Closed metalaureate closed 2 years ago
Thanks - note that in the meantime you can use pip with the included requirements.txt file which also generally works.
Thank you for the note! I couldn’t get the requirements.txt to install locally, which was why I was trying the Docker route (Mac). Any advice?
Not a cog expert, but @zeke might be able to help.
cc @bfirsh @andreasjansson
The error I get when I try to install locally on a fresh conda environment (Python 3.9) is
ERROR: Could not find a version that satisfies the requirement torch==1.9.0+cu102 (from versions: 1.7.1, 1.8.0, 1.8.1, 1.9.0, 1.9.1, 1.10.0)
ERROR: No matching distribution found for torch==1.9.0+cu102
This appears to be the same error that is killing the docker image build.
I got it working (hooray!) locally by removing the +cu102
and remembering to execute the # DO THIS comment for installing diffvg, and also randomly needing to install googletrans
. I think Cog is really awesome idea and I would love to use it for this if we could figure that out (I mean, it is meant to solve for this exact kind of problem case!).
Great - if you don't have GPU then you might want to start with quality=draft which will be less resource intensive.
Any pointers on getting it to build on Cog/Docker?