search

imatge-upc / salgan

SalGAN: Visual Saliency Prediction with Generative Adversarial Networks
https://imatge-upc.github.io/salgan
MIT License
371 stars 105 forks source link

A problem with using pre-trained Discriminator model "[SalGAN Discriminator (3.4 MB)]" #58

Open LJOVO opened 3 years ago

LJOVO commented 3 years ago

Hi,

I had tried to use the pre-trained Discriminator model "[SalGAN Discriminator (3.4 MB)]" but it always raises an error:

ValueError: mismatch: parameter has shape (12288L, 100L) but value to set has shape (12288L, 60L)

Does it mean the "fc4" layer needs to be changed from 100-D to 60-D?

But this let the structure become different from proposed in the paper.

I would be very grateful if you can help me to solve this problem.

Best, Ljovo