Bug#915186: RFA: caffe -- Fast, open framework for Deep Learning
Package: wnpp
Severity: normal
X-Debbugs-CC: debian-science@lists.debian.org
As time goes by, the DL framework is more and more likely an educational
code base. It's architecture and design is old, compared to PyTorch or
TensorFlow 2.0.
The last time I use caffe for neural network training is several months
ago on a toy dataset, and don't plan to use it anymore. Based on that,
I'll be no longer the right maintainer for caffe and caffe-contrib for
the post-Buster period. It would be better if someone is willing to
take it over in the future -- so I can put more time on some more
important packages, and new frameworks.
Currently this bug is an RFA for both src:caffe and src:caffe-contrib.
After the Buster release, the status of src:caffe will still be RFA, but
the status of src:caffe-contrib will become Orphan. I've totally lost
my interest in packaging CUDA compiler products.
To adopt Caffe, please also make sure to review upstream's installation
guide for Debian (I wrote that) and keep it up-to-date as long as
possible.
Reply to: