From 4c569f6a31af0e31a12c6b926e9e43e1ea6d7ee0 Mon Sep 17 00:00:00 2001 From: lwaekfjlk <1125027232@qq.com> Date: Thu, 21 Sep 2023 21:05:27 +0000 Subject: [PATCH] enable dummy running --- dummy_run.sh | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 dummy_run.sh diff --git a/dummy_run.sh b/dummy_run.sh new file mode 100644 index 00000000..dc419c48 --- /dev/null +++ b/dummy_run.sh @@ -0,0 +1,24 @@ +torchrun --nproc_per_node=4 --master_port=20001 fastchat/train/train_mem.py \ + --model_name_or_path ./vicuna-7b \ + --data_path data/dummy_conversation.json \ + --bf16 True \ + --output_dir output_vicuna \ + --num_train_epochs 3 \ + --per_device_train_batch_size 2 \ + --per_device_eval_batch_size 2 \ + --gradient_accumulation_steps 16 \ + --evaluation_strategy "no" \ + --save_strategy "steps" \ + --save_steps 1200 \ + --save_total_limit 10 \ + --learning_rate 2e-5 \ + --weight_decay 0. \ + --warmup_ratio 0.03 \ + --lr_scheduler_type "cosine" \ + --logging_steps 1 \ + --fsdp "full_shard auto_wrap" \ + --fsdp_transformer_layer_cls_to_wrap 'LlamaDecoderLayer' \ + --tf32 True \ + --model_max_length 2048 \ + --gradient_checkpointing True \ + --lazy_preprocess True