llama-factory/config/models/Qwen2-72B-Instruct.yaml

### model
model_name_or_path: Qwen/Qwen2-72B-Instruct
quantization_bit: 4

### method
stage: sft
do_train: true
finetuning_type: lora
lora_target: all

### dataset
dataset: alpaca_mgtv_p2
template: qwen
cutoff_len: 8192
max_samples: 25000
overwrite_cache: true
preprocessing_num_workers: 16

### output
output_dir: saves/Qwen2-72B-Instruct
logging_steps: 5
save_steps: 35
plot_loss: true
# overwrite_output_dir: true

### train
per_device_train_batch_size: 16
gradient_accumulation_steps: 8
learning_rate: 1.0e-4
num_train_epochs: 2.0
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: 35

report_to: wandb
run_name: Qwen2-72B-Instruct_lora_sft