摘要:
论文地址:https://arxiv.org/abs/1709.06030 1. 论文思想 利用强化学习,对网络进行裁剪,从Layer Removal和Layer Shrinkage两个维度进行裁剪。 一个是对层判断是否进行裁剪,一个是判断一层中的参数的裁剪。 2. 原理图 3. 实现细节 将层信息 阅读全文
摘要:
论文地址:https://arxiv.org/abs/1507.06149 1. 主要思想 权值矩阵对应的两列i,j,如果差异很小或者说没有差异的话,就把j列与i列上(合并,也就是去掉j列),然后在下一层中把第j行的权值累加在第i像。 这个过程就想象一下隐藏层中少一个单元,对权值矩阵的影响。 整体思 阅读全文
摘要:
参考论文:ImageNet Classification with Deep Convolutional Neural Networks 1.特点 1.1 ReLU Nonlinearity的提出 ReLU是非饱和非线性函数,f(x) = max(0, x),收敛速度比饱和激活函数快。 优缺点: Re 阅读全文
摘要:
1. Question Given a collection of distinct numbers, return all possible permutations. For example, [1,2,3] have the following permutations: [ [1,2,3], 阅读全文
摘要:
1. Question Implement next permutation, which rearranges numbers into the lexicographically next greater permutation of numbers. If such arrangement i 阅读全文
摘要:
1. Parameter pruning and sharing 1.1 Quantization and Binarization "Compressing deep convolutional networks using vector quantization" "Quantized conv 阅读全文
摘要:
论文地址:https://arxiv.org/abs/1602.07360 模型地址:https://github.com/DeepScale/SqueezeNet 1. 论文思想 提出一种新的卷积 组合方式 替代原来的3 3的卷积。类似于bottleneck layer减少参数数目。但是不太像Mo 阅读全文
摘要:
Convolutional Neural Networks "ImageNet Models" "Architecture Design" "Activation Functions" "Visualization" "Fast Convolution" "Low Rank Filter Appro 阅读全文
摘要:
1. Question Given an array of integers sorted in ascending order, find the starting and ending position of a given target value. Your algorithm's runt 阅读全文
摘要:
1. Question You are given an n x n 2D matrix representing an image. Rotate the image by 90 degrees (clockwise). Note: You have to rotate the image in 阅读全文