search

HumanCompatibleAI / evaluating-rewards

Library to compare and evaluate reward functions
https://arxiv.org/abs/2006.13900
Apache License 2.0
61 stars 7 forks source link

Integrate rollout return with combined_distances and add checkpoint comparison figure plotting #50

Closed AdamGleave closed 3 years ago

codecov[bot] commented 3 years ago

Codecov Report

Merging #50 (cdc1a51) into master (dca9d76) will decrease coverage by 0.61%. The diff coverage is 74.93%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master      #50      +/-   ##
==========================================
- Coverage   83.47%   82.86%   -0.62%     
==========================================
  Files          68       69       +1     
  Lines        4598     4796     +198     
==========================================
+ Hits         3838     3974     +136     
- Misses        760      822      +62
Impacted Files Coverage Δ
src/evaluating_rewards/scripts/distances/npec.py 79.52% <0.00%> (ø)
...ating_rewards/scripts/rewards/train_adversarial.py 0.00% <0.00%> (ø)
...ing_rewards/scripts/pipeline/combined_distances.py 66.45% <60.58%> (ø)
src/evaluating_rewards/scripts/rl_common.py 80.00% <80.00%> (ø)
src/evaluating_rewards/distances/common_config.py 90.38% <86.66%> (-5.92%) :arrow_down:
...aluating_rewards/scripts/pipeline/train_experts.py 84.00% <89.47%> (+1.39%) :arrow_up:
...uating_rewards/scripts/distances/rollout_return.py 96.82% <96.82%> (ø)
src/evaluating_rewards/rewards/base.py 95.85% <100.00%> (+0.01%) :arrow_up:
src/evaluating_rewards/rewards/comparisons.py 89.65% <100.00%> (+0.36%) :arrow_up:
src/evaluating_rewards/rewards/preferences.py 99.27% <100.00%> (ø)
... and 10 more

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 dca9d76...cdc1a51. Read the comment docs.