diff --git a/.gitignore b/.gitignore index 39e9066..9a7fbea 100644 --- a/.gitignore +++ b/.gitignore @@ -9,3 +9,5 @@ build .DS_Store /private/ +output_qwen/ +output_qwen*/ \ No newline at end of file diff --git a/finetune/finetune_lora_single_gpu.sh b/finetune/finetune_lora_single_gpu.sh index 972bcf0..7c99e4b 100644 --- a/finetune/finetune_lora_single_gpu.sh +++ b/finetune/finetune_lora_single_gpu.sh @@ -41,7 +41,7 @@ python finetune.py \ --data_path $DATA \ --bf16 True \ --output_dir output_qwen \ - --num_train_epochs 5 \ + --num_train_epochs 16 \ --per_device_train_batch_size 2 \ --per_device_eval_batch_size 1 \ --gradient_accumulation_steps 8 \ @@ -56,7 +56,7 @@ python finetune.py \ --lr_scheduler_type "cosine" \ --logging_steps 1 \ --report_to "none" \ - --model_max_length 512 \ + --model_max_length 3072 \ --lazy_preprocess True \ --gradient_checkpointing \ --use_lora