Yangqing Jia
Yangqing Jia
Thanks @soumith for organizing this effort! I think this would definitely help us advance the field to the next level. I am also very interested in benchmarking not only the...
Should be fixed in #1690 now - feel free to reopen if it does not work.
cc/ @salexspb - maybe this is something we can abstract out in the predictor interface?
Ongoing PR #1827 aims to address the issue for (a) no-cuda, shared or static, and (b) cuda, static cases. CUDA + shared is currently not working very well, and the...
Assigning a few folks to take a look - possibly due to additional dependencies we introduced (such as ACL)?
(Also, it is possible it's due to the removal of "-s", cc @Maratyszcza
I am 95% towards dealing with cuda + MSVC. Right now MSVC 2017 and CUDA9 do not work for any framework unfortunately. See https://devtalk.nvidia.com/default/topic/1027299/cuda-setup-and-installation/cuda-9-failed-to-support-the-latest-visual-studio-2017-version-15-5/ . For CUDA9 and VS2015 I...
For now, if you sync beyond https://github.com/caffe2/caffe2/commit/4372b8d04b203c85f68c144df82dae560887ef13 , at least you will be able to overcome the MSB6006 issue.
/cc @akyrola Aapo, what do you think is the best way for us to do finetuning, especially with an existing model? My gut feeling is that, one can manually feed...
Update: PR #1793 is out to make windows build a bit better.