[Refactor] Support passing arguments to loss from head.

[Ezra-Yu]

Motivation

the forward function of loss(CE, Focal Loss and etc) have parameters like weight and so on. but users can not pass weight by using function head.forward_train.

Modification

Add **kwargs in heads, and add kwargs in loss function in Clshead.

BC-breaking (Optional)

No.

Use cases (Optional)

weight = torch.tensor( [0.5, 0.5] )
head.forward_train(imgs, gt_labels, weight=weight........)

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 #523 (465f535) into master (dc35eb6) will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #523   +/-   ##
=======================================
  Coverage   79.48%   79.48%           
=======================================
  Files         106      106           
  Lines        5975     5975           
  Branches      968      968           
=======================================
  Hits         4749     4749           
  Misses       1095     1095           
  Partials      131      131           

Flag	Coverage Δ
unittests	79.48% <100.00%> (ø)

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

Impacted Files	Coverage Δ
mmcls/models/heads/cls_head.py	83.33% <100.00%> (ø)
mmcls/models/heads/linear_head.py	85.18% <100.00%> (ø)
mmcls/models/heads/multi_label_head.py	71.05% <100.00%> (ø)
mmcls/models/heads/multi_label_linear_head.py	67.85% <100.00%> (ø)
mmcls/models/heads/stacked_head.py	97.10% <100.00%> (ø)
mmcls/models/heads/vision_transformer_head.py	93.18% <100.00%> (ø)

---

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 dc35eb6...465f535. Read the comment docs.

[mzr1996]

LGTM