Issues

• Adapting REPA-E to Discrete Autoregressive Modeling

  End2End-Diffusion/REPA-E#31 · Filter by author LinB203 opened on Jul 1, 2026

• Training collapse around 90k steps when reproducing REPA-E SD-VAE end-to-end recipe

  End2End-Diffusion/REPA-E#30 · Filter by author Jerry-Terrasse opened on Jun 5, 2026

• Clarification request: VAE Reconstruction Loss discrepancy between paper ( L MSE L MSE ​ ) and open-sourced config (L1)

  End2End-Diffusion/REPA-E#29 · Filter by author XIONGPEILIN opened on Apr 29, 2026

• How to correctly normalize VAE encoder latents before feeding them into SiT (inverse transform of denormalize_latents)?

  End2End-Diffusion/REPA-E#27 · Filter by author Zhou-Weichen opened on Nov 25, 2025

• Diffusion Checkpoints for T2I?

  End2End-Diffusion/REPA-E#26 · Filter by author zheyuanhu01 opened on Nov 12, 2025
  1

• Missing VAE Discriminators for T2I Models (Qwen, SD3.5, FLUX)?

  End2End-Diffusion/REPA-E#25 · Filter by author leonary opened on Oct 30, 2025

• Generative image appearance Noise Artifacts

  End2End-Diffusion/REPA-E#24 · Filter by author mumumu-wss opened on Oct 29, 2025