qwen4B_models/qwen_4B_d2_iter3_model/lora.log

[2024-07-15 12:07:05,506] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:07,358] [WARNING] [runner.py:202:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
[2024-07-15 12:07:07,410] [INFO] [runner.py:571:main] cmd = /ML-A100/team/mm/zhangge/anaconda3/envs/improve/bin/python -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=29500 --enable_each_rank_log=None ../../src/train_bash.py --deepspeed ../deepspeed/ds_z3_config.json --stage sft --do_train --model_name_or_path /ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B --dataset qwen_4B_d2_iter3_model --dataset_dir ../../data --template qwen_like --finetuning_type lora --lora_target all --lora_rank 8 --lora_alpha 16 --lora_dropout 0.05 --output_dir /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model --overwrite_cache --overwrite_output_dir --cutoff_len 1024 --preprocessing_num_workers 8 --per_device_train_batch_size 1 --per_device_eval_batch_size 1 --gradient_accumulation_steps 2 --lr_scheduler_type cosine --logging_steps 10 --warmup_steps 20 --save_steps 100 --eval_steps 100 --evaluation_strategy steps --load_best_model_at_end --learning_rate 5e-5 --num_train_epochs 2.0 --max_samples 3000 --val_size 0.1 --ddp_timeout 180000000 --plot_loss --bf16
[2024-07-15 12:07:09,422] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:10,330] [INFO] [launch.py:138:main] 0 NCCL_IB_PCI_RELAXED_ORDERING=1
[2024-07-15 12:07:10,330] [INFO] [launch.py:138:main] 0 NCCL_DEBUG=INFO
[2024-07-15 12:07:10,330] [INFO] [launch.py:138:main] 0 NCCL_SOCKET_IFNAME=eth1
[2024-07-15 12:07:10,330] [INFO] [launch.py:138:main] 0 NCCL_IB_HCA=mlx5_1:1,mlx5_2:1,mlx5_3:1,mlx5_4:1
[2024-07-15 12:07:10,330] [INFO] [launch.py:138:main] 0 NCCL_IB_GID_INDEX=7
[2024-07-15 12:07:10,330] [INFO] [launch.py:138:main] 0 NCCL_IB_RETRY_CNT=7
[2024-07-15 12:07:10,330] [INFO] [launch.py:138:main] 0 NCCL_IB_DISABLE=0
[2024-07-15 12:07:10,330] [INFO] [launch.py:138:main] 0 NCCL_VERSION=2.19.3
[2024-07-15 12:07:10,330] [INFO] [launch.py:138:main] 0 NCCL_IB_TIMEOUT=23
[2024-07-15 12:07:10,330] [INFO] [launch.py:145:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
[2024-07-15 12:07:10,330] [INFO] [launch.py:151:main] nnodes=1, num_local_procs=8, node_rank=0
[2024-07-15 12:07:10,330] [INFO] [launch.py:162:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
[2024-07-15 12:07:10,330] [INFO] [launch.py:163:main] dist_world_size=8
[2024-07-15 12:07:10,330] [INFO] [launch.py:165:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
[2024-07-15 12:07:17,240] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:17,240] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:17,247] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:17,248] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:17,249] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:17,249] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:17,265] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:17,265] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2024-07-15 12:07:19,813] [INFO] [comm.py:637:init_distributed] cdb=None
[2024-07-15 12:07:19,813] [INFO] [comm.py:637:init_distributed] cdb=None
[2024-07-15 12:07:19,813] [INFO] [comm.py:637:init_distributed] cdb=None
[2024-07-15 12:07:19,813] [INFO] [comm.py:637:init_distributed] cdb=None
[2024-07-15 12:07:19,813] [INFO] [comm.py:637:init_distributed] cdb=None
[2024-07-15 12:07:19,813] [INFO] [comm.py:637:init_distributed] cdb=None
[2024-07-15 12:07:19,813] [INFO] [comm.py:637:init_distributed] cdb=None
[2024-07-15 12:07:19,813] [INFO] [comm.py:637:init_distributed] cdb=None
[2024-07-15 12:07:19,813] [INFO] [comm.py:668:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
07/15/2024 12:07:21 - WARNING - llmtuner.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training.
07/15/2024 12:07:21 - INFO - llmtuner.hparams.parser - Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
07/15/2024 12:07:21 - WARNING - llmtuner.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training.
07/15/2024 12:07:21 - INFO - llmtuner.hparams.parser - Process rank: 1, device: cuda:1, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
07/15/2024 12:07:21 - WARNING - llmtuner.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training.
07/15/2024 12:07:21 - INFO - llmtuner.hparams.parser - Process rank: 7, device: cuda:7, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
07/15/2024 12:07:21 - WARNING - llmtuner.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training.
07/15/2024 12:07:21 - INFO - llmtuner.hparams.parser - Process rank: 6, device: cuda:6, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
07/15/2024 12:07:21 - WARNING - llmtuner.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training.
07/15/2024 12:07:21 - INFO - llmtuner.hparams.parser - Process rank: 3, device: cuda:3, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
07/15/2024 12:07:21 - WARNING - llmtuner.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training.
07/15/2024 12:07:21 - INFO - llmtuner.hparams.parser - Process rank: 2, device: cuda:2, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
07/15/2024 12:07:21 - WARNING - llmtuner.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training.
07/15/2024 12:07:21 - INFO - llmtuner.hparams.parser - Process rank: 4, device: cuda:4, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
07/15/2024 12:07:21 - WARNING - llmtuner.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training.
07/15/2024 12:07:21 - INFO - llmtuner.hparams.parser - Process rank: 5, device: cuda:5, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
07/15/2024 12:07:21 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
07/15/2024 12:07:21 - INFO - llmtuner.data.loader - Loading dataset /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_outputs/qwen_4B_iter2_d2_10k_output_filtered_evaluated_filtered.json...
07/15/2024 12:07:21 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json.
07/15/2024 12:07:21 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
07/15/2024 12:07:21 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
07/15/2024 12:07:21 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
07/15/2024 12:07:21 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
07/15/2024 12:07:21 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
07/15/2024 12:07:21 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
07/15/2024 12:07:21 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
t-20240715144706-mwmm8-worker-0:144068:144068 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144068:144068 [0] NCCL INFO Bootstrap : Using eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144068:144068 [0] NCCL INFO cudaDriverVersion 12030
NCCL version 2.18.1+cuda12.1
t-20240715144706-mwmm8-worker-0:144071:144071 [3] NCCL INFO cudaDriverVersion 12030
t-20240715144706-mwmm8-worker-0:144071:144071 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144070:144070 [2] NCCL INFO cudaDriverVersion 12030
t-20240715144706-mwmm8-worker-0:144070:144070 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144074:144074 [6] NCCL INFO cudaDriverVersion 12030
t-20240715144706-mwmm8-worker-0:144073:144073 [5] NCCL INFO cudaDriverVersion 12030
t-20240715144706-mwmm8-worker-0:144074:144074 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144072:144072 [4] NCCL INFO cudaDriverVersion 12030
t-20240715144706-mwmm8-worker-0:144073:144073 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144072:144072 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144071:144071 [3] NCCL INFO Bootstrap : Using eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144075:144075 [7] NCCL INFO cudaDriverVersion 12030
t-20240715144706-mwmm8-worker-0:144075:144075 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144069:144069 [1] NCCL INFO cudaDriverVersion 12030
t-20240715144706-mwmm8-worker-0:144070:144070 [2] NCCL INFO Bootstrap : Using eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144069:144069 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144074:144074 [6] NCCL INFO Bootstrap : Using eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144073:144073 [5] NCCL INFO Bootstrap : Using eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144072:144072 [4] NCCL INFO Bootstrap : Using eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144075:144075 [7] NCCL INFO Bootstrap : Using eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144069:144069 [1] NCCL INFO Bootstrap : Using eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Plugin Path : /opt/hpcx/nccl_rdma_sharp_plugin/lib/libnccl-net.so
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO P2P plugin IBext
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO NCCL_IB_PCI_RELAXED_ORDERING set by environment to 1.
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO NET/IB : Using [0]mlx5_1:1/RoCE [1]mlx5_2:1/RoCE [2]mlx5_3:1/RoCE [3]mlx5_4:1/RoCE [RO]; OOB eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Plugin Path : /opt/hpcx/nccl_rdma_sharp_plugin/lib/libnccl-net.so
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO P2P plugin IBext
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Plugin Path : /opt/hpcx/nccl_rdma_sharp_plugin/lib/libnccl-net.so
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO P2P plugin IBext
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Plugin Path : /opt/hpcx/nccl_rdma_sharp_plugin/lib/libnccl-net.so
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO P2P plugin IBext
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO NCCL_IB_PCI_RELAXED_ORDERING set by environment to 1.
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO NET/IB : Using [0]mlx5_1:1/RoCE [1]mlx5_2:1/RoCE [2]mlx5_3:1/RoCE [3]mlx5_4:1/RoCE [RO]; OOB eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Plugin Path : /opt/hpcx/nccl_rdma_sharp_plugin/lib/libnccl-net.so
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO P2P plugin IBext
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Plugin Path : /opt/hpcx/nccl_rdma_sharp_plugin/lib/libnccl-net.so
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO P2P plugin IBext
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO NCCL_IB_PCI_RELAXED_ORDERING set by environment to 1.
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO NET/IB : Using [0]mlx5_1:1/RoCE [1]mlx5_2:1/RoCE [2]mlx5_3:1/RoCE [3]mlx5_4:1/RoCE [RO]; OOB eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO NCCL_IB_PCI_RELAXED_ORDERING set by environment to 1.
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO NET/IB : Using [0]mlx5_1:1/RoCE [1]mlx5_2:1/RoCE [2]mlx5_3:1/RoCE [3]mlx5_4:1/RoCE [RO]; OOB eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Plugin Path : /opt/hpcx/nccl_rdma_sharp_plugin/lib/libnccl-net.so
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO P2P plugin IBext
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Plugin Path : /opt/hpcx/nccl_rdma_sharp_plugin/lib/libnccl-net.so
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO P2P plugin IBext
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to eth1
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO NCCL_IB_PCI_RELAXED_ORDERING set by environment to 1.
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO NET/IB : Using [0]mlx5_1:1/RoCE [1]mlx5_2:1/RoCE [2]mlx5_3:1/RoCE [3]mlx5_4:1/RoCE [RO]; OOB eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO NCCL_IB_PCI_RELAXED_ORDERING set by environment to 1.
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO NET/IB : Using [0]mlx5_1:1/RoCE [1]mlx5_2:1/RoCE [2]mlx5_3:1/RoCE [3]mlx5_4:1/RoCE [RO]; OOB eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO NCCL_IB_PCI_RELAXED_ORDERING set by environment to 1.
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO NET/IB : Using [0]mlx5_1:1/RoCE [1]mlx5_2:1/RoCE [2]mlx5_3:1/RoCE [3]mlx5_4:1/RoCE [RO]; OOB eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO NCCL_IB_PCI_RELAXED_ORDERING set by environment to 1.
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO NET/IB : Using [0]mlx5_1:1/RoCE [1]mlx5_2:1/RoCE [2]mlx5_3:1/RoCE [3]mlx5_4:1/RoCE [RO]; OOB eth1:172.25.40.22<0>
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,00000000,ffffffff,00000000
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO NVLS multicast support is not available on dev 4
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,00000000,ffffffff,00000000
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO NVLS multicast support is not available on dev 7
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,00000000,ffffffff,00000000
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,00000000,ffffffff,00000000
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO NVLS multicast support is not available on dev 6
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Setting affinity for GPU 0 to ffffffff,00000000,ffffffff
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO NVLS multicast support is not available on dev 0
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO NVLS multicast support is not available on dev 5
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Setting affinity for GPU 2 to ffffffff,00000000,ffffffff
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO NVLS multicast support is not available on dev 2
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Setting affinity for GPU 1 to ffffffff,00000000,ffffffff
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Setting affinity for GPU 3 to ffffffff,00000000,ffffffff
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO NVLS multicast support is not available on dev 1
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO NVLS multicast support is not available on dev 3
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 00/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 01/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 02/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 03/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 04/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 05/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 06/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 07/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 08/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 09/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 10/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 11/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 12/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 13/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 14/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 15/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 00/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 01/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 02/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 00/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 00/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 00/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 00/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 00/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 03/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 00/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 00/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 01/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 01/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 01/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 01/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 04/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 01/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 01/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 01/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 02/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 02/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 02/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 02/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 05/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 02/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 02/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 02/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 03/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 03/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 03/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 03/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 06/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 03/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 03/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 03/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 04/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 04/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 04/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 04/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 07/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 04/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 04/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 05/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 05/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 05/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 05/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 08/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 05/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 04/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 05/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 05/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 06/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 06/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 06/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 06/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 06/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 06/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 09/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 06/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 07/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 07/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 07/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 07/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 07/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 07/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 10/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 07/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 08/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 08/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 08/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 08/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 08/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 11/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 08/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 08/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 09/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 09/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 09/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 09/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 09/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 09/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 12/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 09/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 10/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 10/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 10/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 10/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 10/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 13/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 10/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 10/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 11/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 11/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 11/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 11/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 11/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 14/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 11/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 11/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 12/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 12/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 12/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 12/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 12/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 15/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 12/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 12/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 13/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 13/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 13/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 13/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 13/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 13/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 13/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 14/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 14/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 14/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 14/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 14/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 14/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 15/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 15/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 15/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 15/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 14/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 15/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 15/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Channel 15/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 00/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 01/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 02/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 03/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 04/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 05/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 06/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 07/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 08/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 09/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 10/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 11/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 12/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 13/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 14/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 00/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 01/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Channel 15/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 02/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 00/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 00/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 00/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 00/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 03/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 04/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 05/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 06/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 07/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 08/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 01/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 00/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 01/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 02/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 02/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 01/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 01/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 01/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 03/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 03/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 02/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 02/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 02/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 09/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 10/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 11/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 12/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 13/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 14/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Channel 15/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 03/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 04/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 03/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 04/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 03/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 05/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 04/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 04/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 05/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 04/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 05/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 05/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 05/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 06/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 06/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 06/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 07/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 06/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 06/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 07/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 07/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 07/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 07/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 08/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 08/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 08/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 08/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 08/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 09/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 09/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 09/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 10/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 10/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 09/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 10/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 09/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 11/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 10/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 10/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 11/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 11/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 12/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 12/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 11/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 12/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 11/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 12/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 13/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 13/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 13/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 12/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 13/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 13/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 14/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 14/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 14/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 14/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 14/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Channel 15/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Channel 15/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Channel 15/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Channel 15/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Channel 15/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144070:145402 [2] NCCL INFO comm 0xe31f5d0 rank 2 nranks 8 cudaDev 2 busId 29000 commId 0x107e21748b2da532 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144074:145403 [6] NCCL INFO comm 0xc3b2fb0 rank 6 nranks 8 cudaDev 6 busId e0000 commId 0x107e21748b2da532 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144069:145404 [1] NCCL INFO comm 0xd9a0200 rank 1 nranks 8 cudaDev 1 busId 13000 commId 0x107e21748b2da532 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144075:145407 [7] NCCL INFO comm 0xc5ac280 rank 7 nranks 8 cudaDev 7 busId e4000 commId 0x107e21748b2da532 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144072:145405 [4] NCCL INFO comm 0xd6bbbe0 rank 4 nranks 8 cudaDev 4 busId c5000 commId 0x107e21748b2da532 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144073:145406 [5] NCCL INFO comm 0xd5560f0 rank 5 nranks 8 cudaDev 5 busId ca000 commId 0x107e21748b2da532 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144071:145401 [3] NCCL INFO comm 0xc7c15a0 rank 3 nranks 8 cudaDev 3 busId 2d000 commId 0x107e21748b2da532 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144068:145393 [0] NCCL INFO comm 0xd513680 rank 0 nranks 8 cudaDev 0 busId d000 commId 0x107e21748b2da532 - Init COMPLETE
07/15/2024 12:08:06 - INFO - llmtuner.data.loader - Loading dataset /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_outputs/qwen_4B_iter2_d2_10k_output_filtered_evaluated_filtered.json...
07/15/2024 12:08:06 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json.
07/15/2024 12:08:06 - INFO - llmtuner.data.loader - Loading dataset /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_outputs/qwen_4B_iter2_d2_10k_output_filtered_evaluated_filtered.json...
07/15/2024 12:08:06 - INFO - llmtuner.data.loader - Loading dataset /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_outputs/qwen_4B_iter2_d2_10k_output_filtered_evaluated_filtered.json...
07/15/2024 12:08:06 - INFO - llmtuner.data.loader - Loading dataset /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_outputs/qwen_4B_iter2_d2_10k_output_filtered_evaluated_filtered.json...
07/15/2024 12:08:06 - INFO - llmtuner.data.loader - Loading dataset /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_outputs/qwen_4B_iter2_d2_10k_output_filtered_evaluated_filtered.json...
07/15/2024 12:08:06 - INFO - llmtuner.data.loader - Loading dataset /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_outputs/qwen_4B_iter2_d2_10k_output_filtered_evaluated_filtered.json...
07/15/2024 12:08:06 - INFO - llmtuner.data.loader - Loading dataset /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_outputs/qwen_4B_iter2_d2_10k_output_filtered_evaluated_filtered.json...
07/15/2024 12:08:06 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json.
07/15/2024 12:08:06 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json.
07/15/2024 12:08:06 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json.
07/15/2024 12:08:06 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json.
07/15/2024 12:08:06 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json.
07/15/2024 12:08:06 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json.
input_ids:
[151644, 8948, 198, 2610, 525, 264, 10950, 17847, 13, 151645, 198, 151644, 872, 198, 35127, 279, 6291, 311, 279, 2701, 23606, 25, 220, 19, 87, 481, 220, 21, 284, 220, 16, 15, 151645, 198, 151644, 77091, 198, 87, 284, 220, 19, 151643]
inputs:
<|im_start|>system
You are a helpful assistant.<|im_end|>
<|im_start|>user
Give the solution to the following equation: 4x - 6 = 10<|im_end|>
<|im_start|>assistant
x = 4<|endoftext|>
label_ids:
[-100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, -100, 87, 284, 220, 19, 151643]
labels:
x = 4<|endoftext|>
[2024-07-15 12:09:17,165] [INFO] [partition_parameters.py:349:__exit__] finished initializing model - num_params = 483, num_elems = 3.95B
07/15/2024 12:09:20 - INFO - llmtuner.model.patcher - Gradient checkpointing enabled.
07/15/2024 12:09:20 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
07/15/2024 12:09:20 - INFO - llmtuner.model.utils - Found linear modules: gate_proj,v_proj,o_proj,up_proj,q_proj,down_proj,k_proj
07/15/2024 12:09:20 - INFO - llmtuner.model.patcher - Gradient checkpointing enabled.
07/15/2024 12:09:20 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
07/15/2024 12:09:20 - INFO - llmtuner.model.utils - Found linear modules: gate_proj,down_proj,q_proj,up_proj,o_proj,k_proj,v_proj
07/15/2024 12:09:20 - INFO - llmtuner.model.patcher - Gradient checkpointing enabled.
07/15/2024 12:09:20 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
07/15/2024 12:09:20 - INFO - llmtuner.model.utils - Found linear modules: gate_proj,down_proj,q_proj,o_proj,k_proj,up_proj,v_proj
07/15/2024 12:09:20 - INFO - llmtuner.model.patcher - Gradient checkpointing enabled.
07/15/2024 12:09:20 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
07/15/2024 12:09:20 - INFO - llmtuner.model.utils - Found linear modules: k_proj,v_proj,o_proj,up_proj,down_proj,q_proj,gate_proj
07/15/2024 12:09:20 - INFO - llmtuner.model.patcher - Gradient checkpointing enabled.
07/15/2024 12:09:20 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
07/15/2024 12:09:20 - INFO - llmtuner.model.utils - Found linear modules: gate_proj,up_proj,o_proj,k_proj,v_proj,q_proj,down_proj
07/15/2024 12:09:20 - INFO - llmtuner.model.patcher - Gradient checkpointing enabled.
07/15/2024 12:09:20 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
07/15/2024 12:09:20 - INFO - llmtuner.model.utils - Found linear modules: down_proj,q_proj,o_proj,gate_proj,v_proj,k_proj,up_proj
07/15/2024 12:09:20 - INFO - llmtuner.model.patcher - Gradient checkpointing enabled.
07/15/2024 12:09:20 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
07/15/2024 12:09:20 - INFO - llmtuner.model.utils - Found linear modules: q_proj,gate_proj,v_proj,o_proj,down_proj,up_proj,k_proj
07/15/2024 12:09:21 - INFO - llmtuner.model.patcher - Gradient checkpointing enabled.
07/15/2024 12:09:21 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
07/15/2024 12:09:21 - INFO - llmtuner.model.utils - Found linear modules: q_proj,gate_proj,down_proj,up_proj,v_proj,o_proj,k_proj
07/15/2024 12:09:29 - INFO - llmtuner.model.loader - trainable params: 15646720 || all params: 3966016000 || trainable%: 0.3945
07/15/2024 12:09:29 - INFO - llmtuner.model.loader - trainable params: 15646720 || all params: 3966016000 || trainable%: 0.3945
07/15/2024 12:09:29 - INFO - llmtuner.model.loader - trainable params: 15646720 || all params: 3966016000 || trainable%: 0.3945
07/15/2024 12:09:30 - INFO - llmtuner.model.loader - trainable params: 15646720 || all params: 3966016000 || trainable%: 0.3945
07/15/2024 12:09:30 - INFO - llmtuner.model.loader - trainable params: 15646720 || all params: 3966016000 || trainable%: 0.3945
07/15/2024 12:09:30 - INFO - llmtuner.model.loader - trainable params: 15646720 || all params: 3966016000 || trainable%: 0.3945
07/15/2024 12:09:30 - INFO - llmtuner.model.loader - trainable params: 15646720 || all params: 3966016000 || trainable%: 0.3945
07/15/2024 12:09:30 - INFO - llmtuner.model.loader - trainable params: 15646720 || all params: 3966016000 || trainable%: 0.3945
[2024-07-15 12:09:31,165] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed info: version=0.13.0, git-hash=unknown, git-branch=unknown
[2024-07-15 12:09:31,210] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed Flops Profiler Enabled: False
[2024-07-15 12:09:31,215] [INFO] [logging.py:96:log_dist] [Rank 0] Using client Optimizer as basic optimizer
[2024-07-15 12:09:31,215] [INFO] [logging.py:96:log_dist] [Rank 0] Removing param_group that has no 'params' in the basic Optimizer
[2024-07-15 12:09:31,268] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed Basic Optimizer = AdamW
[2024-07-15 12:09:31,268] [INFO] [utils.py:56:is_zero_supported_optimizer] Checking ZeRO support for optimizer=AdamW type=<class 'torch.optim.adamw.AdamW'>
[2024-07-15 12:09:31,268] [INFO] [logging.py:96:log_dist] [Rank 0] Creating fp16 ZeRO stage 3 optimizer, MiCS is enabled False, Hierarchical params gather False
[2024-07-15 12:09:31,268] [INFO] [logging.py:96:log_dist] [Rank 0] Creating torch.bfloat16 ZeRO stage 3 optimizer
[2024-07-15 12:09:31,475] [INFO] [utils.py:791:see_memory_usage] Stage 3 initialize beginning
[2024-07-15 12:09:31,476] [INFO] [utils.py:792:see_memory_usage] MA 1.61 GB         Max_MA 3.03 GB         CA 1.72 GB         Max_CA 3 GB 
[2024-07-15 12:09:31,477] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 7.95 GB, percent = 0.4%
[2024-07-15 12:09:31,486] [INFO] [stage3.py:128:__init__] Reduce bucket size 6553600
[2024-07-15 12:09:31,486] [INFO] [stage3.py:129:__init__] Prefetch bucket size 5898240
[2024-07-15 12:09:31,667] [INFO] [utils.py:791:see_memory_usage] DeepSpeedZeRoOffload initialize [begin]
[2024-07-15 12:09:31,668] [INFO] [utils.py:792:see_memory_usage] MA 1.61 GB         Max_MA 1.61 GB         CA 1.72 GB         Max_CA 2 GB 
[2024-07-15 12:09:31,669] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 7.95 GB, percent = 0.4%
Parameter Offload: Total persistent parameters: 9525760 in 641 params
[2024-07-15 12:09:32,098] [INFO] [utils.py:791:see_memory_usage] DeepSpeedZeRoOffload initialize [end]
[2024-07-15 12:09:32,099] [INFO] [utils.py:792:see_memory_usage] MA 1.59 GB         Max_MA 1.61 GB         CA 1.72 GB         Max_CA 2 GB 
[2024-07-15 12:09:32,099] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 7.95 GB, percent = 0.4%
[2024-07-15 12:09:32,278] [INFO] [utils.py:791:see_memory_usage] Before creating fp16 partitions
[2024-07-15 12:09:32,279] [INFO] [utils.py:792:see_memory_usage] MA 1.59 GB         Max_MA 1.59 GB         CA 1.72 GB         Max_CA 2 GB 
[2024-07-15 12:09:32,279] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 7.95 GB, percent = 0.4%
[2024-07-15 12:09:32,715] [INFO] [utils.py:791:see_memory_usage] After creating fp16 partitions: 1
[2024-07-15 12:09:32,715] [INFO] [utils.py:792:see_memory_usage] MA 1.59 GB         Max_MA 1.59 GB         CA 1.67 GB         Max_CA 2 GB 
[2024-07-15 12:09:32,716] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 7.98 GB, percent = 0.4%
[2024-07-15 12:09:32,931] [INFO] [utils.py:791:see_memory_usage] Before creating fp32 partitions
[2024-07-15 12:09:32,931] [INFO] [utils.py:792:see_memory_usage] MA 1.59 GB         Max_MA 1.59 GB         CA 1.67 GB         Max_CA 2 GB 
[2024-07-15 12:09:32,932] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 7.98 GB, percent = 0.4%
[2024-07-15 12:09:33,115] [INFO] [utils.py:791:see_memory_usage] After creating fp32 partitions
[2024-07-15 12:09:33,115] [INFO] [utils.py:792:see_memory_usage] MA 1.6 GB         Max_MA 1.6 GB         CA 1.67 GB         Max_CA 2 GB 
[2024-07-15 12:09:33,116] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 7.98 GB, percent = 0.4%
[2024-07-15 12:09:33,308] [INFO] [utils.py:791:see_memory_usage] Before initializing optimizer states
[2024-07-15 12:09:33,308] [INFO] [utils.py:792:see_memory_usage] MA 1.6 GB         Max_MA 1.6 GB         CA 1.67 GB         Max_CA 2 GB 
[2024-07-15 12:09:33,309] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 8.21 GB, percent = 0.4%
[2024-07-15 12:09:33,550] [INFO] [utils.py:791:see_memory_usage] After initializing optimizer states
[2024-07-15 12:09:33,550] [INFO] [utils.py:792:see_memory_usage] MA 1.61 GB         Max_MA 1.63 GB         CA 1.71 GB         Max_CA 2 GB 
[2024-07-15 12:09:33,551] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 8.24 GB, percent = 0.4%
[2024-07-15 12:09:33,551] [INFO] [stage3.py:482:_setup_for_real_optimizer] optimizer state initialized
[2024-07-15 12:09:33,970] [INFO] [utils.py:791:see_memory_usage] After initializing ZeRO optimizer
[2024-07-15 12:09:33,971] [INFO] [utils.py:792:see_memory_usage] MA 1.63 GB         Max_MA 1.63 GB         CA 1.71 GB         Max_CA 2 GB 
[2024-07-15 12:09:33,971] [INFO] [utils.py:799:see_memory_usage] CPU Virtual Memory:  used = 8.44 GB, percent = 0.4%
[2024-07-15 12:09:33,971] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed Final Optimizer = AdamW
[2024-07-15 12:09:33,971] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed using client LR scheduler
[2024-07-15 12:09:33,971] [INFO] [logging.py:96:log_dist] [Rank 0] DeepSpeed LR Scheduler = None
[2024-07-15 12:09:33,971] [INFO] [logging.py:96:log_dist] [Rank 0] step=0, skipped=0, lr=[0.0], mom=[(0.9, 0.999)]
[2024-07-15 12:09:33,976] [INFO] [config.py:984:print] DeepSpeedEngine configuration:
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   activation_checkpointing_config  {
    "partition_activations": false, 
    "contiguous_memory_optimization": false, 
    "cpu_checkpointing": false, 
    "number_checkpoints": null, 
    "synchronize_checkpoint_boundary": false, 
    "profile": false
}
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   aio_config ................... {'block_size': 1048576, 'queue_depth': 8, 'thread_count': 1, 'single_submit': False, 'overlap_events': True}
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   amp_enabled .................. False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   amp_params ................... False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   autotuning_config ............ {
    "enabled": false, 
    "start_step": null, 
    "end_step": null, 
    "metric_path": null, 
    "arg_mappings": null, 
    "metric": "throughput", 
    "model_info": null, 
    "results_dir": "autotuning_results", 
    "exps_dir": "autotuning_exps", 
    "overwrite": true, 
    "fast": true, 
    "start_profile_step": 3, 
    "end_profile_step": 5, 
    "tuner_type": "gridsearch", 
    "tuner_early_stopping": 5, 
    "tuner_num_trials": 50, 
    "model_info_path": null, 
    "mp_size": 1, 
    "max_train_batch_size": null, 
    "min_train_batch_size": 1, 
    "max_train_micro_batch_size_per_gpu": 1.024000e+03, 
    "min_train_micro_batch_size_per_gpu": 1, 
    "num_tuning_micro_batch_sizes": 3
}
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   bfloat16_enabled ............. True
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   checkpoint_parallel_write_pipeline  False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   checkpoint_tag_validation_enabled  True
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   checkpoint_tag_validation_fail  False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   comms_config ................. <deepspeed.comm.config.DeepSpeedCommsConfig object at 0x7f01edfbf370>
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   communication_data_type ...... None
[2024-07-15 12:09:33,977] [INFO] [config.py:988: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_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'layer_reduction': {'enabled': False}}
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   curriculum_enabled_legacy .... False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   curriculum_params_legacy ..... False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   data_efficiency_config ....... {'enabled': False, 'seed': 1234, 'data_sampling': {'enabled': False, 'num_epochs': 1000, 'num_workers': 0, 'curriculum_learning': {'enabled': False}}, 'data_routing': {'enabled': False, 'random_ltd': {'enabled': False, 'layer_token_lr_schedule': {'enabled': False}}}}
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   data_efficiency_enabled ...... False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   dataloader_drop_last ......... False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   disable_allgather ............ False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   dump_state ................... False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   dynamic_loss_scale_args ...... None
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   eigenvalue_enabled ........... False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   eigenvalue_gas_boundary_resolution  1
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   eigenvalue_layer_name ........ bert.encoder.layer
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   eigenvalue_layer_num ......... 0
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   eigenvalue_max_iter .......... 100
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   eigenvalue_stability ......... 1e-06
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   eigenvalue_tol ............... 0.01
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   eigenvalue_verbose ........... False
[2024-07-15 12:09:33,977] [INFO] [config.py:988:print]   elasticity_enabled ........... False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   flops_profiler_config ........ {
    "enabled": false, 
    "recompute_fwd_factor": 0.0, 
    "profile_step": 1, 
    "module_depth": -1, 
    "top_modules": 1, 
    "detailed": true, 
    "output_file": null
}
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   fp16_auto_cast ............... None
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   fp16_enabled ................. False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   fp16_master_weights_and_gradients  False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   global_rank .................. 0
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   grad_accum_dtype ............. None
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   gradient_accumulation_steps .. 2
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   gradient_clipping ............ 1.0
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   gradient_predivide_factor .... 1.0
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   graph_harvesting ............. False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   hybrid_engine ................ enabled=False max_out_tokens=512 inference_tp_size=1 release_inference_cache=False pin_parameters=True tp_gather_partition_size=8
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   initial_dynamic_scale ........ 1
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   load_universal_checkpoint .... False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   loss_scale ................... 1.0
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   memory_breakdown ............. False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   mics_hierarchial_params_gather  False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   mics_shard_size .............. -1
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   monitor_config ............... tensorboard=TensorBoardConfig(enabled=False, output_path='', job_name='DeepSpeedJobName') wandb=WandbConfig(enabled=False, group=None, team=None, project='deepspeed') csv_monitor=CSVConfig(enabled=False, output_path='', job_name='DeepSpeedJobName') enabled=False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   nebula_config ................ {
    "enabled": false, 
    "persistent_storage_path": null, 
    "persistent_time_interval": 100, 
    "num_of_version_in_retention": 2, 
    "enable_nebula_load": true, 
    "load_path": null
}
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   optimizer_legacy_fusion ...... False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   optimizer_name ............... None
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   optimizer_params ............. None
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   pipeline ..................... {'stages': 'auto', 'partition': 'best', 'seed_layers': False, 'activation_checkpoint_interval': 0, 'pipe_partitioned': True, 'grad_partitioned': True}
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   pld_enabled .................. False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   pld_params ................... False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   prescale_gradients ........... False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   scheduler_name ............... None
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   scheduler_params ............. None
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   seq_parallel_communication_data_type  torch.float32
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   sparse_attention ............. None
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   sparse_gradients_enabled ..... False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   steps_per_print .............. inf
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   train_batch_size ............. 16
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   train_micro_batch_size_per_gpu  1
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   use_data_before_expert_parallel_  False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   use_node_local_storage ....... False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   wall_clock_breakdown ......... False
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   weight_quantization_config ... None
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   world_size ................... 8
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   zero_allow_untested_optimizer  True
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   zero_config .................. stage=3 contiguous_gradients=True reduce_scatter=True reduce_bucket_size=6553600 use_multi_rank_bucket_allreduce=True allgather_partitions=True allgather_bucket_size=500,000,000 overlap_comm=True 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=5898240 param_persistence_threshold=25600 model_persistence_threshold=sys.maxsize max_live_parameters=1000000000 max_reuse_distance=1000000000 gather_16bit_weights_on_model_save=True stage3_gather_fp16_weights_on_model_save=False ignore_unused_parameters=True legacy_stage1=False round_robin_gradients=False zero_hpz_partition_size=1 zero_quantized_weights=False zero_quantized_nontrainable_weights=False zero_quantized_gradients=False mics_shard_size=-1 mics_hierarchical_params_gather=False memory_efficient_linear=True pipeline_loading_checkpoint=False override_module_apply=True
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   zero_enabled ................. True
[2024-07-15 12:09:33,978] [INFO] [config.py:988:print]   zero_force_ds_cpu_optimizer .. True
[2024-07-15 12:09:33,979] [INFO] [config.py:988:print]   zero_optimization_stage ...... 3
[2024-07-15 12:09:33,979] [INFO] [config.py:974:print_user_config]   json = {
    "train_batch_size": 16, 
    "train_micro_batch_size_per_gpu": 1, 
    "gradient_accumulation_steps": 2, 
    "gradient_clipping": 1.0, 
    "zero_allow_untested_optimizer": true, 
    "fp16": {
        "enabled": false, 
        "loss_scale": 0, 
        "loss_scale_window": 1000, 
        "initial_scale_power": 16, 
        "hysteresis": 2, 
        "min_loss_scale": 1
    }, 
    "bf16": {
        "enabled": true
    }, 
    "zero_optimization": {
        "stage": 3, 
        "overlap_comm": true, 
        "contiguous_gradients": true, 
        "sub_group_size": 1.000000e+09, 
        "reduce_bucket_size": 6.553600e+06, 
        "stage3_prefetch_bucket_size": 5.898240e+06, 
        "stage3_param_persistence_threshold": 2.560000e+04, 
        "stage3_max_live_parameters": 1.000000e+09, 
        "stage3_max_reuse_distance": 1.000000e+09, 
        "stage3_gather_16bit_weights_on_model_save": true
    }, 
    "steps_per_print": inf
}
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Using network IBext
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,00000000,ffffffff,00000000
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,00000000,ffffffff,00000000
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,00000000,ffffffff,00000000
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Setting affinity for GPU 3 to ffffffff,00000000,ffffffff
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO NVLS multicast support is not available on dev 3
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,00000000,ffffffff,00000000
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO NVLS multicast support is not available on dev 6
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO NVLS multicast support is not available on dev 5
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Setting affinity for GPU 2 to ffffffff,00000000,ffffffff
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO NVLS multicast support is not available on dev 7
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Setting affinity for GPU 1 to ffffffff,00000000,ffffffff
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Setting affinity for GPU 0 to ffffffff,00000000,ffffffff
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO NVLS multicast support is not available on dev 2
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO NVLS multicast support is not available on dev 1
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO NVLS multicast support is not available on dev 0
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO NVLS multicast support is not available on dev 4
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 00/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 01/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 02/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 03/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 04/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 05/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 06/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 07/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 08/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 09/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 10/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 11/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 12/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 13/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 14/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 15/16 :    0   1   2   3   4   5   6   7
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO P2P Chunksize set to 524288
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 00/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 01/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 02/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 03/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 04/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 05/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 06/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 07/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 08/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 09/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 10/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 11/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 12/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 00/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 01/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 02/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 03/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 04/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 05/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 13/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 14/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 15/0 : 4[c5000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 06/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 07/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 08/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 09/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 10/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 11/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 12/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 13/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 14/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 15/0 : 6[e0000] -> 7[e4000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 00/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 00/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 00/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 00/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 00/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 00/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 01/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 01/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 01/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 01/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 01/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 01/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 02/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 02/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 02/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 03/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 03/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 02/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 04/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 02/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 04/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 03/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 03/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 03/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 04/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 05/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 06/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 02/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 05/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 05/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 03/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 06/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 04/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 06/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 04/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 04/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 07/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 07/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 07/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 05/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 05/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 05/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 08/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 06/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 08/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 06/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 08/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 06/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 09/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 07/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 09/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 07/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 09/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 07/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 10/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 08/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 10/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 09/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 11/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 11/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 10/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 08/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 08/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 10/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 12/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 12/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 11/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 09/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 09/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 11/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 12/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 13/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 13/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 14/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 13/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 14/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Channel 15/0 : 0[d000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 12/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 10/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 10/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 15/0 : 2[29000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 14/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 13/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 11/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 11/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 15/0 : 1[13000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 12/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 12/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 14/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 13/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 15/0 : 3[2d000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 13/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 14/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 14/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 15/0 : 7[e4000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 15/0 : 5[ca000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Connected all rings
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 00/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 01/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 02/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 03/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 04/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 05/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 00/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 01/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 02/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 03/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 06/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 07/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 04/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 05/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 06/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 07/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 08/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 09/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 00/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 01/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 02/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 03/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 04/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 08/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 09/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 10/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 11/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 12/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 13/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 14/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Channel 15/0 : 1[13000] -> 0[d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 10/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 00/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 11/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 01/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 05/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 06/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 07/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 08/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 09/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 10/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 11/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 12/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 00/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 02/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 13/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 03/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 04/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 12/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 13/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 14/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Channel 15/0 : 2[29000] -> 1[13000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 01/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 02/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 03/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 00/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 04/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 14/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 01/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 05/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 05/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Channel 15/0 : 7[e4000] -> 6[e0000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 02/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 06/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 03/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 06/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 00/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 04/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 07/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 01/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 05/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 07/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 08/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 02/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 06/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 08/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 09/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 09/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 07/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 03/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 10/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 08/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 04/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 10/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 11/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 09/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 05/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 11/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 12/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 13/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 14/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 10/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 12/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 06/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 13/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 11/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 07/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 14/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 12/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 08/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Channel 15/0 : 3[2d000] -> 2[29000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 13/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 14/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Channel 15/0 : 6[e0000] -> 5[ca000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 09/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Channel 15/0 : 5[ca000] -> 4[c5000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 10/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 11/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 12/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 13/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 14/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Channel 15/0 : 4[c5000] -> 3[2d000] via P2P/IPC/read
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO Connected all trees
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO 16 coll channels, 0 nvls channels, 16 p2p channels, 16 p2p channels per peer
t-20240715144706-mwmm8-worker-0:144074:147008 [6] NCCL INFO comm 0x7ff3c8cce310 rank 6 nranks 8 cudaDev 6 busId e0000 commId 0xfba2b6f962dccd22 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144073:147011 [5] NCCL INFO comm 0x7f9db0cc5f40 rank 5 nranks 8 cudaDev 5 busId ca000 commId 0xfba2b6f962dccd22 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144068:147006 [0] NCCL INFO comm 0x7efeb8d0d350 rank 0 nranks 8 cudaDev 0 busId d000 commId 0xfba2b6f962dccd22 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144072:147012 [4] NCCL INFO comm 0x7f0cd0cc33b0 rank 4 nranks 8 cudaDev 4 busId c5000 commId 0xfba2b6f962dccd22 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144071:147009 [3] NCCL INFO comm 0x7f793e9a76c0 rank 3 nranks 8 cudaDev 3 busId 2d000 commId 0xfba2b6f962dccd22 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144075:147010 [7] NCCL INFO comm 0x7fae32a4d5a0 rank 7 nranks 8 cudaDev 7 busId e4000 commId 0xfba2b6f962dccd22 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144070:147007 [2] NCCL INFO comm 0x7f38a8ccafb0 rank 2 nranks 8 cudaDev 2 busId 29000 commId 0xfba2b6f962dccd22 - Init COMPLETE
t-20240715144706-mwmm8-worker-0:144069:147013 [1] NCCL INFO comm 0x7ef940d72ba0 rank 1 nranks 8 cudaDev 1 busId 13000 commId 0xfba2b6f962dccd22 - Init COMPLETE
{'loss': 1.5243, 'grad_norm': 2.836832841758166, 'learning_rate': 2.5e-05, 'epoch': 0.13}
{'loss': 1.2748, 'grad_norm': 1.6973649342513468, 'learning_rate': 5e-05, 'epoch': 0.26}
{'loss': 0.9, 'grad_norm': 2.341722149848631, 'learning_rate': 4.929528920808854e-05, 'epoch': 0.39}
{'loss': 0.8365, 'grad_norm': 2.965102617927073, 'learning_rate': 4.722088621637309e-05, 'epoch': 0.52}
{'loss': 0.7014, 'grad_norm': 1.0955355901170196, 'learning_rate': 4.389373935885646e-05, 'epoch': 0.65}
{'loss': 0.8207, 'grad_norm': 1.7442390587113106, 'learning_rate': 3.9501422739279956e-05, 'epoch': 0.78}
{'loss': 0.8187, 'grad_norm': 3.2816271852725056, 'learning_rate': 3.4291561391508185e-05, 'epoch': 0.92}
{'loss': 0.7125, 'grad_norm': 1.7842302227509361, 'learning_rate': 2.8557870956832132e-05, 'epoch': 1.05}
{'loss': 0.6435, 'grad_norm': 0.673806281752005, 'learning_rate': 2.2623598917395438e-05, 'epoch': 1.18}
{'loss': 0.6777, 'grad_norm': 2.5479983295530624, 'learning_rate': 1.682330091706446e-05, 'epoch': 1.31}
{'eval_loss': 0.7803115844726562, 'eval_runtime': 12.4961, 'eval_samples_per_second': 10.883, 'eval_steps_per_second': 1.36, 'epoch': 1.31}
[2024-07-15 12:15:17,486] [INFO] [logging.py:96:log_dist] [Rank 0] [Torch] Checkpoint global_step100 is about to be saved!
[2024-07-15 12:15:17,516] [INFO] [logging.py:96:log_dist] [Rank 0] Saving model checkpoint: /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/zero_pp_rank_0_mp_rank_00_model_states.pt
[2024-07-15 12:15:17,516] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/zero_pp_rank_0_mp_rank_00_model_states.pt...
[2024-07-15 12:15:17,542] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/zero_pp_rank_0_mp_rank_00_model_states.pt.
[2024-07-15 12:15:17,553] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt...
[2024-07-15 12:15:17,591] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt.
[2024-07-15 12:15:17,592] [INFO] [engine.py:3477:_save_zero_checkpoint] zero checkpoint saved /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
[2024-07-15 12:15:17,634] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step100 is ready now!
{'loss': 0.6782, 'grad_norm': 1.1151334522530318, 'learning_rate': 1.148397956361007e-05, 'epoch': 1.44}
{'loss': 0.67, 'grad_norm': 1.0655316451808559, 'learning_rate': 6.906649047373246e-06, 'epoch': 1.57}
{'loss': 0.6896, 'grad_norm': 1.3976425544779711, 'learning_rate': 3.3493649053890326e-06, 'epoch': 1.7}
{'loss': 0.6471, 'grad_norm': 0.9518115917493006, 'learning_rate': 1.0126756596375686e-06, 'epoch': 1.83}
{'loss': 0.6339, 'grad_norm': 1.1220129037148685, 'learning_rate': 2.831652042480093e-08, 'epoch': 1.96}
[2024-07-15 12:17:59,109] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/zero_pp_rank_0_mp_rank_00_model_states.pt...
[2024-07-15 12:17:59,124] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/zero_pp_rank_0_mp_rank_00_model_states.pt.
[2024-07-15 12:17:59,125] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/zero_pp_rank_0_mp_rank_00_model_states.pt...
[2024-07-15 12:17:59,138] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/zero_pp_rank_0_mp_rank_00_model_states.pt.
[2024-07-15 12:17:59,166] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt...
[2024-07-15 12:17:59,178] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/checkpoint-100/global_step100/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt.
[2024-07-15 12:17:59,178] [INFO] [engine.py:3019:_get_all_zero_checkpoint_state_dicts] successfully read 8 ZeRO state_dicts for rank 0
[2024-07-15 12:17:59,190] [INFO] [engine.py:2951:_load_zero_checkpoint] loading 8 zero partition checkpoints for rank 0
{'train_runtime': 505.2345, 'train_samples_per_second': 4.822, 'train_steps_per_second': 0.301, 'train_loss': 0.8229627185746243, 'epoch': 1.99}
***** train metrics *****
  epoch                    =     1.9869
  total_flos               =    18895GF
  train_loss               =      0.823
  train_runtime            = 0:08:25.23
  train_samples_per_second =      4.822
  train_steps_per_second   =      0.301
Figure saved at: /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/training_loss.png
Figure saved at: /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model/training_eval_loss.png
***** eval metrics *****
  epoch                   =     1.9869
  eval_loss               =     0.7803
  eval_runtime            = 0:00:08.39
  eval_samples_per_second =     16.208
  eval_steps_per_second   =      2.026
[2024-07-15 12:18:14,030] [INFO] [launch.py:347:main] Process 144072 exits successfully.
[2024-07-15 12:18:14,031] [INFO] [launch.py:347:main] Process 144070 exits successfully.
[2024-07-15 12:18:15,032] [INFO] [launch.py:347:main] Process 144071 exits successfully.
[2024-07-15 12:18:15,032] [INFO] [launch.py:347:main] Process 144073 exits successfully.
[2024-07-15 12:18:15,032] [INFO] [launch.py:347:main] Process 144069 exits successfully.
[2024-07-15 12:18:15,033] [INFO] [launch.py:347:main] Process 144074 exits successfully.
[2024-07-15 12:18:15,033] [INFO] [launch.py:347:main] Process 144075 exits successfully.
[2024-07-15 12:18:15,033] [INFO] [launch.py:347:main] Process 144068 exits successfully.