Use CPU version of PyTorch in Docker image

[c-bata]

Motivation

Currently the installed PyTorch wheel package in Docker image is different with tests-integration action.

• Dockerfile:

  optuna/Dockerfile

  Line 19 in fba23b1

  pip install ${PIP_OPTIONS} -e '.[benchmark, checking, document, integration, optional, test]' -f https://download.pytorch.org/whl/torch_stable.html; \

• tests-integration:

  optuna/.github/workflows/tests-integration.yml

  Line 60 in fba23b1

  pip install --progress-bar off .[integration] --extra-index-url https://download.pytorch.org/whl/cpu

Description of the changes

Change the PyTorch wheel package installed in Docker image.

[c-bata]

@not522 Could you review this PR? I think this can be merged with just your approval.

[codecov]

Codecov Report

Merging #5019 (66a9a0a) into master (ab87e40) will increase coverage by 0.01%.
Report is 1 commits behind head on master.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #5019      +/-   ##
==========================================
+ Coverage   89.36%   89.37%   +0.01%     
==========================================
  Files         203      203              
  Lines       15046    15046              
==========================================
+ Hits        13446    13448       +2     
+ Misses       1600     1598       -2     

see 1 file with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[not522]

LGTM!
Build Docker Image / dockerimage (3.7, dev) will fail, but #5009 addresses this issue.