paper-with-me

홈 › Papers

Constant Memory Attention Block

2023-06-21 · Leo Feng, Frederick Tung, Hossein Hajimirsadeghi, Yoshua Bengio, Mohamed Osama Ahmed

Modern foundation model architectures rely on attention mechanisms to effectively capture context. However, these methods require linear or quadratic memory in terms of the number of inputs/datapoints, limiting their applicability in low-compute domains. In this work, we propose Constant Memory Attention Block (CMAB), a novel general-purpose attention block that computes its output in constant memory and performs updates in constant computation. Highlighting CMABs efficacy, we introduce methods for Neural Processes and Temporal Point Processes. Empirically, we show our proposed methods achieve results competitive with state-of-the-art while being significantly more memory efficient.

📄 PDF Abstract BibTeX arXiv:2306.12599

Code (0)

등록된 구현이 없습니다.

Tasks

Point Processes

Similar Papers 제목 키워드 기반

Memory Efficient Neural Processes via Constant Memory Attention Block

2023-05-23 · Leo Feng, Frederick Tung, Hossein Hajimirsadeghi, Yoshua Bengio 외

Neural Processes (NPs) are popular meta-learning methods for efficiently modelling predictive uncertainty. Recent state-of-the-art methods, however, leverage expensive attention mechanisms, limiting their applications, p…

Meta-Learning

SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer

2025-09-29 · Junsong Chen, Yuyang Zhao, Jincheng Yu, Ruihang Chu 외 arxiv

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…

Video GenerationVideo Alignment

Various Lengths, Constant Speed: Efficient Language Modeling with Lightning Attention

2024-05-27 · Zhen Qin, Weigao Sun, Dong Li, Xuyang Shen 외

We present Lightning Attention, the first linear attention implementation that maintains a constant training speed for various sequence lengths under fixed memory consumption. Due to the issue with cumulative summation o…

GPULanguage ModelingLanguage Modelling

Echo: KV-Cache-Free Associative Recall with Spectral Koopman Operators

2026-05-07 · Anupama Sridhar, Alexander Johansen arxiv

Long chain-of-thought reasoning and agentic tool-calling produce traces spanning tens of thousands of tokens, yet Transformer KV caches grow linearly with sequence length, creating a memory bottleneck on commodity hardwa…

Tensor Memory: Fixed-Size Recurrent State for Long-Horizon Transformers

2026-05-26 · Kabir Swain, Sijie Han, Daniel Karl I. Weidele, Mauro Martino 외 arxiv

Transformers process images and videos by flattening space and time into long token sequences. While attention and KV caching preserve past features, their memory grows with sequence length and they lack an explicit, per…