Improvise_a_Jazz_Solo_with_an_LSTM_Network_v4 - music_inference_model

Course Q&A Deep Learning Specialization Sequence Models
1

Dear,
I have problem with Week 1 Assignment - music_inference_model
outputs=
for t in range(Ty)
a, _, c = LSTM_cell(inputs=x, initial_state=[a0, c0])
out = densor(a)
outputs.append(out)
x = tf.one_hot(tf.math.argmax(out,axis=-1), depth=n_values)
inference_model = Model([x0, a0, c0], outputs)

Error as below:
ValueError: Input 0 of layer lstm is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: [None, 90]

Anyone have experience on this issue? I took me almost 1 night to solve it but I still got stuck here.
Thank you

2

Two problems:
The initial state is [a, c], not [a0, c0]
You’re missing Step 2.E.

3

Thank you TMosh. Problems solved.