google-deepmind / google-deepmind/rlax
Epsilon keeps getting reintialized in the example script
Open
- Dominant language
- Python
- Stars
- 1.4k
- Forks
- 109
- Avg merge
- 2h 15m
- Merged PRs (30d)
- 1
Description
In [simple_dqn.py](https://github.com/deepmind/rlax/blob/master/examples/simple_dqn.py), the epsilon schedule is chosen by actor_state.count however, that keeps getting reinitialized after every episode in [experiment.py](https://github.com/deepmind/rlax/blob/master/examples/experiment.py). So the epsilon schedule does not work.
Contributor guide
Assessment
This issue has not been assessed yet.