AnyFlow: Generate High‑Quality Video in 4 Steps and Keep Improving with More Sampling

AnyFlow introduces a flow‑map distillation framework that enables video diffusion models to produce high‑quality results in just four sampling steps while still gaining quality as the number of steps increases, supporting both causal and bidirectional architectures and scaling up to 14 B parameters.

bidirectional videocausal videofew-step generation

0 likes · 14 min read

AnyFlow: Generate High‑Quality Video in 4 Steps and Keep Improving with More Sampling

SuanNi

Apr 12, 2026 · Artificial Intelligence

How TDM‑R1 Achieves 4‑Step Image Generation that Beats 80‑Step Models

Researchers from HKUST, CUHK and XiaoHongShu introduced TDM‑R1, a reinforcement‑learning‑based method that enables 4‑step diffusion image generation to surpass 80‑step models in speed, fidelity, and complex instruction adherence, as demonstrated on the GenEval benchmark and multiple quality metrics.

AI image synthesisbenchmarkingdiffusion models

0 likes · 9 min read

How TDM‑R1 Achieves 4‑Step Image Generation that Beats 80‑Step Models

AIWalker

Apr 7, 2025 · Artificial Intelligence

TurboFill: High‑Quality Image Inpainting in Just 4 Steps

TurboFill introduces a fast image‑inpainting model that trains a repair adapter on a few‑step text‑to‑image diffusion backbone, achieving state‑of‑the‑art results with only four diffusion steps while dramatically reducing computational cost.

TurboFillcomputer visiondiffusion models

0 likes · 17 min read

TurboFill: High‑Quality Image Inpainting in Just 4 Steps