Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/86941
Note: Links to docs will display an error until the docs builds have been completed. ❌ 3 Failures, 1 PendingAs of commit ccdbb27: The following jobs have failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Dockerfile
Outdated
There was a problem hiding this comment.
Does pytorch conda channel contain pytorch-cuda packages?
There was a problem hiding this comment.
These 2 channels should contain the right packages:
ARG CUDA_CHANNEL=nvidia
ARG INSTALL_CHANNEL=pytorch-nightly
Nightly contains pytorch-cuda
Pytorch not yet
b93eb95 to
2cbb202
Compare
Docker builds fix Fix docker file Testing Advance Docker builds to cuda 11.6 Fix typo Fix typo Fix cuda version test Tune substitution test test fix test test test test testing
33bcc85 to
7302d2f
Compare
docker.Makefile
Outdated
| CUDA_CHANNEL = nvidia | ||
| # The conda channel to use to install pytorch / torchvision | ||
| INSTALL_CHANNEL ?= pytorch | ||
| INSTALL_CHANNEL ?= pytorch-nightly |
There was a problem hiding this comment.
oh wait we shouldn't change the defaults for this
There was a problem hiding this comment.
There might be users who rely on the default behavior so changing this could be very problematic
There was a problem hiding this comment.
The issue with pytorch channel is that it does not contain pytorch-cuda yet. I think publishing pytorch-cuda package to pytorch channel should fix this issue.
docker.Makefile
Outdated
| CUDA_CHANNEL = nvidia | ||
| # The conda channel to use to install pytorch / torchvision | ||
| INSTALL_CHANNEL ?= pytorch | ||
| INSTALL_CHANNEL ?= pytorch-nightly |
There was a problem hiding this comment.
There might be users who rely on the default behavior so changing this could be very problematic
|
Please note current failure of Build Official Docker Image is because of this issue: |
|
@pytorchmergebot merge -f "some build failures expected" |
Merge startedYour change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
|
@pytorchmergebot revert -c "Workflow is passing but installs CUDA 11.3 PyTorch rather then 11.6" |
|
❌ 🤖 pytorchbot command failed: Try |
|
@pytorchmergebot revert -m "Workflow is passing but installs CUDA 11.3 PyTorch rather then 11.6" -c weird |
|
@pytorchbot successfully started a revert job. Check the current status here. |
|
@atalman your PR has been successfully reverted. |
This reverts commit c5de535. Reverted #86941 on behalf of https://github.com/atalman due to Workflow is passing but installs CUDA 11.3 PyTorch rather then 11.6
Fixes following: https://github.com/pytorch/pytorch/actions/runs/3242695506/jobs/5316334351 crash in Docker builds introduced by: #82682 The PR seems to introduce some changes not compatible with cuda 11.3 which is used by our Docker builds This is a reland of original pr: #86941 (Created this new PR to start fresh) Which was reverted because conda install, installed wrong version of pytorch. It installed pytorch for cuda 11.3 still rather then 11.6 This should be fixed now with Release 1.13 Pull Request resolved: #87858 Approved by: https://github.com/seemethere, https://github.com/malfet, https://github.com/izaitsevfb
Fixes following: https://github.com/pytorch/pytorch/actions/runs/3242695506/jobs/5316334351 crash in Docker builds introduced by: pytorch#82682 The PR seems to introduce some changes not compatible with cuda 11.3 which is used by our Docker builds This is a reland of original pr: pytorch#86941 (Created this new PR to start fresh) Which was reverted because conda install, installed wrong version of pytorch. It installed pytorch for cuda 11.3 still rather then 11.6 This should be fixed now with Release 1.13 Pull Request resolved: pytorch#87858 Approved by: https://github.com/seemethere, https://github.com/malfet, https://github.com/izaitsevfb
Fixes following: https://github.com/pytorch/pytorch/actions/runs/3242695506/jobs/5316334351 crash in Docker builds introduced by: pytorch#82682 The PR seems to introduce some changes not compatible with cuda 11.3 which is used by our Docker builds This is a reland of original pr: pytorch#86941 (Created this new PR to start fresh) Which was reverted because conda install, installed wrong version of pytorch. It installed pytorch for cuda 11.3 still rather then 11.6 This should be fixed now with Release 1.13 Pull Request resolved: pytorch#87858 Approved by: https://github.com/seemethere, https://github.com/malfet, https://github.com/izaitsevfb
Fixes following:
https://github.com/pytorch/pytorch/actions/runs/3242695506/jobs/5316334351
crash in Docker builds introduced by: #82682
The PR seems to introduce some changes not compatible with cuda 11.3 which is used by our Docker builds