Fix two issues introduced in AutoGPTQ deprecation#3014
Fix two issues introduced in AutoGPTQ deprecation#3014githubnemo merged 5 commits intohuggingface:mainfrom
Conversation
Merging PR huggingface#2932 introduced two bugs that lead to failing CPU and GPU pipelines. Firstly, a merge on main in the PR without a follow-up CI run re-introduced a deleted AutoGPTQ code branch which is now removed again. Secondly, gptqmodel seems, just like EETQ, to need a non-isolated build environment to find external dependencies like PyTorch to be installed correctly. As this was not present, the nightly slow CI wasn't run.
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
Apparently `hf-doc-builder` doesn't expose its dependency to `requests` in the `setup.py` (it does in `pyproject.toml`). For some reason `requests` is not installed anymore (some other dependency removed it probably), so we're getting CI errors.
githubnemo
force-pushed
the
issue/fix-reintroduced-autogptq-code
branch
from
72c0c2c to
98f828e
Compare
The docker test build failed because of escaped values passed to the next task. Avoiding wrapping the value in an env variable fixes this issue.
|
@sayakpaul I will attempt to fix the BPE |
| ALL_CHANGED_FILES: ${{ steps.changed-files.outputs.all_changed_files }} | ||
| run: | | ||
| echo "matrix=${ALL_CHANGED_FILES}" >> $GITHUB_OUTPUT | ||
| echo "matrix=${{ steps.changed-files.outputs.all_changed_files }}" >> $GITHUB_OUTPUT |
There was a problem hiding this comment.
I think the previous implement is for security reasons?
There was a problem hiding this comment.
Yes, it was and it is indeed a potential vector since file names are controlled by a potential attacker - even though the output from the changed-files plugin is JSON encoded I wouldn't bet it is shell escaped as well. I've updated with a fix that uses shell escaping through env vars (like before) but fixes the escaping issue.
Escaping the changed files is indeed relevant since it is a potential intruder controlled vector.
sayakpaul
left a comment
sayakpaul
left a comment
There was a problem hiding this comment.
Thanks for fixing this on priority. Approving given the tests pass.
|
HTTP errors are unrelated but The same goes for the failing docker build. This is best handled separately. |
3 participants
Merging PR #2932 introduced two bugs that lead to failing CPU and GPU pipelines.
Firstly, a merge on main in the PR without a follow-up CI run re-introduced a deleted AutoGPTQ code branch which is now removed again.
Secondly, gptqmodel seems, just like EETQ, to need a non-isolated build environment to find external dependencies like PyTorch to be installed correctly. As this was not present, the nightly slow CI wasn't run.
Additional CI fixes:
DeprecationWarningfrom diffusers via transformersrequestsdependency forhf-doc-builderto fix quality CI check