[Feature] Support CPU training

[HIT-cwh]

Hi! Related to open-mmlab/mmdet#7016, CPU training in MMRazor is supported in this pr.

We modified the train api in mmrazor/apis/mmdet(mmseg)/train.py while kept mmrazor/apis/mmcls/train.py the same.

Now we can use the CPU to train/debug our model and test our model with batch size >=2. Before running the program we need to export CUDA_VISIBLE_DEVICES=-1 to disable GPU visibility.

[codecov]

Codecov Report

Merging #62 (b26ee0c) into master (9596379) will increase coverage by 0.00%.
The diff coverage is 0.00%.

❗ Current head b26ee0c differs from pull request most recent head a71f2d4. Consider uploading reports for the commit a71f2d4 to get more accurate results

@@           Coverage Diff           @@
##           master      #62   +/-   ##
=======================================
  Coverage   55.60%   55.61%           
=======================================
  Files          81       81           
  Lines        2942     2938    -4     
  Branches      544      542    -2     
=======================================
- Hits         1636     1634    -2     
+ Misses       1231     1228    -3     
- Partials       75       76    +1     

Flag	Coverage Δ
unittests	55.61% <0.00%> (+<0.01%)	⬆️

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

Impacted Files	Coverage Δ
mmrazor/apis/mmdet/train.py	0.00% <0.00%> (ø)
mmrazor/apis/mmseg/train.py	0.00% <0.00%> (ø)
mmrazor/models/mutators/one_shot_mutator.py	92.59% <0.00%> (-3.71%)	⬇️

---

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 9596379...a71f2d4. Read the comment docs.