Hi ,
I have implemented a custom Loss function using Tensorflow operations. I tested it and it was working fine. However, when I wanted to add this loss to my VAE model and then fit the model, I get errors. The technical details are in this link https://stackoverflow.com/questions/52311186/how-model-fit-works-in-keras.
So, I was wondering how the model.fit works in keras ? Does it instantiate a new session ?
I couldn't find any useful link for that.
Thank you in advance.
The keras official documentation says:
You can either pass the name of an existing loss function, or pass a TensorFlow/Theano symbolic function that returns a scalar for each data-point and takes the following two arguments:
- y_true: True labels. TensorFlow/Theano tensor.
- y_pred: Predictions. TensorFlow/Theano tensor of the same shape as y_true.
The actual optimized objective is the mean of the output array across all datapoints.
(https://keras.io/losses/)
You pass the loss function in the model.compile:
Example: model.compile( loss='mean_squared_error', optimizer='sgd' ).
Only then, after you compile you model, you can call the model.fit() operator.
But, if your question is mainly about how model.fit works, this link can help. Is the implementation of the training details of keras, includind the fit.
https://github.com/keras-team/keras/blob/master/keras/engine/training.py
Is this answer usefull?
Regards!
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs