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
tl;dr: BoTorch will stop publishing a botorch conda package to the -c pytorch conda channel. However, there will still be a conda package available on the -c conda-forge channel for those users who need it
Supporting publishing our own conda package has caused an outsized maintenance burden over the years, especially more recently. conda builds are slow, the setup is brittle and often breaks. There were a share of bugs with the conda dependency solver in the past. Recently, our conda tests broke and we haven't been able to figure out why. At this point we can't really justify this added maintenance burden given the limited benefit of publishing the conda package, especially since the -c conda-forge channel also publishes a botorch package. Finally, PyTorch is deprecating its official anaconda channel for similar reasons.
BoTorch version 0.12 will have been the last version published to the -c pytorch channel. This will reduce maintenance burden and speed up GitHub Actions CI by requiring fewer testing workflows (and thus also save compute capacity).
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
tl;dr: BoTorch will stop publishing a botorch conda package to the
-c pytorch
conda channel. However, there will still be a conda package available on the-c conda-forge
channel for those users who need itSupporting publishing our own conda package has caused an outsized maintenance burden over the years, especially more recently. conda builds are slow, the setup is brittle and often breaks. There were a share of bugs with the conda dependency solver in the past. Recently, our conda tests broke and we haven't been able to figure out why. At this point we can't really justify this added maintenance burden given the limited benefit of publishing the conda package, especially since the
-c conda-forge
channel also publishes a botorch package. Finally, PyTorch is deprecating its official anaconda channel for similar reasons.BoTorch version 0.12 will have been the last version published to the
-c pytorch
channel. This will reduce maintenance burden and speed up GitHub Actions CI by requiring fewer testing workflows (and thus also save compute capacity).Beta Was this translation helpful? Give feedback.
All reactions