Nothing Special   »   [go: up one dir, main page]

×
Please click here if you are not redirected within a few seconds.
We need to support a massive amount of convolutional neural network (CNN) calculations with limited area and power consumption of data centers by using domain ...
We map this ROM to LUTs on an. FPGA. To reduce the number of LUTs, we have to reduce the number of inputs. Thus, we propose sparse local convolution. (SLC) ...
This problem is referred to as sparse nonnegative convolution, and has received considerable attention in the literature; the fastest algorithms to date run in ...
Bibliographic details on Design Method for an LUT Network-Based CNN with a Sparse Local Convolution.
Thus, we also propose an efficient sparse matrix multiplication algorithm. Based on the fact that the sparse convolutional kernels are fixed after training, we ...
People also ask
This paper proposes an FPGA-based CNN accelerator. The highly reusable accelerator function is designed to construct the optimized convolutional neural network ...
Missing: Method | Show results with:Method
This paper proposes an inference processor with a table look-up method that is 41 times faster than an Arduino embedded processor and introduces ternary ...
In this paper, a look-up table (LUT) based multiplier array is proposed for convolution in CNNs, where multiplications play the critical role. By exploring the ...
This paper proposes an FPGA-based CNN accelerator. The highly reusable accelerator function is designed to construct the optimized convolutional neural network.
This paper gives detailed insight about computation speed acceleration using Stochastic Gradient Decent(SGD) optimization, Fast convolution and exploiting ...