mmcv issue

[lrj16]

I tried everything i can, but still can not fix this problem. I just want to inference and do not need to train, but this line:

from mmdet.apis import inference_detector, init_detector

always report error: ModuleNotFoundError: No module named 'mmcv._ext'
My env:

torch.version
'2.2.2'
print(f"CUDA version used by PyTorch: {torch.version.cuda}")
CUDA version used by PyTorch: 12.1

mmcv: 2.1.0

How to fix this bug? any suggestion? thanks a lot.

[name-liang]

可以尝试pip install mmcv==2.1.0 -f https://download.openmmlab.com/mmcv/dist/cu121/torch2.4/index.html ，对torch2.4+cu121有效

[lrj16]

可以尝试pip install mmcv==2.1.0 -f https://download.openmmlab.com/mmcv/dist/cu121/torch2.4/index.html ，对torch2.4+cu121有效

多谢啦，我尝试过不行。已经替换成别的方法了

[buaa-csc]

可以尝试pip install mmcv==2.1.0 -f https://download.openmmlab.com/mmcv/dist/cu121/torch2.4/index.html，对torch2.4+cu121有效

多谢啦，我尝试过不行。已经替换成其他方法了

请问您使用了什么方法

[lrj16]

可以尝试pip install mmcv==2.1.0 -f https://download.openmmlab.com/mmcv/dist/cu121/torch2.4/index.html，对torch2.4+cu121有效

多谢啦，我尝试过不行。已经替换成其他方法了

请问您使用了什么方法

我用了poseViT了，Wan2.2开源的代码中就有模型和相关推理代码。