internvl2-2b lora inference报错 - Githubissues

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

4.34k stars 381 forks source link

internvl2-2b lora inference报错 #2470

Closed cornfield404 closed 4 days ago

cornfield404 commented 4 days ago

参考之前的issue安装了transformers 4.44.0，但又产生了新的错误[ERROR:swift] import lmdeploy_utils error: No module named 'transformers.models.mllama'

happyday-lkj commented 2 days ago

你是怎么解决的呢

cornfield404 commented 2 days ago

你是怎么解决的呢

修改了lmdeploy源码，删掉了对mllama的依赖

happyday-lkj commented 2 days ago

你是怎么解决的呢

修改了lmdeploy源码，删掉了对mllama的依赖

好吧，我创建了两个环境，一个lmdeploy,一个vllm