Abstract:A neural network accelerated optimization method for FPGA hardware platform is proposed. The method realizes the optimized deployment of neural network algorithms for FPGA hardware platforms from three aspects: computational speed, flexible transplantation, and development methods. Replacing multiplication based on Mitchell algorithm not only breaks through the speed bottleneck of neural network hardware acceleration caused by long multiplication period, but also makes the parallel acceleration of neural network algorithm get rid of the dependence on the number of hardware multipliers in FPGA, which can give full play to the advantages of FPGA parallel acceleration and maximize the computing speed. Based on the configurable strategy of neural network parameters, the number of network layers and nodes within layers can be adjusted according to different logical resource of FPGA, improving the flexibility of neural network transplantation. The adoption of HLS development method overcomes the shortcomings of RTL method in designing complex neural network algorithms, such as high difficulty in development and long development cycle. Using the Cyclone V SE 5CSEBA6U23I7 FPGA as the target device, a parameter configurable BP neural network was designed based on the proposed method. The usage of logical resources such as ALUT, Flip-Flop, RAM, and DSP were 39.6%, 40%, 56.9%, and 18.3% of the pre-optimized ones, respectively. The feasibility of the proposed method was verified using MNIST digital recognition and facial recognition as application scenarios. Compare to pre-optimization, the test time of MNIST number recognition is reduced to 67.58%, and the success rate was lost 0.195%. The test time for facial recognition applications was reduced to 69.571%, and the success rate of combining LDA algorithm was lost within 4%.

Implementation and Optimization of the Accelerator Based on FPGA Hardware for LSTM Network

A Power-Efficient Accelerator Based on FPGAs for LSTM Network

A Near Memory Computing FPGA Architecture for Neural Network Acceleration

An Fpga-Based Lstm Acceleration Engine For Deep Learning Frameworks

An LSTM Acceleration Engine for FPGAs Based on Caffe Framework

An Instruction-Driven Batch-Based High-Performance Resource-Efficient LSTM Accelerator on FPGA

FPGA-based Accelerator for Long Short-Term Memory Recurrent Neural Networks

A Cloud Server Oriented FPGA Accelerator for LSTM Recurrent Neural Network

A High Energy-Efficiency FPGA-Based LSTM Accelerator Architecture Design by Structured Pruning and Normalized Linear Quantization

Cloud Server Oriented FPGA Accelerator for Long Short-Term Memory Recurrent Neural Networks

Design of Low Power LSTM Neural Network Accelerator Based on FPGA

Adaptive design and implementation of automatic modulation recognition accelerator

FPGA Acceleration of LSTM Based on Data for Test Flight.

A Spiking LSTM Accelerator for Automatic Speech Recognition Application Based on FPGA

Energy-Efficient And High-Throughput Fpga-Based Accelerator For Convolutional Neural Networks

E-RNN: Design Optimization for Efficient Recurrent Neural Networks in FPGAs

E-LSTM: an Efficient Hardware Architecture for Long Short-Term Memory

A neural network accelerated optimization method for FPGA

A compression strategy to accelerate LSTM meta-learning on FPGA

A Ubiquitous Machine Learning Accelerator With Automatic Parallelization on FPGA

F-LSTM: FPGA-Based Heterogeneous Computing Framework for Deploying LSTM-Based Algorithms