Using the BVLC/caffe with Python 3, OpenCV 4, CUDA 11.8.0 and cuDNN 8.
255
The source repository duruyao/caffe is forked from BVLC/caffe and updated to use Python 3, OpenCV 4, CUDA 11.X and cuDNN 8.
The following table lists some major differences between my own Docker image duruyao/caffe and the official Docker image bvlc/caffe.
| bvlc/caffe | duruyao/caffe |
|---|---|
| ubuntu 16.04 | ubuntu 22.04 |
| cuda 8.0 | cuda 11.8.0 |
| cudnn 6 | cudnn 8 |
| python 2 | python 3 |
| opencv 3 | opencv 4 |
When using only Caffe, the official Docker image bvlc/caffe is still available.
However, if you want to use Caffe and other popular AI frameworks (such as the latest ONNX and TensorFlow) in the same image, please build your own image based on duruyao/caffe, or refer to the Dockerfile.
docker pull duruyao/caffe:cpu
docker pull duruyao/caffe:gpu
For more information about Caffe, please visit the official website caffe.berkeleyvision.org and the original repository BVLC/caffe.
Content type
Image
Digest
sha256:dcf9b2c29…
Size
707.2 MB
Last updated
about 3 years ago
docker pull duruyao/caffe:cpu