explainingai-code/StableDiffusion-PyTorch

Issues

• How to improve the reconstruction of high-frequency details in the VQVAE training?

  #25 opened 3 months ago by vadori
  15

• Can this code use for image super-resolution or restoration?

  #27 opened 2 months ago by wendeyy
  6

• Image size in custom dataset

  #28 opened 2 months ago by MariaYsern
  2

• Error loading vgg weights

  #23 opened 3 months ago by Recoil77
  2

• Is there a way to configure this to train on Mel Spectrograms?

  #26 opened 3 months ago by Aman-Khokhar18
  1

• Why is `disc_fake_loss` applied twice in train_vqvae.py?

  #24 opened 3 months ago by whikwon
  2

• How to train a text-conditioned ldm for MNIST dataset?

  #19 opened 4 months ago by nusratdeeptee
  5

• How to train the model with the image2image mode

  #21 opened 4 months ago by dangdinh17
  2

• How to improve text-conditioned generation?

  #17 opened 5 months ago by Nikita-Sherstnev
  1

• How to modify config files to generate higher resolution images

  #13 opened 7 months ago by vavasthi
  8

• Losses for conditional diffusion models

  #15 opened 6 months ago by vinayak-sharan
  6

• ldm for inpainting

  #16 opened 6 months ago by ath90
  1

• How to condition based on multiple features?

  #14 opened 6 months ago by vinodrajendran001
  2

• Bug when saving Latent information?

  #10 opened 8 months ago by jpmcarvalho
  2

• Unable to run

  #12 opened 8 months ago by mognc
  29

• The training loss is not decreasing

  #11 opened 8 months ago by 100daggerz
  1

• Running out of memory... best number of samples for custom data sets?

  #8 opened 8 months ago by pgtinsley
  6

• Unexpected output after sampling using Conditional LDM

  #9 opened 8 months ago by randomaccess2023
  6

• Question about sampling data

  #7 opened 8 months ago by jpmcarvalho
  2

• Question - Why using VAE and not just UNet to predict the final output image?

  #6 opened 8 months ago by jpmcarvalho
  3

• why the ldm can't generate images well?

  #3 opened 9 months ago by awais00012
  2

• It wuold be greatly appreciated if model ckpts could be provided

  #5 opened 9 months ago by CatLoves
  3

• Why your model generated mnist images are noises?

  #4 opened 9 months ago by CatLoves
  3