DI-engine v0.2.2
Env (dizoo)
- apple key to door treasure env (#128)
- bsuite memory benchmark (#138)
- polish atari impala config
Algorithm
- Guided Cost IRL algorithm (#57)
- ICM exploration algorithm (#41)
- MP-DQN hybrid action space algorithm (#131)
- add loss statistics and polish r2d3 pong config (#126)
Enhancement
- add renew env mechanism in env manager and update timeout mechanism (#127) (#134)
Fix
- async subprocess env manager reset bug (#137)
- keepdims name bug in model wrapper
- on-policy ppo value norm bug
- GAE and RND unittest bug
- hidden state wrapper h tensor compatibility
- naive buffer auto config create bug
Style
- add supporters list
New Repo Feature
Contributors: @PaParaZz1 @puyuan1996 @RobinC94 @LikeJulia @Will-Nie @Weiyuhong-1998 @timothijoe @davide97l @lichuminglcm @YinminZhang