相关论文: Flash Multi-Head Feed-Forward Network

Transformers with linear attention allow for efficient parallel training but can simultaneously be formulated as an RNN with 2D (matrix-valued) hidden states, thus enjoying linear-time inference complexity. However, linear attention…

Despite many recent works on Mixture of Experts (MoEs) for resource-efficient Transformer language models, existing methods mostly focus on MoEs for feedforward layers. Previous attempts at extending MoE to the self-attention layer fail to…

Recent work has shown that feed-forward networks (FFNs) in pre-trained Transformers are a key component, storing various linguistic and factual knowledge. However, the computational patterns of FFNs are still unclear. In this work, we study…

The explosive growth of Large Language Models (LLMs), such as GPT-4 with 1.8 trillion parameters, demands a fundamental rethinking of data center architecture to ensure scalability, efficiency, and cost-effectiveness. Our work provides a…

State-of-the-art results on neural machine translation often use attentional sequence-to-sequence models with some form of convolution or recursion. Vaswani et al. (2017) propose a new architecture that avoids recurrence and convolution…

While Transformer self-attention offers strong parallelism, the Key-Value (KV) cache grows linearly with sequence length and becomes a bottleneck for inference efficiency. Multi-head latent attention was recently developed to compress the…

As modern LLMs support thousands to millions of tokens, KV caches grow to hundreds of gigabytes, stressing memory capacity and bandwidth. Existing solutions, such as KV cache pruning and offloading, alleviate these but underutilize hardware…

Large language models (LLMs) are popular around the world due to their powerful understanding capabilities. As the core component of LLMs, accelerating Transformer through parallelization has gradually become a hot research topic. Mask…

Memory is fundamental to intelligence, enabling learning, reasoning, and adaptability across biological and artificial systems. While Transformer architectures excel at sequence modeling, they face critical limitations in long-range context…

Large language models (LLMs) have demonstrated exceptional proficiency in understanding and generating human language, but efficient inference on resource-constrained embedded devices remains challenging due to large model sizes and…

FPGA-based data processing in datacenters is increasing in popularity due to the demands of modern workloads and the ensuing necessity for specialization in hardware. Driven by this trend, vendors are rapidly adapting reconfigurable devices…

Multi-head attention is a driving force behind state-of-the-art transformers, which achieve remarkable performance across a variety of natural language processing (NLP) and computer vision tasks. It has been observed that for many…

Dominant pan-sharpening frameworks simply concatenate the MS stream and the PAN stream once at a specific level. This way of fusion neglects the multi-level spectral-spatial correlation between the two streams, which is vital to improving…

We study the problem of efficient generative inference for Transformer models, in one of its most challenging settings: large deep models, with tight latency targets and long sequence lengths. Better understanding of the engineering…

Federated meta-learning (FML) has emerged as a promising paradigm to cope with the data limitation and heterogeneity challenges in today's edge learning arena. However, its performance is often limited by slow convergence and corresponding…

The rapid growth of LLMs demands high-throughput, memory-capacity-intensive inference on resource-constrained edge devices, where single-batch decoding remains fundamentally memory-bound. Existing out-of-core GPU-based and SSD-like…

Current multi-modal image fusion methods typically rely on task-specific models, leading to high training costs and limited scalability. While generative methods provide a unified modeling perspective, they often suffer from slow inference…

Mainstream Transformer-based large language models face major efficiency bottlenecks: training computation scales quadratically with sequence length, and inference memory grows linearly, limiting long-context processing. Building large…

We propose a novel receiver architecture that preserves the performance benefits of multiport selection in fluid-antenna systems while requiring only a very small number of radio-frequency (RF) chains. The resulting fluid-antenna hybrid…

Vision Transformer(ViT) is now dominating many vision tasks. The drawback of quadratic complexity of its token-wise multi-head self-attention (MHSA), is extensively addressed via either token sparsification or dimension reduction (in…