全量微调后，如何起预测服务？目前只看到lora微调后起预测服务的示例

[JoyceChen36]

Reminder

• I have read the README and searched the existing issues.

System Info

• llamafactory version: 0.9.2.dev0
• Platform: Linux-5.10.0-1.0.0.41-x86_64-with-glibc2.17
• Python version: 3.11.11
• PyTorch version: 2.5.0+cu121 (GPU)
• Transformers version: 4.46.1
• Datasets version: 3.1.0
• Accelerate version: 1.0.1
• PEFT version: 0.12.0
• TRL version: 0.9.6
• GPU type: NVIDIA A100-PCIE-40GB
• DeepSpeed version: 0.15.4
• vLLM version: 0.6.3.post1

Reproduction

model_name_or_path: model_sft/
template: qwen
finetuning_type: full

目前sft完的数据格式为：
.
├── added_tokens.json
├── all_results.json
├── checkpoint-1267
├── config.json
├── generation_config.json
├── global_step1267
├── latest
├── merges.txt
├── README.md
├── special_tokens_map.json
├── tokenizer_config.json
├── tokenizer.json
├── trainer_log.jsonl
├── trainer_state.json
├── training_args.bin
├── training_loss.png
├── train_results.json
├── vocab.json
└── zero_to_fp32.py

Expected behavior

No response

Others

No response

[hiyouga]

https://github.com/hiyouga/LLaMA-Factory/blob/main/examples/inference/llama3_full_sft.yaml