calahealthgpt / scripts /test_train.sh
alexshengzhili's picture
Upload folder using huggingface_hub
e72aedf
raw
history blame
719 Bytes
torchrun --nproc_per_node=4 --master_port=20001 fastchat/train/train.py \
--model_name_or_path ~/model_weights/llama-7b \
--data_path ~/datasets/sampled.json \
--fp16 True \
--output_dir output_vicuna \
--num_train_epochs 3 \
--per_device_train_batch_size 2 \
--per_device_eval_batch_size 2 \
--gradient_accumulation_steps 1 \
--evaluation_strategy "steps" \
--eval_steps 2 \
--save_strategy "steps" \
--save_steps 1200 \
--learning_rate 2e-5 \
--weight_decay 0. \
--warmup_ratio 0.03 \
--lr_scheduler_type "cosine" \
--logging_steps 1 \
--tf32 False \
--model_max_length 2048 \
--gradient_checkpointing True \
--lazy_preprocess True