Installing specific version of mmcv with find-links fails with new pip-20.3

I’m not sure what’s happening.

With pip 20.1.1 I was able to run:

pip install mmcv_full==1.0.5+torch1.6.0+cpu -f https://openmmlab.oss-accelerate.aliyuncs.com/mmcv/dist/index.html

and it works fine.

After upgrading pip to 20.3 I run the same command and get:

Looking in links: https://openmmlab.oss-accelerate.aliyuncs.com/mmcv/dist/index.html
ERROR: Exception:
Traceback (most recent call last):
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_vendor/resolvelib/resolvers.py", line 171, in _merge_into_criterion
    crit = self.state.criteria[name]
KeyError: 'mmcv-full'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_internal/cli/base_command.py", line 210, in _main
    status = self.run(options, args)
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_internal/cli/req_command.py", line 180, in wrapper
    return func(self, options, args)
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_internal/commands/install.py", line 319, in run
    reqs, check_supported_wheels=not options.target_dir
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_internal/resolution/resolvelib/resolver.py", line 122, in resolve
    requirements, max_rounds=try_to_avoid_resolution_too_deep,
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_vendor/resolvelib/resolvers.py", line 445, in resolve
    state = resolution.resolve(requirements, max_rounds=max_rounds)
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_vendor/resolvelib/resolvers.py", line 310, in resolve
    name, crit = self._merge_into_criterion(r, parent=None)
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_vendor/resolvelib/resolvers.py", line 173, in _merge_into_criterion
    crit = Criterion.from_requirement(self._p, requirement, parent)
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_vendor/resolvelib/resolvers.py", line 82, in from_requirement
    if not cands:
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_vendor/resolvelib/structs.py", line 124, in __bool__
    return bool(self._sequence)
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_internal/resolution/resolvelib/found_candidates.py", line 96, in __bool__
    return any(self)
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_internal/resolution/resolvelib/found_candidates.py", line 20, in _deduplicated_by_version
    for candidate in candidates:
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_internal/resolution/resolvelib/factory.py", line 203, in iter_index_candidates
    version=ican.version,
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_internal/resolution/resolvelib/factory.py", line 145, in _make_candidate_from_link
    link, template, factory=self, name=name, version=version,
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_internal/resolution/resolvelib/candidates.py", line 290, in __init__
    wheel_version = Version(wheel.version)
  File "/home/joncrall/.local/conda/envs/py36/lib/python3.6/site-packages/pip/_vendor/packaging/version.py", line 298, in __init__
    raise InvalidVersion("Invalid version: '{0}'".format(version))
pip._vendor.packaging.version.InvalidVersion: Invalid version: '1.0.5+torch1.6.0+cpu'

OS is Ubuntu 20.04.

[Erotemic]

Running a version with +cpu also fails:

pip install mmcv_full==1.0.5+torch1.6.0 -f https://openmmlab.oss-accelerate.aliyuncs.com/mmcv/dist/index.html

However, removing both torch and cpu specification works:

pip install mmcv_full -f https://openmmlab.oss-accelerate.aliyuncs.com/mmcv/dist/index.html

That works, but it takes a very long time and maxes out several of my cpu cores. (Its probably solving some dependency resolution problem). I’m thinking the new 2020 resolver broke something.

Yup: turning on the old resolver fixed it:

pip install mmcv_full==1.0.5+cpu -f https://openmmlab.oss-accelerate.aliyuncs.com/mmcv/dist/index.html --use-deprecated=legacy-resolver

Not sure if the error with the new resolver is in pip or if mmcv versions need to be changed to accommodate it.

[hellock]

Hi, the same problem occurred to me. I’m using pytorch 1.7.0 with cudatoolkit 10.2 and the command

‘pip install mmcv-full==latest+torch1.7.0+cu102 -f https://download.openmmlab.com/mmcv/dist/index.html’

gives the same error as specified using pip 20.3. It works properly with pip 20.2.4

As a workaround suggested by @Erotemic, you can use pip install mmcv-full==latest+torch1.7.0+cu102 -f https://openmmlab.oss-accelerate.aliyuncs.com/mmcv/dist/index.html --use-deprecated=legacy-resolver.