [2022-12-26 21:53:41,882] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_134_mp_rank_00_optim_states.pt 16: [2022-12-26 21:53:41,882] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_128_mp_rank_00_optim_states.pt 16: [2022-12-26 21:53:41,882] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_133_mp_rank_00_optim_states.pt 16: [2022-12-26 21:53:41,882] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_130_mp_rank_00_optim_states.pt 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 16: [2022-12-26 21:53:41,882] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_132_mp_rank_00_optim_states.pt 16: [2022-12-26 21:53:41,882] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_135_mp_rank_00_optim_states.pt 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 16: [2022-12-26 21:53:41,882] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_131_mp_rank_00_optim_states.pt 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 16: [2022-12-26 21:53:41,882] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 24: [2022-12-26 21:53:41,852] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt. 24: [2022-12-26 21:53:41,852] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt. 24: [2022-12-26 21:53:41,852] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt. 24: [2022-12-26 21:53:41,853] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_196_mp_rank_00_optim_states.pt 24: [2022-12-26 21:53:41,853] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_192_mp_rank_00_optim_states.pt 24: [2022-12-26 21:53:41,853] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_195_mp_rank_00_optim_states.pt 24: [2022-12-26 21:53:41,853] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 24: [2022-12-26 21:53:41,853] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 24: [2022-12-26 21:53:41,853] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 24: [2022-12-26 21:53:41,863] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt. 24: [2022-12-26 21:53:41,863] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt. 24: [2022-12-26 21:53:41,863] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt. 24: [2022-12-26 21:53:41,863] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_199_mp_rank_00_optim_states.pt 24: [2022-12-26 21:53:41,863] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt. 24: [2022-12-26 21:53:41,863] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 24: [2022-12-26 21:53:41,863] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_194_mp_rank_00_optim_states.pt 24: [2022-12-26 21:53:41,863] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_198_mp_rank_00_optim_states.pt 24: [2022-12-26 21:53:41,863] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_193_mp_rank_00_optim_states.pt 24: [2022-12-26 21:53:41,863] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 24: [2022-12-26 21:53:41,863] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 24: [2022-12-26 21:53:41,863] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 9: [2022-12-26 21:53:41,893] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt. 9: [2022-12-26 21:53:41,894] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_78_mp_rank_00_optim_states.pt 9: [2022-12-26 21:53:41,894] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 18: [2022-12-26 21:53:41,899] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt. 18: [2022-12-26 21:53:41,900] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_147_mp_rank_00_optim_states.pt 18: [2022-12-26 21:53:41,900] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: [2022-12-26 21:53:41,900] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt 0: [2022-12-26 21:53:41,900] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt. 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt. 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt. 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt. 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt. 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt. 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt. 26: [2022-12-26 21:53:41,913] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_208_mp_rank_00_optim_states.pt 26: [2022-12-26 21:53:41,913] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_210_mp_rank_00_optim_states.pt 26: [2022-12-26 21:53:41,913] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_212_mp_rank_00_optim_states.pt 26: [2022-12-26 21:53:41,913] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_209_mp_rank_00_optim_states.pt 26: [2022-12-26 21:53:41,913] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_215_mp_rank_00_optim_states.pt 26: [2022-12-26 21:53:41,913] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_213_mp_rank_00_optim_states.pt 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 26: [2022-12-26 21:53:41,913] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_211_mp_rank_00_optim_states.pt 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 26: [2022-12-26 21:53:41,913] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 31: [2022-12-26 21:53:41,937] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt. 31: [2022-12-26 21:53:41,937] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_250_mp_rank_00_optim_states.pt 31: [2022-12-26 21:53:41,937] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 25: [2022-12-26 21:53:41,941] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:41,941] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_206_mp_rank_00_optim_states.pt 25: [2022-12-26 21:53:41,941] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 7: [2022-12-26 21:53:41,941] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt. 7: [2022-12-26 21:53:41,941] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_58_mp_rank_00_optim_states.pt 7: [2022-12-26 21:53:41,941] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 13: [2022-12-26 21:53:41,944] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt. 13: [2022-12-26 21:53:41,944] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_109_mp_rank_00_optim_states.pt 13: [2022-12-26 21:53:41,944] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 24: [2022-12-26 21:53:41,952] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt. 24: [2022-12-26 21:53:41,953] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_197_mp_rank_00_optim_states.pt 24: [2022-12-26 21:53:41,953] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 22: [2022-12-26 21:53:41,955] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt. 22: [2022-12-26 21:53:41,955] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_180_mp_rank_00_optim_states.pt 22: [2022-12-26 21:53:41,955] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 29: [2022-12-26 21:53:41,959] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt. 29: [2022-12-26 21:53:41,960] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_235_mp_rank_00_optim_states.pt 29: [2022-12-26 21:53:41,960] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 23: [2022-12-26 21:53:41,965] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt. 23: [2022-12-26 21:53:41,966] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_185_mp_rank_00_optim_states.pt 23: [2022-12-26 21:53:41,966] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 21: [2022-12-26 21:53:41,967] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt. 21: [2022-12-26 21:53:41,967] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_168_mp_rank_00_optim_states.pt 21: [2022-12-26 21:53:41,967] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 16: [2022-12-26 21:53:41,971] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt. 8: [2022-12-26 21:53:41,971] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt. 16: [2022-12-26 21:53:41,972] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_129_mp_rank_00_optim_states.pt 16: [2022-12-26 21:53:41,972] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 8: [2022-12-26 21:53:41,972] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_65_mp_rank_00_optim_states.pt 8: [2022-12-26 21:53:41,972] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 12: [2022-12-26 21:53:41,972] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt. 12: [2022-12-26 21:53:41,972] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_102_mp_rank_00_optim_states.pt 12: [2022-12-26 21:53:41,973] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 17: [2022-12-26 21:53:41,973] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt. 17: [2022-12-26 21:53:41,974] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_141_mp_rank_00_optim_states.pt 17: [2022-12-26 21:53:41,974] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 19: [2022-12-26 21:53:41,975] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt. 19: [2022-12-26 21:53:41,975] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_159_mp_rank_00_optim_states.pt 19: [2022-12-26 21:53:41,975] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 6: [2022-12-26 21:53:41,978] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt. 6: [2022-12-26 21:53:41,978] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_55_mp_rank_00_optim_states.pt 6: [2022-12-26 21:53:41,978] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 1: [2022-12-26 21:53:41,975] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. 1: [2022-12-26 21:53:41,976] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt 1: [2022-12-26 21:53:41,976] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 10: [2022-12-26 21:53:41,979] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt. 10: [2022-12-26 21:53:41,979] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_87_mp_rank_00_optim_states.pt 10: [2022-12-26 21:53:41,979] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 28: [2022-12-26 21:53:41,980] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt. 28: [2022-12-26 21:53:41,980] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_230_mp_rank_00_optim_states.pt 28: [2022-12-26 21:53:41,980] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 26: [2022-12-26 21:53:41,981] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt. 26: [2022-12-26 21:53:41,981] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_214_mp_rank_00_optim_states.pt 26: [2022-12-26 21:53:41,981] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 4: [2022-12-26 21:53:41,982] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt. 4: [2022-12-26 21:53:41,982] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_34_mp_rank_00_optim_states.pt 4: [2022-12-26 21:53:41,982] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 3: [2022-12-26 21:53:41,983] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt. 3: [2022-12-26 21:53:41,983] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt 3: [2022-12-26 21:53:41,983] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 5: [2022-12-26 21:53:41,984] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt. 5: [2022-12-26 21:53:41,984] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_45_mp_rank_00_optim_states.pt 5: [2022-12-26 21:53:41,984] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 27: [2022-12-26 21:53:41,985] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt. 27: [2022-12-26 21:53:41,985] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_217_mp_rank_00_optim_states.pt 27: [2022-12-26 21:53:41,985] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 30: [2022-12-26 21:53:41,987] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt. 30: [2022-12-26 21:53:41,987] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_245_mp_rank_00_optim_states.pt 30: [2022-12-26 21:53:41,987] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 13: [2022-12-26 21:53:41,991] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt. 13: [2022-12-26 21:53:41,991] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_104_mp_rank_00_optim_states.pt 13: [2022-12-26 21:53:41,991] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 14: [2022-12-26 21:53:41,993] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt. 14: [2022-12-26 21:53:41,993] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_114_mp_rank_00_optim_states.pt 14: [2022-12-26 21:53:41,993] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 31: [2022-12-26 21:53:41,993] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt. 31: [2022-12-26 21:53:41,993] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_252_mp_rank_00_optim_states.pt 31: [2022-12-26 21:53:41,993] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 11: [2022-12-26 21:53:41,996] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt. 11: [2022-12-26 21:53:41,996] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_93_mp_rank_00_optim_states.pt 11: [2022-12-26 21:53:41,996] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 25: [2022-12-26 21:53:41,998] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt. 22: [2022-12-26 21:53:41,998] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:41,998] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_207_mp_rank_00_optim_states.pt 25: [2022-12-26 21:53:41,998] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 22: [2022-12-26 21:53:41,998] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_182_mp_rank_00_optim_states.pt 22: [2022-12-26 21:53:41,998] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 5: [2022-12-26 21:53:41,999] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt. 5: [2022-12-26 21:53:41,999] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_43_mp_rank_00_optim_states.pt 5: [2022-12-26 21:53:41,999] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 10: [2022-12-26 21:53:42,001] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt. 10: [2022-12-26 21:53:42,001] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_81_mp_rank_00_optim_states.pt 10: [2022-12-26 21:53:42,001] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 20: [2022-12-26 21:53:42,002] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt. 20: [2022-12-26 21:53:42,002] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_160_mp_rank_00_optim_states.pt 20: [2022-12-26 21:53:42,002] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 28: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt. 28: [2022-12-26 21:53:42,003] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_229_mp_rank_00_optim_states.pt 28: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 7: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt. 27: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt. 7: [2022-12-26 21:53:42,003] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_56_mp_rank_00_optim_states.pt 9: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt. 27: [2022-12-26 21:53:42,003] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_222_mp_rank_00_optim_states.pt 27: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 7: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 9: [2022-12-26 21:53:42,003] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_77_mp_rank_00_optim_states.pt 9: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 17: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt. 17: [2022-12-26 21:53:42,003] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_137_mp_rank_00_optim_states.pt 17: [2022-12-26 21:53:42,003] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 21: [2022-12-26 21:53:42,004] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt. 21: [2022-12-26 21:53:42,004] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_170_mp_rank_00_optim_states.pt 21: [2022-12-26 21:53:42,005] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 18: [2022-12-26 21:53:42,005] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt. 18: [2022-12-26 21:53:42,005] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_146_mp_rank_00_optim_states.pt 18: [2022-12-26 21:53:42,005] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 31: [2022-12-26 21:53:42,006] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt. 31: [2022-12-26 21:53:42,006] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_249_mp_rank_00_optim_states.pt 31: [2022-12-26 21:53:42,006] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 1: [2022-12-26 21:53:42,006] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. 1: [2022-12-26 21:53:42,006] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt 1: [2022-12-26 21:53:42,006] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 17: [2022-12-26 21:53:42,008] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt. 17: [2022-12-26 21:53:42,008] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_136_mp_rank_00_optim_states.pt 17: [2022-12-26 21:53:42,008] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 15: [2022-12-26 21:53:42,007] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt. 15: [2022-12-26 21:53:42,007] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_124_mp_rank_00_optim_states.pt 15: [2022-12-26 21:53:42,007] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 3: [2022-12-26 21:53:42,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt. 29: [2022-12-26 21:53:42,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt. 29: [2022-12-26 21:53:42,011] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_233_mp_rank_00_optim_states.pt 3: [2022-12-26 21:53:42,011] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt 29: [2022-12-26 21:53:42,011] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 29: [2022-12-26 21:53:42,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt. 3: [2022-12-26 21:53:42,011] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 29: [2022-12-26 21:53:42,011] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_232_mp_rank_00_optim_states.pt 29: [2022-12-26 21:53:42,011] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 1: [2022-12-26 21:53:42,011] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. 1: [2022-12-26 21:53:42,012] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt 1: [2022-12-26 21:53:42,012] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 8: [2022-12-26 21:53:42,012] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt. 8: [2022-12-26 21:53:42,012] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_68_mp_rank_00_optim_states.pt 8: [2022-12-26 21:53:42,012] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 27: [2022-12-26 21:53:42,012] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt. 27: [2022-12-26 21:53:42,012] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_223_mp_rank_00_optim_states.pt 27: [2022-12-26 21:53:42,012] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 5: [2022-12-26 21:53:42,013] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt. 5: [2022-12-26 21:53:42,013] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_40_mp_rank_00_optim_states.pt 5: [2022-12-26 21:53:42,013] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 12: [2022-12-26 21:53:42,014] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt. 20: [2022-12-26 21:53:42,015] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt. 12: [2022-12-26 21:53:42,015] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_96_mp_rank_00_optim_states.pt 12: [2022-12-26 21:53:42,015] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 20: [2022-12-26 21:53:42,015] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_164_mp_rank_00_optim_states.pt 20: [2022-12-26 21:53:42,015] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 6: [2022-12-26 21:53:42,015] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt. 6: [2022-12-26 21:53:42,015] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_51_mp_rank_00_optim_states.pt 6: [2022-12-26 21:53:42,015] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 9: [2022-12-26 21:53:42,015] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt. 9: [2022-12-26 21:53:42,015] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_79_mp_rank_00_optim_states.pt 9: [2022-12-26 21:53:42,015] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 25: [2022-12-26 21:53:42,015] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt. 25: [2022-12-26 21:53:42,015] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_202_mp_rank_00_optim_states.pt 25: [2022-12-26 21:53:42,016] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 11: [2022-12-26 21:53:42,017] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt. 8: [2022-12-26 21:53:42,017] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt. 11: [2022-12-26 21:53:42,017] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_94_mp_rank_00_optim_states.pt 8: [2022-12-26 21:53:42,017] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_70_mp_rank_00_optim_states.pt 11: [2022-12-26 21:53:42,017] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 8: [2022-12-26 21:53:42,017] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 6: [2022-12-26 21:53:42,017] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt. 6: [2022-12-26 21:53:42,017] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_54_mp_rank_00_optim_states.pt 6: [2022-12-26 21:53:42,017] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 7: [2022-12-26 21:53:42,018] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt. 7: [2022-12-26 21:53:42,018] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_63_mp_rank_00_optim_states.pt 7: [2022-12-26 21:53:42,019] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 14: [2022-12-26 21:53:42,019] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt. 14: [2022-12-26 21:53:42,019] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_113_mp_rank_00_optim_states.pt 14: [2022-12-26 21:53:42,019] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 3: [2022-12-26 21:53:42,020] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt. 3: [2022-12-26 21:53:42,020] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt 3: [2022-12-26 21:53:42,020] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 19: [2022-12-26 21:53:42,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt. 19: [2022-12-26 21:53:42,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_154_mp_rank_00_optim_states.pt 19: [2022-12-26 21:53:42,022] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt. 19: [2022-12-26 21:53:42,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 19: [2022-12-26 21:53:42,022] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_155_mp_rank_00_optim_states.pt 19: [2022-12-26 21:53:42,022] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 10: [2022-12-26 21:53:42,023] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt. 10: [2022-12-26 21:53:42,023] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_86_mp_rank_00_optim_states.pt 10: [2022-12-26 21:53:42,023] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 21: [2022-12-26 21:53:42,023] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt. 21: [2022-12-26 21:53:42,023] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_174_mp_rank_00_optim_states.pt 21: [2022-12-26 21:53:42,023] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 12: [2022-12-26 21:53:42,023] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt. 12: [2022-12-26 21:53:42,023] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_97_mp_rank_00_optim_states.pt 12: [2022-12-26 21:53:42,023] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 28: [2022-12-26 21:53:42,024] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt. 28: [2022-12-26 21:53:42,024] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_226_mp_rank_00_optim_states.pt 28: [2022-12-26 21:53:42,024] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 15: [2022-12-26 21:53:42,025] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt. 15: [2022-12-26 21:53:42,025] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_125_mp_rank_00_optim_states.pt 15: [2022-12-26 21:53:42,025] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 15: [2022-12-26 21:53:42,025] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt. 15: [2022-12-26 21:53:42,026] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_121_mp_rank_00_optim_states.pt 15: [2022-12-26 21:53:42,026] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 11: [2022-12-26 21:53:42,040] [INFO] [torch_checkpoint_engine.py:17:save] [Torch] Saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt. 11: [2022-12-26 21:53:42,040] [INFO] [engine.py:3213:_save_zero_checkpoint] bf16_zero checkpoint saved checkpoints_2b855brealtasky/global_step52452/bf16_zero_pp_rank_91_mp_rank_00_optim_states.pt 11: [2022-12-26 21:53:42,040] [INFO] [torch_checkpoint_engine.py:27:commit] [Torch] Checkpoint global_step52452 is ready now! 0: successfully saved checkpoint at iteration 52452 to checkpoints_2b855brealtasky END 2295714: Mon Dec 26 21:54:07 EET 2022