13 skills found
BoooC / CNN Accelerator Based On Eyeriss V2A Flexible and Energy Efficient Accelerator For Sparse Convolution Neural Network
f-dangel / UnfoldNd(N=1,2,3)-dimensional unfold (im2col) and fold (col2im) in PyTorch
gplhegde / Convolution FlavorsImplementation of convolution layer in different flavors
3outeille / CNNumpyA Numpy implementation of a Convolutional Neural Network: slow & fast (im2col/col2im).
OpenHero / Im2colimage to column
piojanu / CUDA Im2col ConvCUDA project for uni subject
f-dangel / EinconvConvolutions and more as einsum for PyTorch
blackccpie / Fastconvfast 2D convolution implementation benchmark
rioyokotalab / Im2col4D image/filter tensor -> 2d matrix
norxander / DE1 SoC HPSFPGAImage to column FPGA implementation (im2col by caffe)
anilsathyan7 / ConvAccAccelerating convolution using numba, cupy and xnor in python
ConvolutedDog / Implicit Im2col For BackpropagationSource code of "BP-Im2col: An implicit Im2col framework for efficient CNN backpropagation", accepted at ICCD 2022 and "CRSPU: Exploit Commonality of Regular Sparsity to Support Various Convolutions on Systolic Arrays", accepted at DATE 2023.
pgZhx / Cuda Parallel Program一个基于CUDA并行编程的项目，包含通用矩阵乘法（GEMM）的并行实现、直接卷积与im2col优化卷积的GPU实现，以及CUBLAS和cuDNN库的性能对比分析。通过实验探索不同并行计算策略的性能差异，并总结优化方法
