ArcFlow Accelerates Diffusion Models 40x with 5% Params

Post LinkedIn

🧠Read original on 机器之心

#few-step-generation #lora-accelerationarcflow

💡40x faster diffusion inference w/ 5% params—game-changer for real-time image gen (code out now)

⚡ 30-Second TL;DR

What Changed

Enables few-step generation by drifting along teacher model's curved trajectories

Why It Matters

Dramatically reduces inference latency for diffusion models, enabling real-time AI image generation in production while minimizing compute costs and parameter overhead.

What To Do Next

Clone the ArcFlow GitHub repo and benchmark 40x speedup on your FLUX or Qwen diffusion models.

Who should care:Researchers & Academics

🧠 Deep Insight

Web-grounded analysis with 9 cited sources.

🔑 Enhanced Key Takeaways

•ArcFlow achieves 2-step (2 NFEs) text-to-image generation, reducing inference from multi-step teacher models to seconds while matching quality on benchmarks[1][2][3].
•Authors are from Fudan University (Zihan Yang, Shuyuan Tu, Licheng Zhang, Yu-Gang Jiang, Zuxuan Wu) and Microsoft Research Asia (Qi Dai), with paper published on arXiv as 2602.09014[3].
•Code, technical report, and basic model checkpoints released on GitHub on 2026-02-09, with upcoming ArcFlow-Qwen-20B and ArcFlow-FLUX-12B models planned[7].

🛠️ Technical Deep Dive

•ArcFlow parameterizes the velocity field as a mixture of continuous momentum processes to capture velocity evolution and form continuous non-linear trajectories within each denoising step[2][3].
•Uses analytical integration of non-linear trajectories to avoid numerical discretization errors, enabling high-precision approximation of teacher model paths[2][3].
•Implements trajectory distillation with lightweight LoRA adapters on teachers like Qwen-Image-20B (20B params) and FLUX.1-dev (12B params)[2][3][7].

🔮 Future ImplicationsAI analysis grounded in cited sources

ArcFlow LoRA adapters will integrate into existing diffusion pipelines without full retraining

LoRA enables lightweight adaptation to pre-trained models, allowing developers to add 2-step speedups to infrastructures like FLUX without overhauling systems[1][2].

2-NFE diffusion distillation will become standard for real-time image generation apps

40x speedup with preserved quality on large-scale models demonstrates viability for edge and production deployment beyond research[2][3].

⏳ Timeline

2026-02

ArcFlow paper published on arXiv (2602.09014)

2026-02-09

Code, technical report, and basic checkpoints released on GitHub

📎 Sources (9)

Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.

🧠Read original article on 机器之心

📰

Weekly AI Recap

Read this week's curated digest of top AI events →

👉Related Updates

Same topic

Explore #few-step-generation

Same product