### model
model_name_or_path: /gemini/code/Qwen2-1.5B
trust_remote_code: true  # 必须开启

### method
stage: sft
do_train: true
finetuning_type: full
deepspeed: /gemini/code/full fine-tuning/ds_z3_config.json

### dataset
dataset_dir: /gemini/code/dataset
dataset: qwen_dataset
template: qwen
cutoff_len: 1024
overwrite_cache: true
preprocessing_num_workers: 16

### output
output_dir: saves/full
logging_steps: 10
save_steps: 100
plot_loss: true
overwrite_output_dir: true

### train
per_device_train_batch_size: 1
gradient_accumulation_steps: 16
learning_rate: 1.0e-5
num_train_epochs: 1.0
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: true
ddp_timeout: 180000000

### eval
val_size: 0.1
per_device_eval_batch_size: 1
eval_strategy: steps
eval_steps: 500