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
When I try to run the 'sortseq_mpa_visualization' demo, I get this error:
ValueError: Cannot assign value to variable ' discrete_agnostic_mp_10/d_k:0': Shape mismatch.The variable shape (100,), and the assigned value shape (10, 100) are incompatible.
This is likely because the demo models are still trained with code that uses the mpa measurement process, while MAVE-NN 2 otherwise uses discrete agnostic...
The text was updated successfully, but these errors were encountered:
Thanks for writing. Are you using the source code from the MAVE-NN2 branch or MAVE-NN? If there's an issue with the code from the main MAVE-NN's branch, it's probably due to updates with other packages (e.g., tensorflow or numpy), if it's MAVE-NN2, then it's very likely due to some mismatch between which source, MAVE-NN or MAVE-NN2, is being used. I'm planning to address both issues in this summer (I've known about some of these issues and have been meaning to fix some of these issues, but I've little to no bandwidth). Can you provide a starter notebook for me to test this with?
When I try to run the 'sortseq_mpa_visualization' demo, I get this error:
ValueError: Cannot assign value to variable ' discrete_agnostic_mp_10/d_k:0': Shape mismatch.The variable shape (100,), and the assigned value shape (10, 100) are incompatible.
This is likely because the demo models are still trained with code that uses the mpa measurement process, while MAVE-NN 2 otherwise uses discrete agnostic...
The text was updated successfully, but these errors were encountered: