Good morning,
I have a problem with submitting my architecture which I trained using a custom loss function.
The first error I got said that I needed to insert my loss function as a custom object when loading the model, which I did even adding the exact definition of the function in the 'model' class, but now the error is 'unknown loss function'.
Hence I ask if you can explain how I have to manage my model.py file in the zip folder for submission.
Thank you very much!
Check for the custom_objects argument of the load_model function (https://www.tensorflow.org/api_docs/python/tf/keras/models/load_model)
Posted by: an2dl.competitions @ Nov. 24, 2022, 11:31 a.m.