English
Related papers

Related papers: Hymba: A Hybrid-head Architecture for Small Langua…

200 papers

Slim attention shrinks the context memory size by 2x for transformer models with MHA (multi-head attention), which can speed up inference by up to 2x for large context windows. Slim attention is an exact, mathematically identical…

Machine Learning · Computer Science 2025-06-04 Nils Graef , Andrew Wasielewski

State space models (SSMs) have recently shown promising results on small-scale sequence and language modelling tasks, rivalling and outperforming many attention-based approaches. In this paper, we propose a multi-head state space (MH-SSM)…

Audio and Speech Processing · Electrical Eng. & Systems 2023-05-29 Yassir Fathullah , Chunyang Wu , Yuan Shangguan , Junteng Jia , Wenhan Xiong , Jay Mahadeokar , Chunxi Liu , Yangyang Shi , Ozlem Kalinli , Mike Seltzer , Mark J. F. Gales

State-space models (SSMs), such as Mamba (Gu & Dao, 2023), have been proposed as alternatives to Transformer networks in language modeling, by incorporating gating, convolutions, and input-dependent token selection to mitigate the quadratic…

Machine Learning · Computer Science 2024-04-26 Jongho Park , Jaeseung Park , Zheyang Xiong , Nayoung Lee , Jaewoong Cho , Samet Oymak , Kangwook Lee , Dimitris Papailiopoulos

Hybrid State-Space models combine Attention with recurrent State-Space Model (SSM) layers, balancing eidetic memory from Attention with compressed fading memory from SSMs. This yields smaller Key-Value caches and faster decoding than…

This paper introduces a novel approach to enhance the capabilities of Large Language Models (LLMs) in processing and understanding extensive text sequences, a critical aspect in applications requiring deep comprehension and synthesis of…

Computation and Language · Computer Science 2023-12-15 Kaiqiang Song , Xiaoyang Wang , Sangwoo Cho , Xiaoman Pan , Dong Yu

Large language models (LLMs) have advanced significantly due to the attention mechanism, but their quadratic complexity and linear memory demands limit their performance on long-context tasks. Recently, researchers introduced Mamba, an…

Computation and Language · Computer Science 2024-10-22 Wangjie You , Zecheng Tang , Juntao Li , Lili Yao , Min Zhang

Large language models (LLMs) struggle with maintaining coherence in extended conversations spanning hundreds of turns, despite performing well within their context windows. This paper introduces HEMA (Hippocampus-Inspired Extended Memory…

Computation and Language · Computer Science 2025-04-24 Kwangseob Ahn

Hybrid sequence models that combine efficient Transformer components with linear sequence modeling blocks are a promising alternative to pure Transformers, but most are still pretrained from scratch and therefore fail to reuse existing…

This paper unveils Dimba, a new text-to-image diffusion model that employs a distinctive hybrid architecture combining Transformer and Mamba elements. Specifically, Dimba sequentially stacked blocks alternate between Transformer and Mamba…

Computer Vision and Pattern Recognition · Computer Science 2024-06-04 Zhengcong Fei , Mingyuan Fan , Changqian Yu , Debang Li , Youqiang Zhang , Junshi Huang

Efficient long-context language modeling remains a significant challenge in Natural Language Processing (NLP). While Transformers dominate language tasks, they struggle with long sequences due to quadratic computational complexity in…

Key-Value (KV) cache memory and bandwidth increasingly dominate large language model inference cost in long-context and long-generation regimes. Architectures such as multi-head latent attention (MLA) and hybrid sliding-window attention…

Computation and Language · Computer Science 2026-04-08 Zhen Cheng , Hao-Bo Yang , Wan-Yi Huang , Jin-Long Li

Large Language Models (LLMs) encounter significant performance bottlenecks in long-sequence tasks due to the computational complexity and memory overhead inherent in the self-attention mechanism. To address these challenges, we introduce…

Artificial Intelligence · Computer Science 2026-02-17 Ziming Wang , Xiang Wang , Kailong Peng , Lang Qin , Juan Gabriel Kostelec , Christos Sourmpis , Axel Laborieux , Qinghai Guo

The Mamba layer offers an efficient selective state space model (SSM) that is highly effective in modeling multiple domains, including NLP, long-range sequence processing, and computer vision. Selective SSMs are viewed as dual models, in…

Machine Learning · Computer Science 2024-04-02 Ameen Ali , Itamar Zimerman , Lior Wolf

In this report, we introduce Falcon-H1, a new series of large language models (LLMs) featuring hybrid architecture designs optimized for both high performance and efficiency across diverse use cases. Unlike earlier Falcon models built…

We present Jamba, a new base large language model based on a novel hybrid Transformer-Mamba mixture-of-experts (MoE) architecture. Specifically, Jamba interleaves blocks of Transformer and Mamba layers, enjoying the benefits of both model…

Attention is all we need as long as we have enough data. Even so, it is sometimes not easy to determine how much data is enough while the models are becoming larger and larger. In this paper, we propose HYDRA heads, lightweight pretrained…

Computation and Language · Computer Science 2021-09-14 Ha-Thanh Nguyen , Vu Tran , Tran-Binh Dang , Minh-Quan Bui , Minh-Phuong Nguyen , Le-Minh Nguyen

We present the first comprehensive study of latent multi-head attention (MLA) for small language models, revealing interesting efficiency-quality trade-offs. Training 30M-parameter GPT models on 100,000 synthetic stories, we benchmark three…

Computation and Language · Computer Science 2025-06-17 Sushant Mehta , Raj Dandekar , Rajat Dandekar , Sreedath Panat

The quadratic computational complexity of MultiHead SelfAttention (MHSA) remains a fundamental bottleneck in scaling Large Language Models (LLMs) for longcontext tasks. While sparse and linearized attention mechanisms attempt to mitigate…

Computation and Language · Computer Science 2025-12-19 Caner Erden

Transformers are the driving force behind today's Large Language Models (LLMs), serving as the foundation for their performance and versatility. Yet, their compute and memory costs grow with sequence length, posing scalability challenges…