I am trying to use the input layer to compute a residue and have the loss function based on that residue. I am trying to use the method described in this answer, however when i try to call the fit method i get this error:
TypeError: in user code:
c:\python38\lib\site-packages\tensorflow\python\keras\engine\training.py:805 train_function *
return step_function(self, iterator)
c:\python38\lib\site-packages\tensorflow\python\keras\engine\training.py:795 step_function **
outputs = model.distribute_strategy.run(run_step, args=(data,))
c:\python38\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:1259 run
return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
c:\python38\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:2730 call_for_each_replica
return self._call_for_each_replica(fn, args, kwargs)
c:\python38\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:3417 _call_for_each_replica
return fn(*args, **kwargs)
c:\python38\lib\site-packages\tensorflow\python\keras\engine\training.py:788 run_step **
outputs = model.train_step(data)
c:\python38\lib\site-packages\tensorflow\python\keras\engine\training.py:755 train_step
loss = self.compiled_loss(
c:\python38\lib\site-packages\tensorflow\python\keras\engine\compile_utils.py:237 __call__
self._loss_metric.update_state(
c:\python38\lib\site-packages\tensorflow\python\keras\utils\metrics_utils.py:90 decorated
update_op = update_state_fn(*args, **kwargs)
c:\python38\lib\site-packages\tensorflow\python\keras\metrics.py:177 update_state_fn
return ag_update_state(*args, **kwargs)
c:\python38\lib\site-packages\tensorflow\python\keras\metrics.py:363 update_state **
sample_weight = weights_broadcast_ops.broadcast_weights(
c:\python38\lib\site-packages\tensorflow\python\ops\weights_broadcast_ops.py:155 broadcast_weights
values = ops.convert_to_tensor(values, name="values")
c:\python38\lib\site-packages\tensorflow\python\profiler\trace.py:163 wrapped
return func(*args, **kwargs)
c:\python38\lib\site-packages\tensorflow\python\framework\ops.py:1540 convert_to_tensor
ret = conversion_func(value, dtype=dtype, name=name, as_ref=as_ref)
c:\python38\lib\site-packages\tensorflow\python\framework\constant_op.py:339 _constant_tensor_conversion_function
return constant(v, dtype=dtype, name=name)
c:\python38\lib\site-packages\tensorflow\python\framework\constant_op.py:264 constant
return _constant_impl(value, dtype, shape, name, verify_shape=False,
c:\python38\lib\site-packages\tensorflow\python\framework\constant_op.py:281 _constant_impl
tensor_util.make_tensor_proto(
c:\python38\lib\site-packages\tensorflow\python\framework\tensor_util.py:435 make_tensor_proto
values = np.asarray(values)
c:\python38\lib\site-packages\numpy\core\_asarray.py:83 asarray
return array(a, dtype, copy=False, order=order)
c:\python38\lib\site-packages\tensorflow\python\keras\engine\keras_tensor.py:273 __array__
raise TypeError(
TypeError: Cannot convert a symbolic Keras input/output to a numpy array. This error may indicate that you're trying to pass a symbolic value to a NumPy call, which is not supported. Or, you may be trying to pass Keras symbolic inputs/outputs to a TF API that does not register dispatching, preventing Keras from automatically converting the API call to a lambda layer in the Functional Model.
This is the loss function i use
def custom_loss(inputs):
def loss(y_true, y_pred):
residue = y_true - inputs
return 1/2 * tf.math.squared_difference(residue, y_pred)
return loss
And this is the way i compile the model:
opt = tf.keras.optimizers.RMSprop(learning_rate=0.1, clipnorm=1)
model.compile(optimizer=opt,
loss=custom_loss(inputs),
metrics=[PSNR, SSIM])
The error arises when i call model.fit(). I want to point out that i am using a generator for a dataset, but I have been using the same approach with different models, that didn't require a custom loss function, and everything works well. It seems that the problem arises because i am using the keras.Input() layer in the loss function, but i don't know how to circumvent this error.
