Issues

• Question about setting eta=1.0 in DDIM — does it reduce to DDPM? Could lower values like 0.5 be used?

  yk7333/d3po#24 · Filter by author hugluckylu opened on Nov 1, 2025

• question about "iterative N-round online approach”

  yk7333/d3po#23 · Filter by author KindredSpirithub opened on Jul 18, 2025

• GPU memory issues

  yk7333/d3po#22 · Filter by author zcyyyc opened on Mar 26, 2025

• Optimization issue/suggestion for "train_d3po.py"

  yk7333/d3po#21 · Filter by author AHHHZ975 opened on Jan 30, 2025
  2

• About the traning loss

  yk7333/d3po#20 · Filter by author nini0919 opened on Jan 10, 2025
  2

• After training diffusion model with human preference dataset.

  yk7333/d3po#18 · Filter by author giovanlee opened on Oct 2, 2024
  3

• Clarification on Human Preference JSON File

  yk7333/d3po#17 · Filter by author jxxtin opened on Sep 20, 2024
  1

• Does D3PO only available for DM which predicts X instead of noise (eps)?

  yk7333/d3po#15 · Filter by author imabackstabber opened on Jul 30, 2024
  1

• How to organize human feedback results into a JSON file?

  yk7333/d3po#14 · Filter by author Mr-Loevan opened on Jul 12, 2024
  3

• Pretrained weight release

  yk7333/d3po#13 · Filter by author Mer9ury opened on Jun 28, 2024
  1

• DPO with existed-images

  yk7333/d3po#12 · Filter by author guoyanan1g opened on Jun 22, 2024
  2

• Combine with dreambooth

  yk7333/d3po#9 · Filter by author thanhhung0112 opened on Feb 19, 2024
  1

• ValueError : Attemting to unscale fp16 Gradients

  yk7333/d3po#6 · Filter by author ADKoishi opened on Jan 16, 2024
  1