I am using tensorboard with LSTMFCNClassifier and getting training data like i should. I instantiate it like this:

tensorboard_callback = tf.keras.callbacks.tensorboard(logdir,histogram_freq=1)

lstm = LSTMFCNClassifier(n_epochs=50, batch_size=4,callbacks=tensorboard_callback )

When i call fit:

lstm.fit(x_train,ytrain)

I get good data from tensorboard for the training. How do I run a validation set so that I can get validation data with my training data. From the docs I tried:

lstm.fit(X_train, y_train).predict(X_test)

This still only gives me training data. Is there function call to the LSTMFCNClassifier that can run validation data so that it gets logged in tensorboard?