Repo

FlashVSR_Plus

Runs one-step diffusion video super-resolution with untiled 4K inference using hybrid sparse attention and streaming decode options.

GitHub ↗Oct 14, 2026Paper ↗

  • cv
  • video-super-resolution
  • diffusion
  • sparse-attention
Qualitative SR comparisons vs baselines plus quality-vs-speed scatter showing core claims
Qualitative SR comparisons vs baselines plus quality-vs-speed scatter showing core claims

What it is

One-step diffusion video super-resolution that can run spatially untiled 4K on a single 24GB RTX 4090, with hybrid sparse attention and reward-based distillation. Upscale videos with a one-step diffusion pipeline that supports untiled 4K inference, streaming long videos, and selectable sparse-attention backends through a CLI or Python API.

Key results

  • Untiled 4K inference on a single 24GB RTX 4090
  • Hybrid attention gives up to 2× DiT acceleration
  • Cuts peak memory by 45% vs standard DiT attention

Shows that high-quality diffusion VSR can reach practical untiled 4K inference on a single consumer GPU.

One useful ML email a week.

Worth your time, not all of your time.

We track new ML work across papers, code, and data, then keep only the pieces worth a closer look. Selected, not collected.

Papers.Data.Code

The Weekly ML Signal

Selected, not collected.

One useful ML email a week: the papers, repos and datasets worth your attention.

Worth your time, not all of your time.