search

yandexdataschool / Practical_RL

A course in reinforcement learning in the wild
The Unlicense
5.92k stars 1.7k forks source link

spring_2024_week_04_approx_rl: a major update #541

Closed zshrav closed 8 months ago

zshrav commented 8 months ago
  1. Removed the preprocessing section from the task. Instead, gymnasium preprocessing is used. uint8 dtype and LazyFrames allow to use a larger buffer, resulting in higher scores.
  2. Included Dueling and Double DQN to the main task.
  3. Included interpretation to the main task.
  4. Modified some tests and added some new ones.
  5. Modified the evaluation points.
  6. Slightly reorganized the code.
review-notebook-app[bot] commented 8 months ago

Check out this pull request on  ReviewNB

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB