Related papers: DraftAttention: Fast Video Diffusion via Low-Resol…

Large language models (LLMs) now support extremely long context windows, but the quadratic complexity of vanilla attention results in significantly long Time-to-First-Token (TTFT) latency. Existing approaches to address this complexity…

Attention accounts for an increasingly dominant fraction of total computation during inference for mixture-of-experts (MoE) models, making efficient acceleration critical. Emerging domain-specific accelerators for large model inference are…

Attention is sparse in vision transformers. We observe the final prediction in vision transformers is only based on a subset of most informative tokens, which is sufficient for accurate image recognition. Based on this observation, we…

Point tracking aims to localize corresponding points across video frames, serving as a fundamental task for 4D reconstruction, robotics, and video editing. Existing methods commonly rely on shallow convolutional backbones such as ResNet…

The computing cost of transformer self-attention often necessitates breaking long documents to fit in pretrained models in document ranking tasks. In this paper, we design Query-Directed Sparse attention that induces IR-axiomatic structures…

We introduce SANA-Video, a small diffusion model that can efficiently generate videos up to 720x1280 resolution and minute-length duration. SANA-Video synthesizes high-resolution, high-quality and long videos with strong text-video…

Efficient video generation models are increasingly vital for multimedia synthetic content generation. Leveraging the Transformer architecture and the diffusion process, video DiT models have emerged as a dominant approach for high-quality…

Distributed attention is a fundamental problem for scaling context window for Large Language Models (LLMs). The state-of-the-art method, Ring-Attention, suffers from scalability limitations due to its excessive communication traffic. This…

Current high-resolution vision-language models encode images as high-resolution image tokens and exhaustively take all these tokens to compute attention, which significantly increases the computational cost. To address this problem, we…

Diffusion Transformers (DiTs) achieve state-of-the-art performance in high-fidelity image and video generation but suffer from expensive inference due to their iterative denoising structure. While prior methods accelerate sampling by…

Transformer-based diffusion models have recently superseded traditional U-Net architectures, with multimodal diffusion transformers (MM-DiT) emerging as the dominant approach in state-of-the-art models like Stable Diffusion 3 and Flux.1.…

State-of-the-art sparse attention methods for reducing decoding latency fall into two main categories: approximate top-$k$ (and its extension, top-$p$) and recently introduced sampling-based estimation. However, these approaches are…

Leveraging long contexts is crucial for advanced AI systems, but attention computation poses a scalability challenge. While scaled dot-product attention (SDPA) exhibits token sparsity, i.e. only a few pivotal tokens significantly contribute…

Attention-based models, exemplified by the Transformer, can effectively model long range dependency, but suffer from the quadratic complexity of self-attention operation, making them difficult to be adopted for high-resolution image…

Diffusion Language Models (DLMs) enable globally coherent, bidirectional, and controllable text generation, offering advantages over traditional autoregressive LLMs, while scaling to ultra-long sequences remains costly. Many existing…

Transformers are considered one of the most important deep learning models since 2018, in part because it establishes state-of-the-art (SOTA) records and could potentially replace existing Deep Neural Networks (DNNs). Despite the remarkable…

The increased model capacity of Diffusion Transformers (DiTs) and the demand for generating higher resolutions of images and videos have led to a significant rise in inference latency, impacting real-time performance adversely. While prior…

Vision Transformers and their variants have achieved remarkable success in diverse visual perception tasks. Despite their effectiveness, they suffer from two significant limitations. First, the quadratic computational complexity of…

While diffusion language models (DLMs) offer a promising alternative to autoregressive models (ARs), existing open-source DLMs suffer from high inference latency. This bottleneck is mainly due to the attention's quadratic complexity with…

This paper presents PipeFusion, an innovative parallel methodology to tackle the high latency issues associated with generating high-resolution images using diffusion transformers (DiTs) models. PipeFusion partitions images into patches and…