Skip to content

📕 Curated list of ressources for learning state-of-the-art deep learning and machine learning!

Notifications You must be signed in to change notification settings

oyane806/curated-ressources-dl-ml

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 

Repository files navigation

Curated list of ressources for deep learning and machine learning - in progress

You can find below a curated list of ressources for deep learning and machine learning. It contains repos, articles, research papers and ebooks. I am trying to keep this list as minimal as possible!

python review

a python quick reference ⭐ article

deep learning ebook

deep learning reference book that you can download for free ⭐ ebook
to get familiar with mathematical notations, the hundred pages machine learning book ebook

random forest

Random forests are very good for classification but cannot extrapolate.
1️⃣ to get a visual intuition of how random forests work article
2️⃣ to code a random forest from scratch ⚙️⚙️ notebook
this is a very good exercise to brush up your coding skills and get an in-depth knowledge of how random forests work
3️⃣ random forest interpretation ⭐⭐⭐ (tree variance feature importance, partial dependence, tree interpreter) article

neural nets

  • activation functions: tanh, relu, leaky relu, sigmoid, softmax
  • initialization: He initialization, Xavier initialization
  • regularization: L2, L1, dropout, data augmentation, early stopping
  • optimization: batch gradient descent or vanilla gradient descent, stochastic gradient deschent, mini batch gradient descent, gradient descent with momentum, RMS prop, Adam optimization, learning rate decay, batch normalization

1️⃣ how to use cyclical learning rates for training neural networks ⭐ paper

time-series

1️⃣ great practical repo to get SOTA results quickly ⭐⭐ notebooks
2️⃣ quick walk through different time-series predictions methods (neural net, cnn, lstm, cnn-lstm, encoder-decoder lstm) article

SOTA algorithms

2017 hive-cote: ensemble of 37 classifiers (no neural net)
2019 rocket: using random convolutional kernels paper

image recognition

  • filter, padding, stride, convolution, pooling (max, average), fully connected layer

cnn architectures

year model size top-1 accuracy top-5 accuracy parameters depth
2012 AlexNet
2014 VGG16 528MB 0.713 0.901 138,357,544 23
2015 InceptionV3/GoogleNet 92MB 0.779 0.937 23,851,784 159
2015 ResNet50 98MB 0.749 0.921 25,636,712

segmentation

2015 unet paper

About

📕 Curated list of ressources for learning state-of-the-art deep learning and machine learning!

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published