Issue trying to reproduce train_generator notebook

[fedeotto]

For now I have been trying to reproduce train_generator.ipnyb notebook. But once I get to run

vae.fit([X_train, y_train], X_train, batch_size=256, epochs=150, validation_split=0.2,callbacks=[reduce_lr,checkpoint,save_loss, DecoderSaveCheckpoint('ding_decoder_best.h5', decoder)])

the following issue appears:

TypeError: in user code:

    File "C:\Users\hp\AppData\Roaming\Python\Python38\site-packages\keras\engine\training.py", line 878, in train_function  *
        return step_function(self, iterator)
    File "C:\Users\hp\AppData\Roaming\Python\Python38\site-packages\keras\engine\training.py", line 867, in step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    File "C:\Users\hp\AppData\Roaming\Python\Python38\site-packages\keras\engine\training.py", line 860, in run_step  **
        outputs = model.train_step(data)
    File "C:\Users\hp\AppData\Roaming\Python\Python38\site-packages\keras\engine\training.py", line 809, in train_step
        loss = self.compiled_loss(
    File "C:\Users\hp\AppData\Roaming\Python\Python38\site-packages\keras\engine\compile_utils.py", line 239, in __call__
        self._loss_metric.update_state(
    File "C:\Users\hp\AppData\Roaming\Python\Python38\site-packages\keras\utils\metrics_utils.py", line 73, in decorated
        update_op = update_state_fn(*args, **kwargs)
    File "C:\Users\hp\AppData\Roaming\Python\Python38\site-packages\keras\metrics.py", line 177, in update_state_fn
        return ag_update_state(*args, **kwargs)
    File "C:\Users\hp\AppData\Roaming\Python\Python38\site-packages\keras\metrics.py", line 451, in update_state  **
        sample_weight = tf.__internal__.ops.broadcast_weights(
    File "C:\Users\hp\AppData\Roaming\Python\Python38\site-packages\keras\engine\keras_tensor.py", line 255, in __array__
        raise TypeError(

    TypeError: You are passing KerasTensor(type_spec=TensorSpec(shape=(), dtype=tf.float32, name=None), name='Placeholder:0', description="created by layer 'tf.cast_4'"), an intermediate Keras symbolic input/output, to a TF API that does not allow registering custom dispatchers, such as `tf.cond`, `tf.function`, gradient tapes, or `tf.map_fn`. Keras Functional model construction only supports TF API calls that *do* support dispatching, such as `tf.math.add` or `tf.reshape`. Other APIs cannot be called directly on symbolic Kerasinputs/outputs. You can work around this limitation by putting the operation in a custom Keras layer `call` and calling that layer on this symbolic input/output.

I have actually created a dedicated environment installing all dependencies using requirements.txt .

EDIT:

OK I think I have solved the issue in the following way:

before constructing the model just run

from tensorflow.python.framework.ops import disable_eager_execution

disable_eager_execution()

Since we are using a custom loss function, I have also specified

experimental_run_tf_function=False

in model.compile()