You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Regarding this issue, we have summarized the following points
1、[3, 3, 16] is [evaluator_num, players_per_team, 4*4(dirction * action)]
2、In policy, we merge the first two dimensions, so it becomes [9,16].
3、When using DP,it will divide 9 into 4 and 5, and it will becomes [4,16] and [5,16].
4、[4,16] and [5,16] can't transform [batch, 3, 16], So it reported an error.
5、The solution is to only use DP on Encode,not on all gobigger model.
When I test the DP mode ,the error happened
Refer to the https://di-engine-docs.readthedocs.io/zh_CN/latest/best_practice/multi_gpu_example.html, I changed the file:
(1)gobigger_no_spatial_config.py
(2)gobigger_vsbot_baseline_simple_main.py
The text was updated successfully, but these errors were encountered: