mittlin Goto Github PK

about_nlp

algorithm_interview_notes-chinese

2018/2019/校招/春招/秋招/算法/机器学习(Machine Learning)/深度学习(Deep Learning)/自然语言处理(NLP)/C/C++/Python/面试笔记

awesome-deep-learning-resources

Rough list of my favorite deep learning resources, useful for revisiting topics or for reference. I have got through all of the content listed there, carefully.

awesome-knowledge-distillation

Awesome Knowledge Distillation

caffe

Caffe: a fast open framework for deep learning.

caffe_test

用训练好的caffe网络模型进行数字，少量汉字，特殊字符（./等）的识别（总共有210类）

data-augmentation-review

List of useful data augmentation resources. You will find here some not common techniques, libraries, links to GitHub repos, papers, and others.

distilling-the-knowledge-in-neural-network

Teaches a student network from the knowledge obtained via training of a larger teacher network

faceswap-gan

A denoising autoencoder + adversarial loss for face swapping.

glibc

glibc mirror

hough_lines

用hough变换检测医疗报告单中间的分栏直线

lihang-code

《统计学习方法》的代码实现

mmdnn

MMdnn is a set of tools to help users inter-operate among different deep learning frameworks. E.g. model conversion and visualization. Convert models between Caffe, Keras, MXNet, Tensorflow, CNTK, PyTorch Onnx and CoreML.

multilabel

multilabel classification（caffe）

mxnet_test

Test mxnet with own trained model,用训练好的网络模型进行数字，少量汉字，特殊字符（./等）的识别（总共有210类）

pytorch-cyclegan-and-pix2pix

Image-to-image translation in PyTorch (e.g. horse2zebra, edges2cats, and more)

shell-scripts

批量数据预处理

synthtext

Code for generating synthetic text images as described in "Synthetic Data for Text Localisation in Natural Images", Ankush Gupta, Andrea Vedaldi, Andrew Zisserman, CVPR 2016.

text-classification-cnn-rnn

CNN-RNN中文文本分类，基于tensorflow

tutorials

Tutorials for using ONNX

unsupervised_image_clustering

The purpose of this study was to develop an algorithm that given a query image finds the “closest” entries to it on a database of images. All images files are of the same format, size and black and white, representing “meaningful shapes”. Throughout the development process, two versions were implemented and will be referred to as Version 1 and Version 2 throughout this paper. Version 1 processed each image by detecting distinct object within the matrix and caching their relevant attributes. The attributes of each database image were then compared to the attributes of each query image, and the k most similar database images were returned for each respective query. While Version 1 had an accuracy of 70\% on small datasets, it failed to scale in runtime on datasets of size 1000 query images and 1000 database images. A new approach was implemented, Version 2, which compared the difference between local binary patterns, and up to 84\% accuracy in 30 seconds on the larger provided dataset.

usbcam

capturing image through Usb2.0 Camera