tsycnh / Mlbasic
深度学习常用优化方法详解
Stars: ✭ 222
Programming Languages
python
139335 projects - #7 most used programming language
Labels
Projects that are alternatives of or similar to Mlbasic
Ai Simplest Network
The simplest form of an artificial neural network explained and demonstrated.
Stars: ✭ 333 (+50%)
Mutual labels: gradient-descent
Learning To Learn By Pytorch
"Learning to learn by gradient descent by gradient descent "by PyTorch -- a simple re-implementation.
Stars: ✭ 31 (-86.04%)
Mutual labels: gradient-descent
Awesome Machine Learning Deep Learning Mathematics
A curated list of mathematics documents ,Concepts, Study Materials , Algorithms and Codes available across the internet for machine learning and deep learning
Stars: ✭ 138 (-37.84%)
Mutual labels: gradient-descent
Neuralnetwork.net
A TensorFlow-inspired neural network library built from scratch in C# 7.3 for .NET Standard 2.0, with GPU support through cuDNN
Stars: ✭ 392 (+76.58%)
Mutual labels: gradient-descent
Mindseye
Neural Networks in Java 8 with CuDNN and Aparapi
Stars: ✭ 8 (-96.4%)
Mutual labels: gradient-descent
Physical Gradient Descent
Code for https://fosterelli.co/executing-gradient-descent-on-the-earth
Stars: ✭ 70 (-68.47%)
Mutual labels: gradient-descent
Neural Network From Scratch
Implementing Multiple Layer Neural Network from Scratch
Stars: ✭ 272 (+22.52%)
Mutual labels: gradient-descent
Ml Course
Starter code of Prof. Andrew Ng's machine learning MOOC in R statistical language
Stars: ✭ 154 (-30.63%)
Mutual labels: gradient-descent
Univariate Linear Regression Gradient Descent Javascript
⭐️ Univariate Linear Regression with Gradient Descent in JavaScript (Vectorized)
Stars: ✭ 9 (-95.95%)
Mutual labels: gradient-descent
Raytracer.jl
Differentiable RayTracing in Julia
Stars: ✭ 110 (-50.45%)
Mutual labels: gradient-descent
Optim
OptimLib: a lightweight C++ library of numerical optimization methods for nonlinear functions
Stars: ✭ 411 (+85.14%)
Mutual labels: gradient-descent
Cppnumericalsolvers
a lightweight C++17 library of numerical optimization methods for nonlinear functions (Including L-BFGS-B for TensorFlow)
Stars: ✭ 638 (+187.39%)
Mutual labels: gradient-descent
Machine Learning Numpy
Gathers Machine learning models using pure Numpy to cover feed-forward, RNN, CNN, clustering, MCMC, timeseries, tree-based, and so much more!
Stars: ✭ 90 (-59.46%)
Mutual labels: gradient-descent
Thesemicolon
This repository contains Ipython notebooks and datasets for the data analytics youtube tutorials on The Semicolon.
Stars: ✭ 345 (+55.41%)
Mutual labels: gradient-descent
Nmflibrary
MATLAB library for non-negative matrix factorization (NMF): Version 1.8.1
Stars: ✭ 153 (-31.08%)
Mutual labels: gradient-descent
Pytorch Meta Optimizer
A PyTorch implementation of Learning to learn by gradient descent by gradient descent
Stars: ✭ 275 (+23.87%)
Mutual labels: gradient-descent
Gmm
Variational Inference in Gaussian Mixture Model
Stars: ✭ 31 (-86.04%)
Mutual labels: gradient-descent
Far Ho
Gradient based hyperparameter optimization & meta-learning package for TensorFlow
Stars: ✭ 161 (-27.48%)
Mutual labels: gradient-descent
Backprop
Heterogeneous automatic differentiation ("backpropagation") in Haskell
Stars: ✭ 154 (-30.63%)
Mutual labels: gradient-descent
Neural Tangents
Fast and Easy Infinite Neural Networks in Python
Stars: ✭ 1,357 (+511.26%)
Mutual labels: gradient-descent
一些关于机器学习优化函数的练习
- p1: 最基本的梯度下降法：gradient descent
- p2: 原始 SGD方法：stochastic gradient descent
- p3: minibatch-SGD方法
- p4 momentum SGD: minibatch-SGD with momentum
- p4 momentum: momentum with SGD
- p5: Nesterov方法
- p6: adagrad
- p7: adadelta
- p8: adam
参考
p1 参考 https://zhuanlan.zhihu.com/p/27297638
p2~pn 参考 http://ruder.io/optimizing-gradient-descent/index.html
单独参考
p5 参考 http://cs231n.github.io/neural-networks-3/
p6 参考 https://zhuanlan.zhihu.com/p/22252270
p7 参考 https://arxiv.org/abs/1212.5701 (原始论文)
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at [email protected].