fix: multi task predict for nested heads

[piercus]

Motivation

I faced a problem when playing with multi task following #1229

• Each head is not called with a data_sample as input (see

  mmpretrain/mmpretrain/models/heads/multi_task_head.py

  Line 119 in e80418a

  task_samples = head.predict(feats)

  ), which will break the process with nested multi_task heads
• when the input data_sample is [MultiTaskDataSample, None] it is not working properly

Modification

• Reimplement data_sample filtering in multi_task_head predict method
• Add a unit test for predict in a nested scenario

Use cases

• Fix the multi-task nested
• I'm working on other use cases like shared layers between multi-tasked heads

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.

[piercus]

@marouaneamz would be glad to get your feedback on this :-)