polish(nyz): polish offpolicy RL multi-gpu DDP training

[PaParaZz1]

Description

• fix random policy bug
• add multi-gpu DDP config transformation
• polish multi-gpu DDP log and statistics (e.g. envstep)

Related Issue

TODO

Check List

• merge the latest version source branch/repo, and resolve all the conflicts
• pass style check
• pass all the tests

[codecov]

Codecov Report

Merging #679 (05ef4dd) into main (6b188c9) will decrease coverage by 0.25%.
The diff coverage is 90.04%.

@@            Coverage Diff             @@
##             main     #679      +/-   ##
==========================================
- Coverage   82.06%   81.81%   -0.25%     
==========================================
  Files         586      588       +2     
  Lines       47515    48148     +633     
==========================================
+ Hits        38991    39392     +401     
- Misses       8524     8756     +232     

Flag	Coverage Δ
unittests	81.81% <90.04%> (-0.25%)	⬇️

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

Impacted Files	Coverage Δ
...g/worker/collector/interaction_serial_evaluator.py	83.56% <33.33%> (+1.98%)	⬆️
ding/worker/collector/sample_serial_collector.py	88.52% <78.57%> (-3.00%)	⬇️
ding/policy/policy_factory.py	69.23% <80.00%> (+1.66%)	⬆️
ding/torch_utils/network/merge.py	84.44% <84.44%> (ø)
ding/worker/replay_buffer/advanced_buffer.py	92.10% <96.15%> (-0.62%)	⬇️
ding/config/config.py	84.48% <100.00%> (ø)
ding/entry/serial_entry.py	96.92% <100.00%> (ø)
...ing/framework/middleware/tests/test_distributer.py	19.37% <100.00%> (-38.32%)	⬇️
ding/torch_utils/data_helper.py	78.51% <100.00%> (ø)
ding/torch_utils/network/__init__.py	100.00% <100.00%> (ø)
... and 4 more

... and 261 files with indirect coverage changes