# SALAD: Achieve High-Sparsity Attention via Efficient Linear Attention Tuning for Video Diffusion Transformer

arXiv: https://arxiv.org/abs/2601.16515 · HTML: https://mlsanity.com/paper/2601.16515/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Placed in Video generation, one of 14 directions in Diffusion models.
- Builds on 37 of its 42 references that are in the map.
- First posted on arXiv 23 Jan 2026.

## Placed in

- Diffusion models / [Video generation](https://mlsanity.com/diffusion/video-generation/)

## Builds on (most cited first, 30 of 37)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer](https://mlsanity.com/paper/2408.06072/) (2024)
- [VBench: Comprehensive Benchmark Suite for Video Generative Models](https://mlsanity.com/paper/2311.17982/) (2023)
- [WorldSimBench: Towards Video Generation Models as World Simulators](https://mlsanity.com/paper/2410.18072/) (2024)
- [LLM-Pruner: On the Structural Pruning of Large Language Models](https://mlsanity.com/paper/2305.11627/) (2023)
- [Latte: Latent Diffusion Transformer for Video Generation](https://mlsanity.com/paper/2401.03048/) (2024)
- [Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention](https://mlsanity.com/paper/2502.11089/) (2025)
- [SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers](https://mlsanity.com/paper/2410.10629/) (2024)
- [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)
- [LongLoRA: Efficient Fine-tuning of Long-Context Large Language Models](https://mlsanity.com/paper/2309.12307/) (2023)
- [Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity](https://mlsanity.com/paper/2502.01776/) (2025)
- [MoBA: Mixture of Block Attention for Long-Context LLMs](https://mlsanity.com/paper/2502.13189/) (2025)
- [SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models](https://mlsanity.com/paper/2411.05007/) (2024)
- [VisionReward: Fine-Grained Multi-Dimensional Human Preference Learning for Image and Video Generation](https://mlsanity.com/paper/2412.21059/) (2024)
- [Koala-36M: A Large-scale Video Dataset Improving Consistency between Fine-grained Conditions and Video Content](https://mlsanity.com/paper/2410.08260/) (2024)
- [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)
- [ViDiT-Q: Efficient and Accurate Quantization of Diffusion Transformers for Image and Video Generation](https://mlsanity.com/paper/2406.02540/) (2024)
- [DiTFastAttn: Attention Compression for Diffusion Transformer Models](https://mlsanity.com/paper/2406.08552/) (2024)
- [Adaptive Caching for Faster Video Generation with Diffusion Transformers](https://mlsanity.com/paper/2411.02397/) (2024)
- [Sparse VideoGen2: Accelerate Video Generation with Sparse Attention via Semantic-Aware Permutation](https://mlsanity.com/paper/2505.18875/) (2025)
- [VSA: Faster Video Diffusion with Trainable Sparse Attention](https://mlsanity.com/paper/2505.13389/) (2025)
- [Q-DiT: Accurate Post-Training Quantization for Diffusion Transformers](https://mlsanity.com/paper/2406.17343/) (2024)
- [SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer](https://mlsanity.com/paper/2509.24695/) (2025)

## Built on by (most cited first, 7 of 7)

- [SVG-EAR: Parameter-Free Linear Compensation for Sparse Video Generation via Error-aware Routing](https://mlsanity.com/paper/2603.08982/) (2026)
- [Attend Locally, Remember Linearly: Linear Attention as Cross-Frame Memory for Autoregressive Video Diffusion](https://mlsanity.com/paper/2605.16579/) (2026)
- [RoPeSLR: 3D RoPE-driven Sparse-LowRank Attention for Efficient Diffusion Transformers](https://mlsanity.com/paper/2605.20659/) (2026)
- [RoLA: Rotary-Positioned Low-Rank Linear Attention for Efficient Diffusion Transformers](https://mlsanity.com/paper/2609.06712/) (2026)
- [RT-Lynx: Putting GEMM Sparsity in the Right Place for Diffusion Models](https://mlsanity.com/paper/2605.26632/) (2026)
- [SparkDiffusion: Mitigating the High-Sparsity Trap --- A Unified Framework for up to 265× Single-GPU Acceleration of Visual Generation](https://mlsanity.com/paper/2609.23153/) (2026)
- [iCATS: Fast Video Generation via Interaction-Aware Sparse Attention and Timestep-Adaptive Sparsity](https://mlsanity.com/paper/2610.11302/) (2026)

## Abstract (arXiv)

Diffusion Transformers have demonstrated remarkable performance in video generation. However, their long input sequences incur substantial latency due to the quadratic complexity of full attention. Various sparse attention mechanisms have been proposed. Training-free approaches are limited to moderate sparsity and thus yield only modest acceleration, whereas training-based methods can reach much higher sparsity but demand substantial data and computation. In this work, we propose SALAD, introducing a lightweight linear attention branch in parallel with the sparse attention. Leveraging a Multi-level Static-Dynamic Scaling Strategy to balance the two branches, our method attains up to 90% sparsity and 1.52-2.03x inference speedup across different models and sequence lengths, while maintaining generation quality comparable to the full attention baseline. Moreover, our finetuning process is highly efficient, requiring only 2,000 video samples, fewer than 1,600 training steps, and no more than 30 GPU hours with a batch size of 8.