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
Recurrent MADQN example is failing with shape errors on Pong environment
To Reproduce
Steps to reproduce the behavior:
1.python3 examples/tf/petting_zoo/atari/pong/recurrent/decentralised/run_madqn.py
Expected behavior
Nodes should run as normal and not be halted by ValueError: Dimensions must be equal, but are 19 and 32 for '{{node SelectV2}} = SelectV2[T=DT_FLOAT](Cast, strided_slice_5, SelectV2/e)' with input shapes: [19,32], [19,32,18], [].
Context (Environment)
Mava latest on dev branch as of 23-02-2022
Ubuntu 20.04 (not relevant)
Additional context
Possible Solution
Fix action masking for this environment
The text was updated successfully, but these errors were encountered:
Describe the bug
Recurrent MADQN example is failing with shape errors on Pong environment
To Reproduce
Steps to reproduce the behavior:
1.python3 examples/tf/petting_zoo/atari/pong/recurrent/decentralised/run_madqn.py
Expected behavior
Nodes should run as normal and not be halted by ValueError: Dimensions must be equal, but are 19 and 32 for '{{node SelectV2}} = SelectV2[T=DT_FLOAT](Cast, strided_slice_5, SelectV2/e)' with input shapes: [19,32], [19,32,18], [].
Context (Environment)
Additional context
Possible Solution
Fix action masking for this environment
The text was updated successfully, but these errors were encountered: