-
Notifications
You must be signed in to change notification settings - Fork 4.3k
Closed
Labels
bugIssue describes a potential bug in ml-agents.Issue describes a potential bug in ml-agents.
Description
I am trying to create the right environment to train ML agents on a simple two-player Noughts & Crosses/tic-tac-toe game, and am transitioning from the default auto-step environment to the manually-called Academy.Instance.EnvironmentStep(), as being more suitable for turn-based games. When I include the EnvironmentStep line and press play, unity simply goes into a hang state and needs a force quit (even if I wrap that code into a coroutine).
Reproducing:
- behavior type set to Heuristic for both players (pre training), with random moves programmed
- Academy.Instance.EnvironmentStep() is called every time the next player has a turn
- if I apply it to the first player only (ie comment out if player 2) - it correctly runs and plays a move (before stopping on player 2 with no signal to prompt the next move)
- as soon as I apply it to both players, and press play, the game hangs, not even running
Environment:
Unity 2019.3.15f1
MLAgents 0.17.0
Tensorflow 1.5.0
Mac OS 10.12.3
Metadata
Metadata
Assignees
Labels
bugIssue describes a potential bug in ml-agents.Issue describes a potential bug in ml-agents.