Atari results are not reproducible

Div99 / IQ-Learn

(NeurIPS '21 Spotlight) IQ-Learn: Inverse Q-Learning for Imitation

https://div99.github.io/IQ-Learn/

Other

199 stars 31 forks source link

Atari results are not reproducible #19

Open mertalbaba opened 5 months ago

mertalbaba commented 5 months ago

I can't reproduce the Atari results reported in your paper. I directly use your code with your hyper parameters, with same package versions. However, I can't reproduce your results. The problem also holds for MuJoCo as I explained in another issue. Can you please provide correct hyperparameters to reproduce your results?

Div99 commented 3 months ago

Hi, sorry I don't have the bandwidth to maintain this code anymore. Please look at using exact versions of pytorch and other libraries as in the repo, and try evaluating with the provided checkpoints