Related papers: VFA: Relieving Vector Operations in Flash Attentio…

The Segment Anything Model (SAM) has advanced interactive segmentation but is limited by the high computational cost on high-resolution images. This requires downsampling to meet GPU constraints, sacrificing the fine-grained details needed…

We consider the problem of Visual Question Answering (VQA). Given an image and a free-form, open-ended, question, expressed in natural language, the goal of VQA system is to provide accurate answer to this question with respect to the…

Vision transformers have shown great success on numerous computer vision tasks. However, its central component, softmax attention, prohibits vision transformers from scaling up to high-resolution images, due to both the computational…

We propose novel attention architectures, Multi-matrix Factorization Attention (MFA) and MFA-Key-Reuse (MFA-KR). Existing variants for standard Multi-Head Attention (MHA), including SOTA methods like MLA, fail to maintain as strong…

LLM decoding is bottlenecked for large batches and long contexts by loading the key-value (KV) cache from high-bandwidth memory, which inflates per-token latency, while the sequential nature of decoding limits parallelism. We analyze the…

The self-attention mechanism traditionally relies on the softmax operator, necessitating positional embeddings like RoPE, or position biases to account for token order. But current methods using still face length generalisation challenges.…

Transformer networks are rapidly becoming SotA in many fields, such as NLP and CV. Similarly to CNN, there is a strong push for deploying Transformer models at the extreme edge, ultimately fitting the tiny power budget and memory footprint…

Gating mechanisms have been widely utilized, from early models like LSTMs and Highway Networks to recent state space models, linear attention, and also softmax attention. Yet, existing literature rarely examines the specific effects of…

Multi-Head Latent Attention (MLA), introduced in DeepSeek-V2, improves the efficiency of large language models by projecting query, key, and value tensors into a compact latent space. This architectural change reduces the KV-cache size and…

While Transformer architecture excel at modeling long-range dependencies contributing to its widespread adoption in vision tasks the quadratic complexity of softmax-based attention mechanisms imposes a major bottleneck, particularly when…

Diffusion Transformers are fundamental for video and image generation, but their efficiency is bottlenecked by the quadratic complexity of attention. While block sparse attention accelerates computation by attending only critical key-value…

Attention calculation is extremely time-consuming for long-sequence inference tasks, such as text or image/video generation, in large models. To accelerate this process, we developed a low-precision, mathematically-equivalent algorithm…

This work focuses on an efficient Agile design methodology for domain-specific accelerators. We employ feature-by-feature enhancement of a vertical development stack and apply it to the TVM/VTA inference accelerator. We have enhanced the…

The efficiency of attention is important due to its quadratic time complexity. We enhance the efficiency of attention through two key contributions: First, we leverage the new FP4 Tensor Cores in Blackwell GPUs to accelerate attention…

Self-attention is at the heart of the popular Transformer architecture, yet suffers from quadratic time and memory complexity. The breakthrough FlashAttention algorithm revealed I/O complexity as the true bottleneck in scaling Transformers.…

Long-context LLMs increasingly rely on extended, reusable prefill prompts for agents and domain Q&A, pushing attention and KV-cache to become the dominant decode-time bottlenecks. While sparse attention reduces computation and transfer…

Diffusion models excel in high-fidelity image generation but face scalability limits due to transformers' quadratic attention complexity. Plug-and-play token reduction methods like ToMeSD and ToFu reduce FLOPs by merging redundant tokens in…

The softmax function is crucial in Transformer attention, which normalizes each row of the attention scores with summation to one, achieving superior performances over other alternative functions. However, the softmax function can face a…

This paper proposes Block-Filtered Long-Context Attention (BFLA), a training-free sparse prefill attention mechanism for long-context inference. BFLA adopts a two-stage design. In Stage 1, query and key sequences are compressed into coarse…

Transformer networks have emerged as the state-of-the-art approach for natural language processing tasks and are gaining popularity in other domains such as computer vision and audio processing. However, the efficient hardware acceleration…