TypeError: Tensor is unhashable. Instead, use tensor.ref() as the key.

[NikhilDupally]

' ' '
15
16
---> 17 model_new = surgeon.operate()
18 return model_new

/usr/local/lib/python3.6/dist-packages/tfkerassurgeon/surgeon.py in operate(self)
152 sub_output_nodes = utils.get_node_inbound_nodes(node)
153 outputs, output_masks = self._rebuild_graph(self.model.inputs,
--> 154 sub_output_nodes)
155
156 # Perform surgery at this node

/usr/local/lib/python3.6/dist-packages/tfkerassurgeon/surgeon.py in _rebuild_graph(self, graph_inputs, output_nodes, graph_input_masks)
264 # Call the recursive _rebuild_rec method to rebuild the submodel up to
265 # each output layer
--> 266 outputs, output_masks = zip(*[_rebuild_rec(n) for n in output_nodes])
267 return outputs, output_masks
268

/usr/local/lib/python3.6/dist-packages/tfkerassurgeon/surgeon.py in (.0)
264 # Call the recursive _rebuild_rec method to rebuild the submodel up to
265 # each output layer
--> 266 outputs, output_masks = zip(*[_rebuild_rec(n) for n in output_nodes])
267 return outputs, output_masks
268

/usr/local/lib/python3.6/dist-packages/tfkerassurgeon/surgeon.py in _rebuild_rec(node)
216 # Check for replaced tensors before any other checks:
217 # these are created by the surgery methods.
--> 218 if node_output in self._replace_tensors.keys():
219 logging.debug('bottomed out at replaced output: {0}'.format(
220 node_output))

/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/ops.py in hash(self)
724 if (Tensor._USE_EQUALITY and executing_eagerly_outside_functions() and
725 (g is None or g.building_function)):
--> 726 raise TypeError("Tensor is unhashable. "
727 "Instead, use tensor.ref() as the key.")
728 else:

TypeError: Tensor is unhashable. Instead, use tensor.ref() as the key.
' ' '

[anklebreaker]

Have the same issue too. TF 2.0.1 and Keras 2.3.1.

Was able to get around it by adding .experimental_ref() to several tensors (e.g. if node_output.experimental_ref() in self._replace_tensors.keys():). However, I ultimately got stuck with other problems down the line

[medicresource]

versions
tensorflow: 2.2.0
keras: 2.3.0-tf

I also had this error:
Tensor is unhashable. Instead, use tensor.ref() as the key.

I got around it by first disabling eager execution
tf.compat.v1.disable_eager_execution()

Then I ran into an error saying
TypeError: list indices must be integers or slices, not ListWrapper

In my case this was fixed by editing surgeon.py
under the BatchNormalization class
index[layers.axis] to index[3]
and
new_input_shape[new_layer.axis]' to 'new_input_shape[3]

[ferova]

versions
tensorflow: 2.2.0
keras: 2.3.0-tf

I also had this error:
Tensor is unhashable. Instead, use tensor.ref() as the key.

I got around it by first disabling eager execution
tf.compat.v1.disable_eager_execution()

Then I ran into an error saying
TypeError: list indices must be integers or slices, not ListWrapper

In my case this was fixed by editing surgeon.py
under the BatchNormalization class
index[layers.axis] to index[3]
and
new_input_shape[new_layer.axis]' to 'new_input_shape[3]

Did you get it to work without problem with the "Add layer" as in #7 ?

[devmike01]

I'm having this issue as well and tf.compat.v1.disable_eager_execution() is no longer working.
tensorflow: 2.12.0
numpy: 1.23.5

Update: I was able to solve it by adding .ref() to my tensor(e.g: loss.ref())