You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Situation: One of the dependencies of xplique is tensorflow which offers a cpu-only package with a smaller footprint: tensorflow-cpu.
The problem is that when using both tensorflow-cpu and xplique, tensorflow is still pulled which results in a very big env which is not needed.
Idea: Push a second version of xplique called xplique-cpu that would have tensorflow-cpu as dependency instead of tensorflow. This would enable us to use tensorflow-cpu and xplique with a smaller disk footprint.
The text was updated successfully, but these errors were encountered:
When installing xplique, you can always do so using the --no-deps flag, which installs only the package (no dependencies are installed alongside). If you create your env, you install tensorflow-cpu and then run a pip install xplique --no-deps, you should have a working version of xplique that uses tensorflow-cpu instead of the regular version.
Should you still encounter an issue, don't hesitate to contact us.
Situation: One of the dependencies of xplique is tensorflow which offers a cpu-only package with a smaller footprint: tensorflow-cpu.
The problem is that when using both tensorflow-cpu and xplique, tensorflow is still pulled which results in a very big env which is not needed.
Idea: Push a second version of xplique called xplique-cpu that would have tensorflow-cpu as dependency instead of tensorflow. This would enable us to use tensorflow-cpu and xplique with a smaller disk footprint.
The text was updated successfully, but these errors were encountered: