| NPROC_PER_NODE=8 \ | |
| PYTORCH_CUDA_ALLOC_CONF='' \ | |
| swift rlhf \ | |
| --rlhf_type grpo \ | |
| --model Qwen/Qwen2.5-7B \ | |
| --tuner_type full \ | |
| --use_vllm true \ | |
| --vllm_mode colocate \ | |
| --vllm_gpu_memory_utilization 0.5 \ | |
| --vllm_max_model_len 2048 \ | |
| --vllm_tensor_parallel_size 4 \ | |
| --dataset AI-MO/NuminaMath-TIR#5000 \ | |
| --load_from_cache_file true \ | |
| --torch_dtype bfloat16 \ | |
| --num_train_epochs 1 \ | |
| --max_length 2048 \ | |
| --per_device_train_batch_size 4 \ | |
| --per_device_eval_batch_size 4 \ | |
| --gradient_accumulation_steps 8 \ | |
| --eval_steps 1000 \ | |
| --save_steps 1000 \ | |
| --learning_rate 1e-6 \ | |
| --save_total_limit 2 \ | |
| --logging_steps 5 \ | |
| --output_dir output \ | |
| --warmup_ratio 0.05 \ | |
| --dataloader_num_workers 4 \ | |
| --max_completion_length 1024 \ | |
| --reward_funcs accuracy format \ | |
| --num_generations 32 \ | |
| --system examples/train/grpo/prompt.txt \ | |
| --deepspeed zero3 \ | |
| --temperature 1.0 \ | |
| --top_p 1.0 \ | |
| --top_k 80 \ | |
| --attn_impl flash_attn \ | |
| --log_completions true \ | |
| --async_generate false \ | |
| --offload_optimizer true \ | |
| --offload_model true \ | |
| --padding_free true \ | |
| --sequence_parallel_size 4 \ | |
| --dataloader_drop_last true \ | |
| --sleep_level 1 | |