Tianshou

hongshaorou/Tianshou

Fork 0

Commit Graph

Author	SHA1	Message	Date
Trinkle23897	34f714a677	Numba acceleration (#193 ) Training FPS improvement (base commit is 94bfb32): test_pdqn: 1660 (without numba) -> 1930 discrete/test_ppo: 5100 -> 5170 since nstep has little impact on overall performance, the unit test result is: GAE: 4.1s -> 0.057s nstep: 0.3s -> 0.15s (little improvement) Others: - fix a bug in ttt set_eps - keep only sumtree in segment tree implementation - dirty fix for asyncVenv check_id test	2020-09-02 13:03:32 +08:00
n+e	140b1c2cab	Improve PER (#159 ) - use segment tree to rewrite the previous PrioReplayBuffer code, add the test - enable all Q-learning algorithms to use PER	2020-08-06 10:26:24 +08:00

Author

SHA1

Message

Date

Trinkle23897

34f714a677

Numba acceleration (#193 )

Training FPS improvement (base commit is 94bfb32):
test_pdqn: 1660 (without numba) -> 1930
discrete/test_ppo: 5100 -> 5170

since nstep has little impact on overall performance, the unit test result is:
GAE: 4.1s -> 0.057s
nstep: 0.3s -> 0.15s (little improvement)

Others:
- fix a bug in ttt set_eps
- keep only sumtree in segment tree implementation
- dirty fix for asyncVenv check_id test

2020-09-02 13:03:32 +08:00

n+e

140b1c2cab

Improve PER (#159 )

- use segment tree to rewrite the previous PrioReplayBuffer code, add the test

- enable all Q-learning algorithms to use PER

2020-08-06 10:26:24 +08:00

2 Commits