[BUG] <title>【多图推理】请问如何使用transformer进行minicpm-v 2.6的多图推理？ - Githubissues

OpenBMB / MiniCPM-V

MiniCPM-V 2.6: A GPT-4V Level MLLM for Single Image, Multi Image and Video on Your Phone

Apache License 2.0

12.75k stars 893 forks source link

[BUG] <title>【多图推理】请问如何使用transformer进行minicpm-v 2.6的多图推理？ #646

Open popoyaya opened 1 month ago

popoyaya commented 1 month ago

是否已有关于该错误的issue或讨论？ | Is there an existing issue / discussion for this?

[X] 我已经搜索过已有的issues和讨论 | I have searched the existing issues / discussions

该问题是否在FAQ中有解答？ | Is there an existing answer for this in FAQ?

[X] 我已经搜索过FAQ | I have searched FAQ

当前行为 | Current Behavior

No response

期望行为 | Expected Behavior

No response

复现方法 | Steps To Reproduce

No response

运行环境 | Environment

- OS:
- Python:
- Transformers:
- PyTorch:
- CUDA (`python -c 'import torch; print(torch.version.cuda)'`):

备注 | Anything else?

No response

shaohua-pan commented 2 weeks ago

可以参考：https://github.com/OpenBMB/MiniCPM-V?tab=readme-ov-file#chat-with-multiple-images