C5_W1_Assignment3- music_inference_model error - ValueError: Layer lstm expects 13 inputs

I am hitting this error in the Jazz music assignment . Researched the forum and tried out different stuffs. Can I please get some help in resolving this? The expected parameter goes up with every run.

**ValueError: Layer lstm expects 13 inputs, but it received 3 input tensors. Inputs received: [<tf.Tensor 'input_8:0' shape=(None, 1, 90) dtype=float32>, <tf.Tensor 'a0_7:0' shape=(None, 64) dtype=float32>, <tf.Tensor 'c0_7:0' shape=(None, 64) dtype=float32>]**

Please post a screen capture image that provides more context (the entire error message, not just that one line).

Please have the error details

Did you try restarting the kernel and re-running all of the cells in the notebook?

I ask because in this notebook, LSTM_cell is a global instance of an LSTM layer. So any changes you make to LSTM_cell in any of the functions throughout the notebook means you have to re-run the whole notebook again.

For example:
Let’s say you modify LSTM_cell in the djmodel() function - that will impact any code in music_inference_model() that uses LSTM_cell. So you have to re-run everyrthing to get them back in sync.

LSTM_cell is created way back in Section 2 of the notebook, before you worked on djmodel().

So that’s the first thing to check.

I had tried that previously . But did that now again. I get a different error now.

ARNING:tensorflow:Functional inputs must come from `tf.keras.Input` (thus holding past layer metadata), they cannot be the output of a previous non-Input layer. Here, a tensor specified as input to "functional_3" was not an Input tensor, it was generated by layer repeat_vector_99.
Note that input tensors are instantiated via `tensor = tf.keras.Input(shape)`.
The tensor that caused the issue was: repeat_vector_99/Tile:0
WARNING:tensorflow:Functional inputs must come from `tf.keras.Input` (thus holding past layer metadata), they cannot be the output of a previous non-Input layer. Here, a tensor specified as input to "functional_3" was not an Input tensor, it was generated by layer lstm.
Note that input tensors are instantiated via `tensor = tf.keras.Input(shape)`.
The tensor that caused the issue was: lstm/PartitionedCall_129:2
WARNING:tensorflow:Functional inputs must come from `tf.keras.Input` (thus holding past layer metadata), they cannot be the output of a previous non-Input layer. Here, a tensor specified as input to "functional_3" was not an Input tensor, it was generated by layer lstm.
Note that input tensors are instantiated via `tensor = tf.keras.Input(shape)`.
The tensor that caused the issue was: lstm/PartitionedCall_129:3
---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
<ipython-input-18-fe42db94b8ce> in <module>
      1 ### YOU CANNOT EDIT THIS CELL
----> 2 inference_model = music_inference_model(LSTM_cell, densor, Ty = 50)

<ipython-input-17-73811659398b> in music_inference_model(LSTM_cell, densor, Ty)
     56 
     57     # Step 3: Create model instance with the correct "inputs" and "outputs" (≈1 line)
---> 58     inference_model =  Model(inputs=[x, a, c], outputs=outputs)
     59 
     60     ### END CODE HERE ###

/opt/conda/lib/python3.7/site-packages/tensorflow/python/keras/engine/training.py in __new__(cls, *args, **kwargs)
    240       # Functional model
    241       from tensorflow.python.keras.engine import functional  # pylint: disable=g-import-not-at-top
--> 242       return functional.Functional(*args, **kwargs)
    243     else:
    244       return super(Model, cls).__new__(cls, *args, **kwargs)

/opt/conda/lib/python3.7/site-packages/tensorflow/python/training/tracking/base.py in _method_wrapper(self, *args, **kwargs)
    455     self._self_setattr_tracking = False  # pylint: disable=protected-access
    456     try:
--> 457       result = method(self, *args, **kwargs)
    458     finally:
    459       self._self_setattr_tracking = previous_value  # pylint: disable=protected-access

/opt/conda/lib/python3.7/site-packages/tensorflow/python/keras/engine/functional.py in __init__(self, inputs, outputs, name, trainable)
    113     #     'arguments during initialization. Got an unexpected argument:')
    114     super(Functional, self).__init__(name=name, trainable=trainable)
--> 115     self._init_graph_network(inputs, outputs)
    116 
    117   @trackable.no_automatic_dependency_tracking

/opt/conda/lib/python3.7/site-packages/tensorflow/python/training/tracking/base.py in _method_wrapper(self, *args, **kwargs)
    455     self._self_setattr_tracking = False  # pylint: disable=protected-access
    456     try:
--> 457       result = method(self, *args, **kwargs)
    458     finally:
    459       self._self_setattr_tracking = previous_value  # pylint: disable=protected-access

/opt/conda/lib/python3.7/site-packages/tensorflow/python/keras/engine/functional.py in _init_graph_network(self, inputs, outputs)
    182       # It's supposed to be an input layer, so only one node
    183       # and one tensor output.
--> 184       assert node_index == 0
    185       assert tensor_index == 0
    186       self._input_layers.append(layer)

AssertionError: 

image

Those are not the correct variables.

Thank you @TMosh that helped clearing all Unit tests .

Now in the predict_and_sample function I hit the below error and running out of ideas.

shape of pred - (1, 90)
shape of indices - (1,)
shape of results - (1, 90)
---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-65-0bfb6bdfdaab> in <module>
      3 results, indices = predict_and_sample(inference_model, x_initializer, a_initializer, c_initializer)
      4 
----> 5 print("np.argmax(results[12]) =", np.argmax(results[12]))
      6 print("np.argmax(results[17]) =", np.argmax(results[17]))
      7 print("list(indices[12:18]) =", list(indices[12:18]))

IndexError: index 12 is out of bounds for axis 0 with size 1```

This is how I process indices ,I suspect the shape of indices is incorrect for some reason. 

indices = np.argmax(pred, axis =-1)

That line of code is fine.

@TMosh I sent my code in DM. Please take a look.

@TMosh I was able to resolve the error. There was a issue in the definition of music_inference_model which I had to fix . Interestingly I had "All tests passed " for music_inference_model even with that issue.

Passing the tests in the notebook does not guarantee your code is perfect. The notebook only tests a few checkpoints.