How to reshape input data correctly for input in CNN model?

Question

I have a CNN model, which input image size is (150, 150). I want to feed array-like data for predict function (tensorflow) like this:

fig = plt.figure(figsize=(1.5, 1.5))
plt.plot(time_values, signal_values, '-o', c='r')
plt.axis('off')
fig.canvas.draw()
data = np.frombuffer(fig.canvas.tostring_rgb(), dtype=np.uint8)
data = data.reshape(fig.canvas.get_width_height()[::-1] + (3,))

And when I try:

CNN_model.predict(data)

I get an error:

WARNING:tensorflow:Model was constructed with shape (None, 150, 150, 3) for input Tensor("input_1:0", shape=(None, 150, 150, 3), dtype=float32), but it was called on an input with incompatible shape (None, 150, 3).
Traceback (most recent call last):

Why my shape is (None, 150, 3), but not (None, 150, 150, 3)?

In the last line, just directly write data.reshape((1, 150, 150, 3)) — Sreekant Shenoy
– Sreekant Shenoy, Commented Apr 20, 2021 at 5:18
@SreekantShenoy thank you a lot. As I understood, the firtst value here (None, 150, 150, 3) is equal to ((1, 150, 150, 3)) — dfdd
– dfdd, Commented Apr 20, 2021 at 5:28

Ewran · Accepted Answer · 2021-04-20 05:29:02Z

1

Just add a new dimension to your array data:

data = data[None, :]

Its shape will then be (1, 150, 150, 3), as expected by tensorflow.

answered Apr 20, 2021 at 5:29

Ewran

3284 silver badges15 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

How to reshape input data correctly for input in CNN model?

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