Cannot convert 'input_4' to EagerTensor of dtype float

Course Q&A TensorFlow: Advanced Techniques Specialization Generative Deep Learning with TensorFlow
1

error message says:

TypeError: Cannot convert ‘input_4’ to EagerTensor of dtype float

i can not figure out where is the error
< removed ipynb >

2

@Khaled_Mohammed,
Since your error is that the input type is wrong, I would suggest double-checking that you are passing the proper preprocessed image to inception() everywhere you call it.

Also, can you remove the assignment .ipynb from your post? It’s against community guidelines to share your code.

4

I also encountered the same error and was stuck for hours, the error message is just a hint but deceiving, you are most likely using the inception model constructor rather than the instance of the inception object within your succeeding functions.

2 Likes
5

Nice debugging, @BwandoWando! Thanks for sticking it out and finding the issue. And thanks for sharing here so others can benefit!

6

Welcome! Thank you as well. More power to DeepLearning.AI!

7

:+1: This tip was helpful. In my specific case I had to pay special attention to always invoke the same global model instance rather than the model builder methods.