We are working on the regression problem of 60 inputs and 60 outputs using Keras.
When creating a custom loss function, the size of y_true,y_pred is unknown and cannot be processed.
The batch size is 128 with 10,000 teacher data, 8,000 of which are training data and 2,000 are test data.
If the batch size is 128 and the output dimension is 60, wouldn't the size be [128,60]?
Would it be possible for you to try something like print(y_true.shape) in a custom loss function?
© 2024 OneMinuteCode. All rights reserved.