Troubleshoot "Invalid input shape for input Tensor"

Question

I am trying to build a model with keras and tensorflow to play Go:

    training_data = encode_from_file_info(training_files)
    testing_data = encode_from_file_info(testing_files)

    input_shape = (1, 19, 19)
    model = Sequential(
        [
            keras.layers.Input(input_shape),
            keras.layers.ZeroPadding2D(padding=3, data_format='channels_first'),
        ]
    )

    model.compile(
        loss="categorical_crossentropy", optimizer="sgd", metrics=["accuracy"]
    )
    model.fit(
        training_data, batch_size=64, epochs=15, verbose=1, validation_data=testing_data
    )

I get the following error when I call model.fit():

Invalid input shape for input Tensor("Cast:0", shape=(None, 19, 19), dtype=float32). Expected shape (None, 1, 19, 19), but input has incompatible shape (None, 19, 19)

I have verified that the generators training_data and testing_data yield tuples of two ndarray with shape (1, 19, 19) and (361,).

I'm new to tensorflow and it is a black box to me. Obviously I'm missing something and making incorrect assumptions about my training data and the neural network. How do I troubleshoot issues like this? What tools are available to debug my model to find the shape mismatch?

Layers expect data to come in batches, so your generator needs to return (batch_size, 1, 19, 19). — xdurch0
– xdurch0, Commented Apr 10 at 5:19

Code-Apprentice · Accepted Answer · 2025-04-10 08:04:53Z

I ended up wrapping my data generator to batch the data:

def grouper(iterable, n, *, incomplete="fill", fillvalue=None):
    """Collect data into non-overlapping fixed-length chunks or blocks."""
    # grouper('ABCDEFG', 3, fillvalue='x') → ABC DEF Gxx
    # grouper('ABCDEFG', 3, incomplete='strict') → ABC DEF ValueError
    # grouper('ABCDEFG', 3, incomplete='ignore') → ABC DEF
    iterators = [iter(iterable)] * n
    match incomplete:
        case "fill":
            return itertools.zip_longest(*iterators, fillvalue=fillvalue)
        case "strict":
            return zip(*iterators, strict=True)
        case "ignore":
            return zip(*iterators)
        case _:
            raise ValueError("Expected fill, strict, or ignore")


def batches(data, batch_size):
    for batch in grouper(data, batch_size):
        features = []
        labels = []
        for feature, label in batch:
            features.append(feature)
            labels.append(label)
        yield np.array(features), np.array(labels)

And then changing the call to model.fit:

    model.fit(
        batches(training_data, batch_size),
        epochs=15,
        verbose=1,
        validation_data=batches(testing_data, batch_size),
    )

Collectives™ on Stack Overflow

Troubleshoot "Invalid input shape for input Tensor"

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