#dit

0 stories this week · 1 this month · 8 in 6 months · mostly AI

Stories per day · 30 days

8 stories, newest first · one card per story

Wave Forcing co-designs generation training and execution for continuous video output achieving 1F1C with 147.6 DiT FPS on eight H200 GPUs.

NEPA-DiT-XL predicts embeddings each denoising step and reaches ImageNet FID 1.32 with about one-third of REPA's compute.

DC-SAE compresses images 32x and speeds diffusion training while beating DC-AE on ImageNet.

Improved distributional diffusion models reach 2.38 FID on ImageNet-256 with one from-scratch DiT.

Persistence Forcing uses specialized diffusion features to reach FID 1.63 on ImageNet 256.

Alibaba’s Qwen team released Qwen-Image-2.1, a 7B open-weight model for image generation and editing.

UltraTex enables 2048-resolution multi-view diffusion for 3D texturing with large training and inference speedups.

Alibaba's Qwen open-sourced Qwen-Image-2.1, a 7B-parameter unified text-to-image generation and editing model with native RGBA transparency.

Stories tagged #dit