tqch/ddpm-torch

Issues

• Potential Bug in likelihood calculation

  #20 opened 5 months ago by HirahTang
  0

• KeyError: 'ema'

  #19 opened 7 months ago by doInfinitely
  1

• Is there any way to speed up training?

  #18 opened 8 months ago by nessessence
  0

• what changes would be needed if we want to use our own custom greyscale(channel=1) dataset having size of 256*256.

  #15 opened a year ago by awais00012
  1

• some checkpoint

  #16 opened a year ago by ShaochengShen
  2

• FID & Loss diverge on CIFAR10?

  #14 opened a year ago by tj-14
  2

• FID for CIFAR10 checkpoint

  #13 opened a year ago by tj-14
  2

• Bug in ddim.py

  #12 opened a year ago by Sirius11311
  1

• DDIM sampler training parameters

  #10 opened a year ago by h-bouzid
  2

• Question about FID calculation

  #11 opened a year ago by taehoon-yoon
  2

• Minor bug in Gaussian Diffusion implementation (q_mean_var)?

  #9 opened a year ago by SeminKim
  2

• Have you tested the FID for CelebaHQ?

  #8 opened a year ago by zhuyu-cs
  2

• pretrain

  #7 opened 2 years ago by Passenger12138
  1

• train my own image datasets

  #6 opened 2 years ago by WFLiu0327
  1

• Meet bugs when adopting the --eval choice for training.

  #5 opened 2 years ago by mo666666
  2

• Vector-conditioning using classifier-free guidance

  #1 opened 2 years ago by jorgejgnz
  4

• Comment in code is wrong. Upsample supports align_corners=True

  #4 opened 2 years ago by Skylion007
  2

• CIFAR 10 config

  #3 opened 2 years ago by rahulvigneswaran
  1

• Bug in target noise

  #2 opened 2 years ago by rahulvigneswaran
  1