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
I'm not familier with conda-forge, are we restricted to use conda-forge packages? Otherwise can we use pytorch provided from its own channel so we can also get pytorch-nightly https://anaconda.org/pytorch-nightly/pytorch
The text was updated successfully, but these errors were encountered:
Hi @fecet, so sorry for the late reply. I'm assuming that you've managed to install a newer version of Pytorch from conda-forge already (e.g. 2.3.1), so not sure if this issue is still relevant?
In general, we discourage mixing channels because there could be incompatibilities in the builds across different channels causing weird undefined symbol errors (see https://conda-forge.org/docs/user/tipsandtricks/#using-multiple-channels for more information). For the Pytorch library in particular, some of the differences between the conda-forge and pytorch channel are mentioned at conda-forge/pytorch-cpu-feedstock#218 (see also my comment at conda-forge/pytorch3d-feedstock#10 (comment)). I'll close this issue for now, but feel free to reopen if you need additional clarification, or comment on the thread in pytorch-cpu-feedstock.
Comment:
I'm not familier with conda-forge, are we restricted to use conda-forge packages? Otherwise can we use pytorch provided from its own channel so we can also get pytorch-nightly https://anaconda.org/pytorch-nightly/pytorch
The text was updated successfully, but these errors were encountered: