hailin
/
jd_train
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

.

This commit is contained in:
hailin 2025-08-29 15:07:58 +08:00
parent b83479b09e
commit 890506c505
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
mm-zero3.sh
View File

@ -1,9 +1,9 @@
deepspeed --hostfile hostfile \ deepspeed --hostfile hostfile \
--num_nodes 6 --num_gpus 4 \ --num_nodes 6 --num_gpus 4 \
/home/test/jd_train/train_sft_ds.py \ /home/test/jd_train/train_sft_ds.py \
--model_name_or_path /home/test/Qwen3-8B \ --model_name_or_path /home/test/Qwen3-32B \
--data_glob "/home/test/datasets/my_corpus/train.jsonl" \ --data_glob "/home/test/datasets/my_corpus/train.jsonl" \
--output_dir /home/test/checkpoints/q3-8b-ds4 \ --output_dir /home/test/checkpoints/q3-32b-ds4 \
--seq_len 512 \ --seq_len 512 \
--per_device_train_batch_size 1 \ --per_device_train_batch_size 1 \
--gradient_accumulation_steps 1 \ --gradient_accumulation_steps 1 \