The Transformer architecture has revolutionized the field of natural language processing (NLP) through its selfattention mechanism. However, its high computational complexity and memory requirement present significant deployment challenges on resource-constrained edge devices. While existing research predominantly focuses on accelerating linear operations via model compression and approximation techniques, the inefficiencies and high deployment costs of nonlinear operations (e.g., Softmax and LayerNorm) remain critically understudied. Although some studies have attempted to mitigate these challenges through techniques such as kernel fusion and integer-only quantization, these approaches still suffer from partial fusion and inefficient quantization with retained division operations, leaving significant efficiency gains unexploited. To bridge these gaps, we propose a fully fused Transformer accelerator that co-optimizes both linear and nonlinear operations while minimizing memory bottlenecks. For linear computations, our design incorporates a deeply optimized compute engine featuring double buffering, an output-stationary tiling strategy, and DSP-packing technology to maximize throughput. For nonlinear operations, we introduce a delayed computation strategy for vector-wise operators, effectively reducing memory bandwidth pressure and dependency stalls. Furthermore, we propose a hardware-efficient, divisionfree integer-only quantization scheme, leveraging$\log 2$quantization for Softmax and a polynomial-enhanced approximation for LayerNorm to eliminate costly floating-point units, thereby significantly reducing latency and resource overhead. Through systematic design space exploration, our solution, deployed on the Zynq Z-7100 platform, achieves 1.376 TOPS for BERT inference, demonstrating a$\mathbf{1. 6 5 - 2. 5 1} \boldsymbol{\times}$higher computational efficiency compared to prior works.

Convolutional Neural Networks (CNNs) have demonstrated remarkable performance across various computer vision tasks. Due to the computational and data-intensive nature of CNNs, Field-Programmable Gate Arrays (FPGAs) are exceptionally well-suited for accelerating the CNN computation process. However, large-scale CNNs such as ResNet-84 contain an enormous number of parameters that exceed the capacity of a single FPGA, rendering the deployment on a single device impractical. In this paper, we develop an automated end-to-end design flow for mapping large-scale CNNs across multiple FPGAs, on the basis of the HLS4ML dataflow architecture. We propose a graph optimization method to streamline the CNN structure and reduce resource consumption. We also summarize a resource allocation algorithm that automatically determines the specific hardware resources necessary for each CNN layer. Furthermore, we introduce a partitioning methodology capable of effectively segmenting CNNs into multiple FPGAs, and providing each subgraph with specific interfaces to support communication between different FPGAs. To validate the methodology, we construct a multi-FPGA platform interconnected via LVDS. We select two typical networks, ResNet-8 and ResNet-84, as the benchmarks for evaluation. The experimental results demonstrate that our approach significantly outperforms existing solutions. It attains an 18.6-fold increase in speed over Vitis AI, a 2.2-fold improvement over FINN, and a 3.4-fold enhancement over the original single-FPGA HLS4ML implementation for ResNet-8. For ResNet-84, our method achieves a remarkable 33.6-fold speedup over Vitis AI. Additionally, when compared to other non-automated multi-FPGA solutions, our methodology still exhibits significant performance improvements.

Motor imagery (MI), as one of the important applications of brain-computer interface (BCI), has lately received great attention. However, current MI researches have not provided satisfactory representations of electroencephalogram (EEG), taking account of the space-time-frequency features for MI classification. Moreover, those models also lack the exploration of attentive spatial, temporal, and spectral dynamics. In this study, we propose TA3D (Two-stream Attention based 3D network), a novel model for MI classification. It mainly consists of two streams: the space-time stream and the space-frequency stream, representing and learning discriminative features in the space-time-frequency dimension. Specifically, each stream contains three key parts: 1) 3D representations of EEG signals depict the spatial information over temporal/spectral distributions; 2) Attention mechanisms adaptively explore attentive dynamics of EEG signals and focus on the most valuable information in separate dimensions; 3) 3D convolutions learn spatial representation, temporal dependence, and spectral dependence. The outputs of the two streams are concatenated for space-time-frequency feature fusion. Extensive experiments implemented on two BCI datasets demonstrate that our model outperforms state-of-the-art MI classification methods.