New Paper Proposes Late-Stage Pixel Refinement for Latent-Free Image Transformers
An arXiv paper titled 'Just Image Transformers with Late-Stage Pixel Refinement' builds on recent latent-free diffusion transformers that train image generators without a latent VAE, avoiding compression-related issues.
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-11-11 05:00 · arXiv cs.CV
Just Image Transformers with Late-Stage Pixel Refinement