search

HumanCompatibleAI / adversarial-policies

Find best-response to a fixed policy in multi-agent RL
MIT License
275 stars 47 forks source link

Add paper hyperparameters as config to train #49

Closed AdamGleave closed 4 years ago

AdamGleave commented 4 years ago

Currently only way to replicate paper is via aprl.multi.train, but this is heavyweight for many purposes. Add the paper hyperparameters to a named config in aprl.train for ease of replication.

codecov[bot] commented 4 years ago

Codecov Report

Merging #49 into master will increase coverage by 0.05%. The diff coverage is 50%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master      #49      +/-   ##
==========================================
+ Coverage   62.87%   62.92%   +0.05%     
==========================================
  Files          58       58              
  Lines        5433     5435       +2     
==========================================
+ Hits         3416     3420       +4     
+ Misses       2017     2015       -2
Impacted Files Coverage Δ
src/aprl/configs/multi/train.py 17.31% <33.33%> (+0.42%) :arrow_up:
src/aprl/train.py 87.17% <60%> (-0.46%) :arrow_down:

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 8d4a286...d6d5a67. Read the comment docs.