beresandras

Hi! All except NNCLR? If I run the Colab notebook using any algorithm, it runs without an error. What changes did you make?

Okay, but there are no hyperparameters with the value of either 32 or 6 in the code (which is in the error message). What is your image shape? What is...

Try to find what the error message refers to, *what* has the shape of (32,) and (6,).

Hi! The difference between DDPM and DDIM is in their sampling process, their training is the same. If you set `stochasticity` to 1.0 in model.py/plot_images(), and sample for more steps...

Hi! Description of the schedule from the Improved DDPM paper:  f(0) = cos(s/(1+s)*pi/2)^2 If s is small, f(0) ~= cos(0)^2 = 1, so we can disregard the...

Hi! After 3-4 epochs you should already see some reasonable images. Did you see loss spikes during training? Did you change anything (network, dataset, resolution, hyperparams)? The loss usually stops...