@slidestart
AlphaZero 大作业展示
实现思路
–
MCTS
def get_action_probs(board, player):
if training:
repeat search num_sim times
else:
return policy predicted by NNet
6/27/23About 1 min
@slidestart
–
def get_action_probs(board, player):
if training:
repeat search num_sim times
else:
return policy predicted by NNet
@slidestart
2023-03-28
–
12th Gen Intel(R) Core(TM) i7-1260P (16)| Task | Iterations | Time | Kernel Time |
|---|---|---|---|
s1 -> t1 |
131072 |
0:01:12.125142 |
40.411 s |
u0 -> u0-aligned |
4096 |
0:00:03.423797 |
1.429 s |
s1 -> u1-aligned |
131072 |
0:01:20.373831 |
44.751 s |
all |
2m44s |