Some problems occurred during the installation of mmdetection according to the official installation tutorial. My environmental information is as follows:
The final error message is as follows:
From the mmcv compiler, you can see the mmcv CUDA compiler: not available.
The reason is that the mmcv full script I use is:
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu102/torch1.9.1/index.html
Use the following script to install:
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu102/torch1.9.0/index.html
The reason can be found on the official website:
The reason is that mmcv full has only been compiled in pytorch 1.X.0.
Similar Posts:
- [Solved] DCNv2 Compile Error: error: identifier “THCudaBlas_SgemmBatched” is undefined
- [Solved] PIP Install torch1.7 Error: ERROR: torch has an invalid wheel, .dist-info directory not found
- Solutions to the phenomena of / dev / sda2 clean and / dev / sda2 recovering journal when starting Ubuntu
- Go error reported unimplemented: 64 bit mode not compiled in and MinGW 64 bit installation error res resolved
- Failed building wheel for dlib [How to Solve]
- The solution of losing vcruntime140.dll in the latest Apache installation
- How to Solve SQL fulltext error
- No compatible servers were found when installing MySQL
- Solution to the error of video codec open failed prompted by TechSmith Camtasia Studio 9 screen recording software
- [Solved] pytorchImportError: numpy.core.multiarray failed to import