tensorflow probabitily wrong output type

Question

UP DATE and ANSWER: My error was using tf.dataset to feed my model like y_hat=model(x_tst) using numpy array data give me the good type of output and work as supposed

I try to run tensorflow probality, but get stuck on output type problem. Followinfg this exemple at cell #3, with

y_hat=model(x_tst)

the y_hat type is: tensorflow_probability.python.layers.internal.distribution_tensor_coercible._TensorCoercible

my input are dataset type build like that (2 heads model)

input_1 = tf.data.Dataset.from_tensor_slices(X)
input_2 = tf.data.Dataset.from_tensor_slices(Xphysio)
output = tf.data.Dataset.from_tensor_slices(y)
combined_dataset = tf.data.Dataset.zip(((input_1, input_2), output))
input_dataset = combined_dataset.batch(32)

but if I use directly (like in exemple)

y_hat=model(x_tst)

instead of (who seem work well and produce results):

y_hat=model.predict(x_tst)

I get this error

TypeError: Inputs to a layer should be tensors. Got: <BatchDataset element_spec=((TensorSpec(shape=(None, 120, 9), dtype=tf.float32, name=None), TensorSpec(shape=(None, 24), dtype=tf.float32, name=None)), TensorSpec(shape=(None,), dtype=tf.float32, name=None))>

If I use model.predict() the next step using tf probability don't work

Jonathan Roy · Accepted Answer · 2023-07-06 10:55:35Z

0

Finaly find it, to use y_hat=model(x_tst) instead of model.predict and get the good type, I need to feed the model with numpy array in lieu of tf.dataset.

answered Jul 6, 2023 at 10:55

Jonathan Roy

4531 gold badge7 silver badges24 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

tensorflow probabitily wrong output type

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related