dedhiaparth98 / capsule-network

A TensorFlow implementation of Capsule Network as described in the paper Dynamic Routing Between Capsules
https://towardsdatascience.com/implementing-capsule-network-in-tensorflow-11e4cca5ecae
MIT License
20 stars 15 forks source link

my dataset (200, 200, 3) #3

Open holyprb opened 6 months ago

holyprb commented 6 months ago

i tried to use it on my dataset, it contain 8 classes and the error said

ValueError: Input 0 of layer \"ConvolutionLayer\" is incompatible with the layer: expected axis -1 of input shape to have value 1, but received input with shape (32, 200, 200, 3)

Arguments received by CapsuleNetwork.call(): • inputs=['tf.Tensor(shape=(32, 200, 200, 3), dtype=float32)', 'tf.Tensor(shape=(32, 8, 10), dtype=float32)']"

i guess it goes wrong in the architecture, but i dont really know in which one. could you help me?