search

modelscope / ms-swift

Use PEFT or Full-parameter to finetune 400+ LLMs or 100+ MLLMs. (LLM: Qwen2.5, Llama3.2, GLM4, Internlm2.5, Yi1.5, Mistral, Baichuan2, DeepSeek, Gemma2, ...; MLLM: Qwen2-VL, Qwen2-Audio, Llama3.2-Vision, Llava, InternVL2, MiniCPM-V-2.6, GLM4v, Xcomposer2.5, Yi-VL, DeepSeek-VL, Phi3.5-Vision, ...)
https://swift.readthedocs.io/zh-cn/latest/Instruction/index.html
Apache License 2.0
4.27k stars 377 forks source link

internvl-chat-v1_5 DPO 报错 #1773

Closed teamwong111 closed 2 months ago

teamwong111 commented 3 months ago

Describe the bug What the bug is, and how to reproduce, better with screenshots(描述bug以及复现过程，最好有截图)

internvl-chat-v1_5 DPO 会报错，可以使用官方示例测试

image

问题在 swift/llm/utils/model.py 4452 行

model, tokenizer = get_model_tokenizer_from_repo(
        model_dir, torch_dtype, model_kwargs, load_model, tokenizer=tokenizer, model_config=model_config, **kwargs)

internvl-chat-v1_5 config.jsonl 里没有配置 AutoModelForCasualLM，所以会报错

hjh0119 commented 2 months ago

升级transformers有这个问题吗

Jintao-Huang commented 2 months ago

解决了嘛，可复现的sh发一下