hongshaorou/Tianshou
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
Tianshou/examples/box2d
History
ChenDRAG 3108b9db0d
Add Timelimit trick to optimize policies (#296) 
* consider timelimit.truncated in calculating returns by default
* remove ignore_done
2021-02-26 13:23:18 +08:00
..
results/sac
sac mujoco result (#246)
2020-11-09 16:43:55 +08:00
acrobot_dualdqn.py
add logger (#295)
2021-02-24 14:48:42 +08:00
bipedal_hardcore_sac.py
add logger (#295)
2021-02-24 14:48:42 +08:00
lunarlander_dqn.py
add logger (#295)
2021-02-24 14:48:42 +08:00
mcc_sac.py
Add Timelimit trick to optimize policies (#296)
2021-02-26 13:23:18 +08:00
README.md
sac mujoco result (#246)
2020-11-09 16:43:55 +08:00

README.md

Bipedal-Hardcore-SAC

  • Our default choice: remove the done flag penalty, will soon converge to ~280 reward within 100 epochs (10M env steps, 3~4 hours, see the image below)
  • If the done penalty is not removed, it converges much slower than before, about 200 epochs (20M env steps) to reach the same performance (~200 reward)