thushv89/attention_keras

Failed to convert object of type

mhamadli opened this issue · 3 comments

mhamadli commented

I'm trying to use this attention implementation to build a model for converting graphemes to phonemes, but I got this error:
TypeError: Failed to convert object of type <class 'list'> to Tensor. Contents: [1, Dimension(None)]. Consider casting elements to a supported type.

any help?

tikhonovpavel commented

Same error
I've tried on nmt model from examples directory on TF 1.15 and 1.12

Have you figured out why is that and how to solve this?

Rushabh1798 commented

I got the same error while using the attention_layer. Did you guys found the solution to it..??
I am using TF 2.0
please comment back the solution. Thanks..

mhamadli commented

I have found a solution for this problem, If you want to load a model that includes custom layers, custom classes or functions, you can pass them to the loading mechanism via the custom_objects.
I just added the following to my code and it worked:

    custom_objects = {
        'AttentionLayer': AttentionLayer
    }
    model = load_model(model_path, custom_objects=custom_objects)