denoising-diffusion-pytorch

denoising-diffusion-pytorch copied to clipboard

Free some resources after each step to avoid OOM

7

Fixes #74

andrewmoise

Comparison between P2-Gamma and No P2-Gamma

2

Hi Lucidrains, Well a bit late but now I successfully conducted a test between using P2-Gamma while training and one without I made a little comparison video: https://www.youtube.com/watch?v=3k68aI5QdVU (Gamma weight...

Mut1nyJD

Running out of memory

1

Any advice on how to deal with running out of GPU memory? I'm just getting started with pytorch / this package, and this is what happens when I try an...

andrewmoise

Running on CIFAR 10

14

Hi, I am trying to train and sample using CIFAR 10 dataset. Below is the code for it. ``` from keras.datasets import mnist import torch from denoising_diffusion_pytorch import Unet, GaussianDiffusion,...

DushyantSahoo

the training for the current code on cifar10 is overfitting. It doesn't get any good FID close to the numbers reported in the paper. It seems the paper uses dropout...

MortezaMardani

Are there any formulation or something?

Josh00-Lu

clip_denoised purpose?

10

Hello, I was wondering why we have to clamp the output of x_start from predict_start_from_noise( ) when sampling: ``` def p_mean_variance(self, x, t, clip_denoised: bool): model_output = self.denoise_fn(x, t) if...

malekinho8

getting NaN loss

12

I tried two datasets that consists of 10k and 30k images respectively. Can you please tell me if I need to change any hyperparameter?

ankanbhunia

Information about FID metric to evaluate the model

1

I trained the network using the training images of Cifar10's 'horse' class (5000 samples). I want to compute FID metric to evaluate the model, computing the "distance" between my generated...

rinodipaola15

Size different from 128

10

Hi, I am trying to feed as argument image size=300 but I'm getting this error (my images are indeed of size 300) assert h == img_size and w == img_size,...

ethancohen123