CODEWITHSUNDEEP
tensorflowkerastensorflow2.0keras-layer
Danial
Danial

Reputation: 612

is incompatible with the layer: expected axis -1 of input shape to have value 1 but received input with shape [None, 256, 256, 3]

I have model, that looks like this :

Model: "sequential_4"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d_170 (Conv2D)          (None, 256, 256, 32)      320       
_________________________________________________________________
batch_normalization_169 (Bat (None, 256, 256, 32)      128       
_________________________________________________________________
activation_166 (Activation)  (None, 256, 256, 32)      0         
_________________________________________________________________
conv2d_171 (Conv2D)          (None, 256, 256, 32)      9248      
_________________________________________________________________
batch_normalization_170 (Bat (None, 256, 256, 32)      128       
_________________________________________________________________
activation_167 (Activation)  (None, 256, 256, 32)      0         
_________________________________________________________________
max_pooling2d_35 (MaxPooling (None, 128, 128, 32)      0         


..............

But it gives me :

ValueError: Input 0 of layer sequential_4 is incompatible with the layer: expected axis -1 of input shape to have value 1 but received input with shape [None, 256, 256, 3]

Properties of my images :

print(imm.dtype)   # float32
print(imm.ndim)    # 3
print(imm.shape)   # (256, 256, 3)

This error gets raised at :

history = model.fit(
    x = train_x, y = train_y, 
    #batch_size=32, 
    #epochs=epochs, 
    #verbose=1, 
    #shuffle=True,
    #validation_split=0.2
)

Trace :

ValueError                                Traceback (most recent call last)
<ipython-input-36-bf5138504d79> in <module>()
      2 
      3 history = model.fit(
----> 4     x = train_x, y = train_y,
      5     #batch_size=32,
      6     #epochs=epochs,

When I remove a single comment from model fit, the error gets one line down.

Upvotes: 2

Views: 11607

Answers (2)

gihan
gihan

Reputation: 151

This model is missing the input layer. Start the model sequence with the input layer.

keras.layers.InputLayer(input_shape=(256, 256, 3))

Upvotes: 1

Uzzal Podder
Uzzal Podder

Reputation: 3205

The image has 3 channels, but the first layer has 32 channels. The first layer should have same channel as input image.

Would you please try by adding a new input layer at the beginning of the model (I mean before conv2d_170 layer).

keras.Input(shape=(256, 256, 3))

Upvotes: 2

Related Questions