One question about the model2.py

[Panos13L]

In your model at model2.py after the 3rd max pooling layer we have a shape of (2,2,128) and then we have a reshape layers with shape (128,3) . How is this possible? Am i miss something?

[Panos13L]

Moreover, in your paper says that the input samples to the ConvNet are of the
following dimensions: the total number of trials (1,280) × 4 frames for
each video of the subjects × 16 width × 16 height × 1 depth, but in the code the images have depth=3.