1
0
mirror of https://github.com/gryf/coach.git synced 2025-12-17 19:20:19 +01:00
Files
coach/benchmarks/acer
2019-03-17 15:33:28 +02:00
..
2019-02-20 23:52:34 +02:00
2019-02-20 23:52:34 +02:00

ACER

Each experiment uses 3 seeds. The parameters used for ACER are the same parameters as described in the original paper, except for the optimizer (changed to ADAM) and learning rate (1e-4) used.

Breakout ACER - 16 workers

coach -p Atari_ACER -lvl breakout -n 16
Breakout ACER

Space Invaders ACER - 16 workers

coach -p Atari_ACER -lvl space_invaders -n 16
Space Invaders ACER

Pong ACER - 16 workers

coach -p Atari_ACER -lvl pong -n 16
Pong ACER