# Astraea: A Token-wise Acceleration Framework for Video Diffusion Transformers

arXiv: https://arxiv.org/abs/2506.05096 · HTML: https://mlsanity.com/paper/2506.05096/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Placed in Video generation, one of 14 directions in Diffusion models.
- Builds on 34 of its 39 references that are in the map.
- First posted on arXiv 5 Jun 2025.

## Placed in

- Diffusion models / [Video generation](https://mlsanity.com/diffusion/video-generation/)
- Efficient LLM inference and serving / [Visual token pruning in multimodal models](https://mlsanity.com/inference/visual-token-pruning-in-multimodal-models/)

## Builds on (most cited first, 30 of 34)

- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [VBench: Comprehensive Benchmark Suite for Video Generative Models](https://mlsanity.com/paper/2311.17982/) (2023)
- [HunyuanVideo: A Systematic Framework For Large Video Generative Models](https://mlsanity.com/paper/2412.03603/) (2024)
- [Diffusion Models: A Comprehensive Survey of Methods and Applications](https://mlsanity.com/paper/2209.00796/) (2022)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [One-step Diffusion with Distribution Matching Distillation](https://mlsanity.com/paper/2311.18828/) (2023)
- [Open-Sora: Democratizing Efficient Video Production for All](https://mlsanity.com/paper/2412.20404/) (2024)
- [Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networks](https://mlsanity.com/paper/2102.00554/) (2021)
- [DeepCache: Accelerating Diffusion Models for Free](https://mlsanity.com/paper/2312.00858/) (2023)
- [Open-Sora Plan: Open-Source Large Video Generation Model](https://mlsanity.com/paper/2412.00131/) (2024)
- [Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model](https://mlsanity.com/paper/2411.19108/) (2024)
- [Token Merging for Fast Stable Diffusion](https://mlsanity.com/paper/2303.17604/) (2023)
- [Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity](https://mlsanity.com/paper/2502.01776/) (2025)
- [A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recommendations](https://mlsanity.com/paper/2308.06767/) (2023)
- [Structural Pruning for Diffusion Models](https://mlsanity.com/paper/2305.10924/) (2023)
- [Delta-DiT: A Training-Free Acceleration Method Tailored for Diffusion Transformers](https://mlsanity.com/paper/2406.01125/) (2024)
- [Accelerating Diffusion Transformers with Token-wise Feature Caching](https://mlsanity.com/paper/2410.05317/) (2024)
- [Cache Me if You Can: Accelerating Diffusion Models through Block Caching](https://mlsanity.com/paper/2312.03209/) (2023)
- [Real-Time Video Generation with Pyramid Attention Broadcast](https://mlsanity.com/paper/2408.12588/) (2024)
- [Neural Architecture Search: Insights from 1000 Papers](https://mlsanity.com/paper/2301.08727/) (2023)
- [Fast Video Generation with Sliding Tile Attention](https://mlsanity.com/paper/2502.04507/) (2025)
- [FasterCache: Training-Free Video Diffusion Model Acceleration with High Quality](https://mlsanity.com/paper/2410.19355/) (2024)
- [Adaptive Caching for Faster Video Generation with Diffusion Transformers](https://mlsanity.com/paper/2411.02397/) (2024)
- [BOOT: Data-free Distillation of Denoising Diffusion Models with Bootstrapping](https://mlsanity.com/paper/2306.05544/) (2023)
- [AutoDiffusion: Training-Free Optimization of Time Steps and Architectures for Automated Diffusion Model Acceleration](https://mlsanity.com/paper/2309.10438/) (2023)
- [Rethinking Token-wise Feature Caching: Accelerating Diffusion Transformers with Dual Feature Caching](https://mlsanity.com/paper/2412.18911/) (2024)
- [BlockDance: Reuse Structurally Similar Spatio-Temporal Features to Accelerate Diffusion Transformers](https://mlsanity.com/paper/2503.15927/) (2025)
- [SmoothCache: A Universal Inference Acceleration Technique for Diffusion Transformers](https://mlsanity.com/paper/2411.10510/) (2024)

## Built on by (most cited first, 11 of 11)

- [Sol Video Inference Engine: Agent-Native Full-Stack Acceleration Framework for Efficient Video Generation](https://mlsanity.com/paper/2606.23743/) (2026)
- [Efficient Video Diffusion Models: Advancements and Challenges](https://mlsanity.com/paper/2604.15911/) (2026)
- [Video Generation Models as World Models: Efficient Paradigms, Architectures and Algorithms](https://mlsanity.com/paper/2603.28489/) (2026)
- [MobileWan: Closing the Quality Gap for Mobile Video Diffusion](https://mlsanity.com/paper/2607.06173/) (2026)
- [No Cache Left Idle: Accelerating diffusion model via Extreme-slimming Caching](https://mlsanity.com/paper/2512.12604/) (2025)
- [ORBIS: Output-Guided Token Reduction with Distribution-Aware Matching for Video Diffusion Acceleration](https://mlsanity.com/paper/2605.22015/) (2026)
- [SCOPE: Subspace Clustering with Online Per-Head Top-K Estimation for Sparse Video Attention](https://mlsanity.com/paper/2608.12780/) (2026)
- [Where Compute Matters: Heterogeneous Attention for Efficient Video Diffusion](https://mlsanity.com/paper/2609.31050/) (2026)
- [RefAdapt-DiT: Adaptive Joint Attention for Reference-Conditioned Diffusion Transformers](https://mlsanity.com/paper/2609.32415/) (2026)
- [Carnator: Fast Text-to-Video Generation with Generation-Native Compatibility-Guided Cross-Request Reuse](https://mlsanity.com/paper/2609.32420/) (2026)
- [Sol-H3: Recursive Self-Improvement for MiniMax-H3 Inference Acceleration on Sol-Engine across Cloud and Edge](https://mlsanity.com/paper/2609.35110/) (2026)

## Abstract (arXiv)

Video diffusion transformers (vDiTs) have made tremendous progress in text-to-video generation, but their high compute demands pose a major challenge for practical deployment. While studies propose acceleration methods to reduce workload at various granularities, they often rely on heuristics, limiting their applicability. We introduce Astraea, a framework that searches for near-optimal configurations for vDiT-based video generation under a performance target. At its core, Astraea proposes a lightweight token selection mechanism and a memory-efficient, GPU-friendly sparse attention strategy, enabling linear savings on execution time with minimal impact on generation quality. Meanwhile, to determine optimal token reduction for different timesteps, we further design a search framework that leverages a classic evolutionary algorithm to automatically determine the distribution of the token budget effectively. Together, Astraea achieves up to 2.4× inference speedup on a single GPU with great scalability (up to 13.2× speedup on 8 GPUs) while achieving up to over 10~dB video quality compared to the state-of-the-art methods (<0.5% loss on VBench compared to baselines).