Revision History
Diffusion Transformer (DiT) · 9 revisions
Sizes are character counts of the article source. The signed number is the change from the previous revision.
Recent edit summaries
Detailed summaries recorded by editors. Generic maintenance summaries are omitted here; every recorded revision remains below. Dates describe the edit, not necessarily the event it covers.
- Correction: Corrected the false claim that DiT won the ICCV 2023 best paper award (it was an oral; the Marr Prize went to ControlNet), replaced fabricated FID and parameter figures in the conditioning table with the paper's numbers, corrected the training hardware (TPU v3 pods, not V100/A100 nodes) and removed unsourced GPU-day estimates, fixed DiT and SD3 positional-encoding descriptions (sine-cosine, not learned or RoPE), and updated Bill Peebles (left OpenAI in April 2026 after Sora's shutdown) and Saining Xie (co-founder and CSO of AMI Labs) (plus 32 minor fixes)
Version 9 · Nov 2, 2026, 07:03 AM
- Update: added a video.
Version 8 · Nov 1, 2026, 08:01 AM
- Improve internal links (2026-10-13)
Version 7 · Oct 13, 2026, 10:19 AM
- Added 6 contextual internal links
Version 6 · Jul 23, 2026, 03:29 PM