InaccessibleTensorError: The tensor 'Tensor("mul:0", shape=(), dtype=float32)' cannot be accessed here

[Crispy13]

In Chapter 12,

class ReconstructingRegressor(keras.models.Model):
    def __init__(self, output_dim, **kwargs):
        super().__init__(**kwargs)
        self.hidden = [keras.layers.Dense(30, activation="selu",
                                          kernel_initializer="lecun_normal")
                       for _ in range(5)]
        self.out = keras.layers.Dense(output_dim)
        # TODO: check https://github.com/tensorflow/tensorflow/issues/26260
        #self.reconstruction_mean = keras.metrics.Mean(name="reconstruction_error")

    def build(self, batch_input_shape):
        n_inputs = batch_input_shape[-1]
        self.reconstruct = keras.layers.Dense(n_inputs)
        super().build(batch_input_shape)

    def call(self, inputs, training=None):
        Z = inputs
        for layer in self.hidden:
            Z = layer(Z)
        reconstruction = self.reconstruct(Z)
        recon_loss = tf.reduce_mean(tf.square(reconstruction - inputs))
        self.add_loss(0.05 * recon_loss)
        #if training:
        #    result = self.reconstruction_mean(recon_loss)
        #    self.add_metric(result)
        return self.out(Z)

Making with the above code and training it raised error:

---------------------------------------------------------------------------
InaccessibleTensorError                   Traceback (most recent call last)
<ipython-input-19-455f20b90fc1> in <module>
      1 model = ReconstructingRegressor(1)
      2 model.compile(loss="mse", optimizer="nadam")
----> 3 history = model.fit(tf.random.normal((10, 10)), tf.random.normal((10, 1)), epochs=2)

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\keras\engine\training.py in fit(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)
   1098                 _r=1):
   1099               callbacks.on_train_batch_begin(step)
-> 1100               tmp_logs = self.train_function(iterator)
   1101               if data_handler.should_sync:
   1102                 context.async_wait()

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\eager\def_function.py in __call__(self, *args, **kwds)
    826     tracing_count = self.experimental_get_tracing_count()
    827     with trace.Trace(self._name) as tm:
--> 828       result = self._call(*args, **kwds)
    829       compiler = "xla" if self._experimental_compile else "nonXla"
    830       new_tracing_count = self.experimental_get_tracing_count()

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\eager\def_function.py in _call(self, *args, **kwds)
    869       # This is the first call of __call__, so we have to initialize.
    870       initializers = []
--> 871       self._initialize(args, kwds, add_initializers_to=initializers)
    872     finally:
    873       # At this point we know that the initialization is complete (or less

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\eager\def_function.py in _initialize(self, args, kwds, add_initializers_to)
    724     self._concrete_stateful_fn = (
    725         self._stateful_fn._get_concrete_function_internal_garbage_collected(  # pylint: disable=protected-access
--> 726             *args, **kwds))
    727 
    728     def invalid_creator_scope(*unused_args, **unused_kwds):

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\eager\function.py in _get_concrete_function_internal_garbage_collected(self, *args, **kwargs)
   2967       args, kwargs = None, None
   2968     with self._lock:
-> 2969       graph_function, _ = self._maybe_define_function(args, kwargs)
   2970     return graph_function
   2971 

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\eager\function.py in _maybe_define_function(self, args, kwargs)
   3359 
   3360           self._function_cache.missed.add(call_context_key)
-> 3361           graph_function = self._create_graph_function(args, kwargs)
   3362           self._function_cache.primary[cache_key] = graph_function
   3363 

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\eager\function.py in _create_graph_function(self, args, kwargs, override_flat_arg_shapes)
   3204             arg_names=arg_names,
   3205             override_flat_arg_shapes=override_flat_arg_shapes,
-> 3206             capture_by_value=self._capture_by_value),
   3207         self._function_attributes,
   3208         function_spec=self.function_spec,

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\framework\func_graph.py in func_graph_from_py_func(name, python_func, args, kwargs, signature, func_graph, autograph, autograph_options, add_control_dependencies, arg_names, op_return_value, collections, capture_by_value, override_flat_arg_shapes)
    988         _, original_func = tf_decorator.unwrap(python_func)
    989 
--> 990       func_outputs = python_func(*func_args, **func_kwargs)
    991 
    992       # invariant: `func_outputs` contains only Tensors, CompositeTensors,

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\eager\def_function.py in wrapped_fn(*args, **kwds)
    632             xla_context.Exit()
    633         else:
--> 634           out = weak_wrapped_fn().__wrapped__(*args, **kwds)
    635         return out
    636 

~\anaconda3\envs\ml\lib\site-packages\tensorflow\python\framework\func_graph.py in wrapper(*args, **kwargs)
    975           except Exception as e:  # pylint:disable=broad-except
    976             if hasattr(e, "ag_error_metadata"):
--> 977               raise e.ag_error_metadata.to_exception(e)
    978             else:
    979               raise

InaccessibleTensorError: in user code:

    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\keras\engine\training.py:805 train_function  *
        return step_function(self, iterator)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\keras\engine\training.py:795 step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:1259 run
        return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:2730 call_for_each_replica
        return self._call_for_each_replica(fn, args, kwargs)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:3417 _call_for_each_replica
        return fn(*args, **kwargs)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\keras\engine\training.py:788 run_step  **
        outputs = model.train_step(data)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\keras\engine\training.py:756 train_step
        y, y_pred, sample_weight, regularization_losses=self.losses)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\keras\engine\compile_utils.py:229 __call__
        reg_loss = math_ops.add_n(regularization_losses)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\util\dispatch.py:201 wrapper
        return target(*args, **kwargs)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\ops\math_ops.py:3572 add_n
        return gen_math_ops.add_n(inputs, name=name)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\ops\gen_math_ops.py:418 add_n
        "AddN", inputs=inputs, name=name)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\framework\op_def_library.py:750 _apply_op_helper
        attrs=attr_protos, op_def=op_def)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\framework\func_graph.py:588 _create_op_internal
        inp = self.capture(inp)
    C:\Users\eck\anaconda3\envs\ml\lib\site-packages\tensorflow\python\framework\func_graph.py:638 capture
        % (tensor, tensor.graph, self))

    InaccessibleTensorError: The tensor 'Tensor("mul:0", shape=(), dtype=float32)' cannot be accessed here: it is defined in another function or code block. Use return values, explicit Python locals or TensorFlow collections to access it. Defined in: FuncGraph(name=build_graph, id=2882832177736); accessed from: FuncGraph(name=train_function, id=2882832146952).

[GeorgeEngland]

Ran in Colab - getting the same problem. It's to do with the add_loss function. add_loss(0.05) throws the same error. add_loss(lambda: 0.05) doesn't throw an error

including the recon_loss always throws the error

[ageron]

Hi @GeorgeEngland ,

Thanks for your feedback. Unfortunately, due to an issue introduced in TF 2.2 (#46858), it is currently not possible to use add_loss() along with the build() method. So I updated the code in the notebook: I now create the reconstruct layer in the constructor instead of the build() method. Unfortunately, this means that the number of units in this layer must be hard-coded (alternatively, it could be passed as an argument to the constructor).

Here's the updated code:

class ReconstructingRegressor(keras.models.Model):
    def __init__(self, output_dim, **kwargs):
        super().__init__(**kwargs)
        self.hidden = [keras.layers.Dense(30, activation="selu",
                                          kernel_initializer="lecun_normal")
                       for _ in range(5)]
        self.out = keras.layers.Dense(output_dim)
        self.reconstruct = keras.layers.Dense(8) # workaround for TF issue #46858
        self.reconstruction_mean = keras.metrics.Mean(name="reconstruction_error")

    #Commented out due to TF issue #46858, see the note above
    #def build(self, batch_input_shape):
    #    n_inputs = batch_input_shape[-1]
    #    self.reconstruct = keras.layers.Dense(n_inputs)
    #    super().build(batch_input_shape)

    def call(self, inputs, training=None):
        Z = inputs
        for layer in self.hidden:
            Z = layer(Z)
        reconstruction = self.reconstruct(Z)
        recon_loss = tf.reduce_mean(tf.square(reconstruction - inputs))
        self.add_loss(0.05 * recon_loss)
        if training:
            result = self.reconstruction_mean(recon_loss)
            self.add_metric(result)
        return self.out(Z)