pytorch / pytorch/vision

Installation fails using PDM

Open
#5,371 3 comments 2 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

topic: binaries
Dominant language
Python
Stars
17.9k
Forks
7.3k
Avg merge
1d 15h
Merged PRs (30d)
13

Description

🐛 Describe the bug

Within a pdm project, running pdm add torchvision results in:

ERRORS:
add torchvision failed:
Traceback (most recent call last):
  File "c:\users\classner\.pyenv\pyenv-win\versions\3.9.0\lib\concurrent\futures\thread.py", line 52, in run
    result = self.fn(*self.args, **self.kwargs)
  File "C:\Users\classner\AppData\Roaming\Python\Python39\site-packages\pdm\installers\synchronizers.py", line 190, in install_candidate 
    self.manager.install(can)
  File "C:\Users\classner\AppData\Roaming\Python\Python39\site-packages\pdm\installers\manager.py", line 38, in install
    installer(candidate.build(), self.environment, candidate.direct_url())
  File "C:\Users\classner\AppData\Roaming\Python\Python39\site-packages\pdm\installers\installers.py", line 176, in install_wheel        
    _install_wheel(
  File "C:\Users\classner\AppData\Roaming\Python\Python39\site-packages\pdm\installers\installers.py", line 280, in _install_wheel       
    for record_elements, stream in source.get_contents():
  File "C:\Users\classner\AppData\Roaming\Python\Python39\site-packages\installer\sources.py", line 170, in get_contents
    assert record is not None, "In {}, {} is not mentioned in RECORD".format(
AssertionError: In C:\Users\classner\AppData\Local\Temp\pip-unpack-49uvth_n\torchvision-0.11.2+cu113-cp39-cp39-win_amd64.whl, torchvision/extension.py is not mentioned in RECORD

It seems the pip RECORD file is not correctly built for torchvision. Similar issues are being detected for other packages, too, for example: https://github.com/microsoft/playwright-python/issues/518 .

Versions

Currently tested with 0.11.2+cu113-cp39-cp39 .

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the reported pdm add torchvision failure with torchvision 0.11.2+cu113-cp39-cp39 and inspect the generated wheel's RECORD entry for torchvision/extension.py. Compare the wheel contents with the installer assertion; done means installation completes successfully and the package metadata is consistent.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
build-system, tooling
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.