[Enhance] New-style CPU training and inference.

[mzr1996]

Motivation

Refers to open-mmlab/mmdetection#7016, and it's based on open-mmlab/mmcv#1621

Modification

As the title.

BC-breaking (Optional)

No, the old style is reserved but deprecated and it will be removed after three months.

Use cases (Optional)

Old style CPU training

python tools/train.py $CONFIG --device cpu

New style CPU training

export CUDA_VISIBLE_DEVICES=-1
python tools/train.py $CONFIG

Checklist

Before PR:

• Pre-commit or other linting tools are used to fix the potential lint issues.
• Bug fixes are fully covered by unit tests, the case that causes the bug should be added in the unit tests.
• The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
• The documentation has been modified accordingly, like docstring or example tutorials.

After PR:

• If the modification has potential influence on downstream or other related projects, this PR should be tested with those projects, like MMDet or MMSeg.
• CLA has been signed and all committers have signed the CLA in this PR.

[codecov]

Codecov Report

Merging #674 (78bdf7d) into dev (a451d63) will increase coverage by 0.88%.
The diff coverage is 0.00%.

@@            Coverage Diff             @@
##              dev     #674      +/-   ##
==========================================
+ Coverage   82.27%   83.15%   +0.88%     
==========================================
  Files         120      126       +6     
  Lines        7045     7630     +585     
  Branches     1228     1332     +104     
==========================================
+ Hits         5796     6345     +549     
- Misses       1075     1095      +20     
- Partials      174      190      +16     

Flag	Coverage Δ
unittests	83.15% <0.00%> (+0.88%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
mmcls/apis/train.py	20.73% <0.00%> (-0.52%)	⬇️
mmcls/utils/__init__.py	100.00% <0.00%> (ø)
mmcls/models/necks/__init__.py	100.00% <0.00%> (ø)
mmcls/models/utils/__init__.py	100.00% <0.00%> (ø)
mmcls/models/backbones/resnet.py	100.00% <0.00%> (ø)
mmcls/models/backbones/__init__.py	100.00% <0.00%> (ø)
mmcls/models/backbones/hrnet.py	87.16% <0.00%> (ø)
mmcls/models/backbones/convnext.py	94.78% <0.00%> (ø)
mmcls/utils/setup_env.py	95.45% <0.00%> (ø)
mmcls/models/necks/hr_fuse.py	100.00% <0.00%> (ø)
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a451d63...78bdf7d. Read the comment docs.

[Ezra-Yu]

Why not keep this interface?
I prefer the use xxx --device cpu rather than export CUDA_VISIBLE_DEVICES=-1 xxxx.

the command is simpler.

[mzr1996]

To use MMDataParallel for CPU training/test, we need to specify CUDA_VISILBE_DEVICES=-1, and in that situation, the --device argument is useless.