I am using TensorFlow and Keras to build a 3D CNN, where the 3rd dimension is time. I have a dataset which contains evolving images with time (51 frames were taken). For testing, I took 10 simulations. My dataset is thus composed of 10 samples which contain 51 frames each, and each frame is 64 x 64 x 1 (it's in grayscale).

My descriptor gives the conditions for the simulation (I had 5 conditions, e.g. temperature, concentration, etc.). I initially had a Y of shape (10, 51, 5) and I reshaped it to (510,5) as the MinMaxScaler required.

I try to create my train/test split and I get this error:

ValueError: Found input variables with inconsistent numbers of samples: [10, 510]

How should my image input shape look like? Each entry in my Y contains the timestep of the snapshot so I can't simply just make one entry correspond to all 51 frames.

I hope to use data augmentation to take more sub images per frame as well (e.g. 16 sub images; dataset of shape (10, 16, 51, 64, 64, 1). Any advice is welcome -- thanks in advance!

  • $\begingroup$ What are you using your model to predict? One value for each of your 10 samples? A value for each frame in each of your samples? $\endgroup$
    – Lynn
    Jul 23 at 5:31


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.