minDiffusion icon indicating copy to clipboard operation
minDiffusion copied to clipboard

Self-contained, minimalistic implementation of diffusion models with Pytorch.

Results 7 minDiffusion issues
Sort by recently updated
recently updated
newest added

I'm having the following samples after training: - 200 epochs for 1 CELEBA image ![ddpm_1img_200epochs](https://user-images.githubusercontent.com/60674219/195911759-49cbdea3-4642-4b9c-a94f-b5bdf0fc82ba.png) - 100 epochs for 1000 CELEBA images ![ddpm_1000imgs_100epochs](https://user-images.githubusercontent.com/60674219/195911798-af59dc02-4be8-4764-9400-6a1db899c888.png) Shouldn't using only 1 image for training...

Hi! Im new to this stuff and don't understand how can i download a custom dataset and how to add it to this model? My questions are: what format should...

Hi! If I have a *.safetensors file, can I load it into minDiffusion? I mean can i use either checkpoint or lora downloaded from internet along with minDiffusion?

Why are you normalizing to 1.414 in unet.py? class Conv3(nn.Module): ... def forward(self, x: torch.Tensor) -> torch.Tensor: x = self.main(x) if self.is_res: x = x + self.conv(x) return x /...

Dear @cloneofsimo and @SilenceMonk, thanks so much for this code! It is beneficial and precisely the missing piece I need to understand diffusion models better. Also, I appreciated that I...

How to install `diffusion_utilities` in local mac or Colab