# Sitemap

A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.

## Markdown

This is a page not in th emain menu

## Mixed precision training

Published:

Source: https://forums.fast.ai/t/mixed-precision-training/20720

Published:

# 3D Detection Get Started

Published:

[TOC]

Published:

Published:

Published:

Published:

Published:

Published:

## Domain Adaptation - A Survey

Published:

Domain Adaptation 是迁移学习的一个分之，其主要解决数据分布有差异的问题，更精确地说就是 covariance shifts，前几天在小组内部做了一次分享，感觉也可以讲 ppt 的内容放到专栏中来，毕竟里面也不涉及任何跟公司有关的信息，也不是我在公司做的项目，而是结合之前的实习经历，以及自己平时的积累做的一个总结，因此希望能够对更多的人有所帮助。

Published:

## How Python’s import works

Published:

I’ve almost never been able to write correct Python import statements on the first go. Behavior is inconsistent between Python 2.7 and Python 3.6 (the two versions that I test here), and there is no single method for guaranteeing that imports will always work. This post is my dive into how to resolve common importing problems. Unless otherwise stated, all examples here work with both Python 2.7 and 3.6.

Published:

Published:

## Domain Adaptation From Past To Future

Published:

Please refer to this DA slides

## DenseBox

Published:

Depwise separable filters.

Published:

## A Survey on 3D object Detection.

Published:

• gray scale matching:

Published:

Published:

## 算法实习这一年（二）

Published:

2017.07 开始到现在（2018.03），贯穿始终的就三件事：实习、面试和搬家。

Published:

Published:

Published:

## Single Shot Multibox Detector

Published:

SSD 算法是一种直接 predict bounding box location by regression 和 predict object class by classification 的 object detection 算法，compared to Faster-RCNN，去掉了 bounding box proposal 以及后续的 pixel/feature resampling。运算速度比起 Faster RCNN 快很多，准确率也要高，it holds when compares to YOLO，当时拿到了 Pascal Object Detection 比赛的 top 1，不过现在已经又被 YOLO9000 等算法超越了。

Published:

## A Survery on Edge Detection

Published:

Published:

the summary file has been formated as a PDF.

Published:

## Backpropagation Through Time(BPTT)

Published:

Please refer to the document here.

Published:

• Time Series

Published:

Published:

Published:

[TOC]

Published:

## Neural Network Optimization Methods

Published:

For intuition, look at the below 2 gif first.

Published:

[TOC]

Published:

Published:

Published:

## Domain Adversarial Training of Neural Networks

Published:

Learning a discriminative classifier or other predictor in the presence of a shift between training and test distributions is known as domain adaptation(DA).

Published:

## Time delay neural network

Published:

TDNN 是一种 ANN 结构，其主要目的是处理 sequential data。TDNN 的 units 独立于时间位移（i.e. sequence position）识别特征，通常用于组建一个更大的模式识别系统。例如，将连续的音频转换为分类号的音素（phoneme）标签 stream 来做语音识别。

Published:

# Why use activation functions?

## GAN from 0 to 1

Published:

This post contains all you need to learn Generative Adversial Networks, including video, tutorials and related papers.

## TensorFlow RNN tutorial for Speech Recognition

Published:

In this post, we’ll provide a short tutorial for training a RNN for speech recognition; we’re including code snippets throughout, and you can find the accompanying GitHub repository here. The software we’re using is a mix of borrowed and inspired code from existing open source projects. Below is a video example of machine speech recognition on a 1906 Edison Phonograph advertisement. The video includes a running trace of sound amplitude, extracted spectrogram, and predicted text.

Published:

# TensorFlow Architecture

## Connectionist Temporal Classification

Published:

Demystifying the Connectionist Temporal Classification Loss

## Open Source Toolkit for Speech Recognition

Published:

Until a few years ago, the state-of-the-art for speech recognition was a phonetic-based approach including separate components for pronunciation, acoustic, and language models. Typically, this consists of n-gram language models combined with Hidden Markov models (HMM).

## Bidirectional RNNs

Published:

Reference: CS224d-lecture8

## End to End Models for Speech Processing

Published:

 CTC - a probabiliistic model p(**Y X**), where

Published:

[TOC]

Published:

[TOC]

Published:

## TensorBoard: Embedding Visualization

Published:

https://www.tensorflow.org/images/embedding-mnist.mp4

Published:

Published:

[TOC]

## word2vec - learning vector representations of words : word embeddings

Published:

In this tutorial we look at the word2vec model by Mikolov et al.

Published:

# Getting Started with MXNet

## Ubuntu 16.04 + CUDA8.0 + OpenCV 3.2.0 + Anaconda Python 3.6

Published:

I bought a new deep learning workstation 3 days ago, then I prepared required environment for doing DL stuff. The first thing I chose to do is installing OpenCV 3.1.

## MXNet Tutorials - NDArray, Symbols, Module, Iterator

Published:

In MXNet, NDArray is the core data sturcture for all mathmatical computations. AN NDArray represents a multidimensional fix-sized homogeneous array, just like numpy.ndarray. And it enables imperative computation.

Published:

[TOC]

Published:

## Configure OpenMP & MPI in Clion on Mac

Published:

I’ve been taking the course Parallel Computing this semester, therefore I want to configure Open MP and MPI on my Mac. This post describes how I configured OpenMP using Homebrew and built from MPI source successfully.

Published:

Petri 网起源：

Published:

## SIFT Theory and Practice

Published:

Matching features across different images in a common problem in computer vision. When all images are similar in nature (same scale, orientation, etc) simple corner detectors can work. But when you have images of different scales and rotations, you need to use the SIFT(Scale Invariant Feature Transform).

Published:

## Caffe Installation on OS X 10.11

Published:

A detailed steps of caffe installation.

Published:

Published:

# 1. Local Image Features

## A tutorial on PCA

Published:

This tutorial doesn’t shy away from explaining the ideas informally, nor does it shy away from the mathematics. It addresses both of them.

## CS231n-lecture1

Published:

CS231n focuses on one of the most importantproblems of visual recognition – image classification.

Published:

## C++ frequently asked questions in Interviews

Published:

[TOC]

Published:

1. 从 https://github.com/poodarchu/miwifi_ss 下载所需的文件，上传至小米路由器 3 中。
2. 更改 miwifi.sh 和 shadowsocks_r3.tar.gz 的权限
3. 执行 ./miwifi.sh
4. 输入代理服务器的配置信息，安装成功。
5. 输入 top 命令，可以看到 ss-dier 进程
6. 输入 /etc/init.d/ss start 开启 shadowsocks，输入 /etc/init.d/ss stop 终止进程。

Published:

Published:

Published:

[TOC]

Published:

Published:

Published:

1. 提前准备

Published:

Published:

• 威斯康星大学麦迪逊分校

Published:

Published:

# Deep Learning Work Station Specs

Published:

[TOC]

Published:

Published:

Published:

## t-SNE 高位数据可视化

Published:

t-distribution:

t-分布通常用于从小样本估计总体呈正态分布且方差未知的整体的均值。如果总体的方差已知，例如在样本数量足够多时，应该用正态分布。

Published:

` To save some of you some time, if you run something and you get an error like this:

## Decision Tree

Published:

What’s shown below is abstracted from Andrew Moore’s lab tutorials, thanks to him!

Published:

Published:

Published:

Published:

## Matrix Derivation - 矩阵求导

Published:

Y = A * X –> DY/DX = A’ Y = X * A –> DY/DX = A Y = A’ * X * B –> DY/DX = A * B’ Y = A’ * X’ * B –> DY/DX = B * A’

d(f*g)/dx=(df’/dx)g+(dg/dx)f’

## Mel Frequency Cepstral Coefficient (MFCC) tutorial

Published:

TERM: MFCC 梅尔频率倒谱系数

Published:

Published:

[TOC]

Published:

# Topic Model Implementation

Published:

Published:

coming。

Published:

## How Qt5, OpenCV 2.4 and Clion integrate with each other - On OS X?

Published:

Qt 版本：Qt 5.7

Clion 版本：Clion 2016.2

OpenCV 版本：OpenCV 2.4.13

## QT 的信号与槽机制介绍

Published:

QT 是一个跨平台的 C++ GUI 应用构架，它提供了丰富的窗口部件集，具有面向对象、易于扩展、真正的组件编程等特点，更为引人注目的是目前 Linux 上最为流行的 KDE 桌面环境就是建立在 QT 库的基础之上。QT 支持下列平台：MS/WINDOWS-95、98、NT 和 2000；UNIX/X11-Linux、Sun Solaris、HP-UX、Digital Unix、IBM AIX、SGI IRIX；EMBEDDED- 支持 framebuffer 的 Linux 平台。伴随着 KDE 的快速发展和普及，QT 很可能成为 Linux 窗口平台上进行软件开发时的 GUI 首选。

## K-Means聚类算法

Published:

K-Means 是最简单的无监督学习算法之一，用于解决我们熟知的聚类问题。整个过程基于一定的先验，将一个给定的数据集分成数个clusters（比如k个）。该算法的主要思想是

Published:

# 聚类算法指南

Published:

Published:

## Caffe Installation on Ubuntu 14.04 with CUDA 8.0 Guidance

Published:

A detailed steps of caffe installation on Ubuntu DeepLearning Workstation.

## CNN and Transfer Learning

Published:

CV related CNN and Transfer Learning.

Published:

Published:

Published:

K近邻算法

Published:

Published:

## vimrc

Published:

Custom vimrc for python users

## Det3D

Published:

A general 3D object detection framework featuring diverse models & datasets support and superior performance.

## cvpack2

Published:

A high efficient 2D object detection framework

## Presentation of the winner approach of the nuScenes 3D object detection challenge at WAD, CVPR 2019

Published:

Details can be found at WAD2019.