CODEWITHSUNDEEP
pythontensorflowkeras
redress
redress

Reputation: 1439

Keras Model Misinterprets Input Data Shape

I have a generator, which yields the following:

yield {'ingredients': ingredients, 'documents': documents}, labels

The yield'd iterator has the following shape:

ingredients.shape (10, 46) documents.shape (10, 46) labels.shape (10,)

Once this iterator is fed through do my model, I get the following:

ValueError: Error when checking input: expected ingredients to have shape (1,) but got array with shape (46,)

Here is the model code which produces the above error:

    # Both inputs are 1-dimensional
    ingredients = Input(
        name='ingredients',
        shape=[1]
    )
    # ingredients.shape (?, 1) 
    documents = Input(
        name='documents',
        shape=[1]
    )
    # documents.shape (?, 1)

    logger.info('ingredients %s documents shape %s', ingredients.shape, documents.shape)

    ingredients_embedding = Embedding(name='ingredients_embedding',
                                      input_dim=training_size,
                                      output_dim=embedded_document_size)(ingredients) 

    # Embedding the document (shape will be (None, 1, embedding_size))
    document_embedding = Embedding(name='documents_embedding',
                                   input_dim=training_size,
                                   output_dim=embedded_document_size)(documents)

Upvotes: 0

Views: 30

Answers (1)

Shubham Panchal
Shubham Panchal

Reputation: 4289

The input_shape mentioned in the ingredients and documents Input layer is ( 1 ). But, the shape of ingredients is ( 10 , 46 ) and that of the documents is ( 10 , 46 ). Here 10 is the number of samples.

You are initializing the model to have an input of shape ( None , 1 ). It should be ( None , 46 ). Hence, you can make these changes.

ingredients = Input( name='ingredients', shape=( 46 , ) ) 
documents = Input( name='documents', shape=( 46 , )

This should fix the error. Actually speaking the input has 46 dimensions or 46 features.

Upvotes: 1

Related Questions