feature(zzh): add IBC algorithm

[ZHZisZZ]

Description

Implicit Behavior Cloning Experiment results

Related Issue

TODO

Check List

• [x] minimal working pipeline
• [x] vanilla energy based model
• [x] autoregressive energy based model
• [x] MCMC optimizer
• [x] multigpu training

[codecov[bot]]

Codecov Report

Merging #401 (1a8a616) into main (17fc41d) will decrease coverage by 0.09%. The diff coverage is 88.92%.

@@            Coverage Diff             @@
##             main     #401      +/-   ##
==========================================
- Coverage   85.76%   85.67%   -0.10%     
==========================================
  Files         532      537       +5     
  Lines       42412    42831     +419     
==========================================
+ Hits        36375    36695     +320     
- Misses       6037     6136      +99     

Flag	Coverage Δ
unittests	85.67% <88.92%> (-0.10%)	:arrow_down:

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

Impacted Files	Coverage Δ
ding/utils/__init__.py	96.00% <ø> (ø)
ding/utils/data/dataset.py	60.08% <66.66%> (+1.79%)	:arrow_up:
...g/worker/collector/interaction_serial_evaluator.py	82.43% <77.10%> (-4.04%)	:arrow_down:
ding/entry/serial_entry_offline.py	89.65% <84.61%> (-2.02%)	:arrow_down:
ding/policy/ibc.py	87.75% <87.75%> (ø)
ding/model/template/ebm.py	94.39% <94.39%> (ø)
ding/model/template/tests/test_ebm.py	100.00% <100.00%> (ø)
ding/policy/__init__.py	100.00% <100.00%> (ø)
ding/policy/command_mode_policy_instance.py	93.04% <100.00%> (+0.12%)	:arrow_up:
ding/policy/mbpolicy/mbsac.py	96.00% <100.00%> (ø)
... and 72 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.