Upload ds_block_large_chinese.sh

ds_block_large_chinese.sh

@@ -0,0 +1,48 @@
+#! /bin/bash
+
+script_path=$(realpath $BASH_SOURCE)
+script_dir=$(dirname $script_path)
+
+config_json="./config/config_block_large_chinese.json"
+gpt_options=" \
+       --block-lm \
+       --task-mask \
+       --bert-prob 0.4 \
+       --gap-sentence-prob 0.3 \
+       --avg-block-length 3 \
+       --gpt-min-ratio 0.25 \
+       --block-mask-prob 0.1 \
+       --short-seq-prob 0.02 \
+       --experiment-name blocklm-large-chinese \
+       --model-parallel-size ${MP_SIZE} \
+       --num-layers 24 \
+       --hidden-size 1024 \
+       --num-attention-heads 16 \
+       --seq-length 512 \
+       --max-position-embeddings 1024 \
+       --save ../model_save/checkpoints/ \
+       --load ../model_save/checkpoints/
+       --log-interval 50 \
+       --eval-interval 1000 \
+       --save-interval 2000 \
+       --train-iters 250000000 \
+       --train-data wudao \
+       --resume-dataloader \
+       --loader-scatter 4 \
+       --no-lazy-loader \
+       --tokenizer-type ChineseSPTokenizer \
+       --fix-command-token \
+       --split 949,50,1 \
+       --distributed-backend nccl \
+       --lr-decay-style cosine \
+       --lr-decay-ratio 0.1 \
+       --lr-decay-iters 200000 \
+       --warmup 0.04 \
+       --checkpoint-activations \
+       --deepspeed-activation-checkpointing \
+       --fp16 \
+"
+gpt_options="${gpt_options}
+               --deepspeed \
+               --deepspeed_config ${config_json} \
+"