Witllm/finetune/llamafactory/inference_qwen3_lora_sft.yaml

5 lines
186 B
YAML

model_name_or_path: Qwen/Qwen3-4B
adapter_name_or_path: saves/qwen3-4b/lora/sft
template: qwen3
infer_backend: huggingface # choices: [huggingface, vllm, sglang]
trust_remote_code: true