Skip to content

cagneymoreau/DL4j_RoadMap

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

#Machine Learning RoadMap for DL4J

This is a collection of progressively harder projects that I learned on and you can review or fix. Some are missing flawed etc When I set out to learn I didnt plan on writing this so its a very ugly collections. Its sequential in nature so it should be a good roadmap for someone else.

If you want to add something to this please don't hesitate to let me know. Im trying to touch on each broad category of machine learning.

Some of the more interesting things don't work or are a work in progress. Also let me know if you want to team up on a harder project or find bugs here.

BigPicture.txt

If your a newb what to expect

Download the DL4J examples and test some stuff out for fun

https://github.com/eclipse/deeplearning4j-examples

Working with INDarrays and datasets

done

basic stat exercises

thought these would be cool

Feed Forward Linear Regression and minibatch

done

CSVClassifierLinear

not included I used the dl4j example repo

Feed Forward Logistic Regression and saving network

done

Debuging Basic - Pulling Weights, params & Gradients

done

Recurrent Regression and split DataSet

done with bug

Dual LSTM and csv manipulation

done two contrived examples of simple stock trader design

OverFitting/EarlyStopping & nontrivial split dataset

done stock example extended

Convolutions Cifar

not included I used the dl4j example repo

autoencoder/unsupervised

done stocks example

vae

network/hper params testing/tuning

convolution autoencoder

done celeb a dataset

GrowAble Gan and weight sharing and comp graph debug

super important idea because it shows how networks can start learning simple things and advance to more complicated without direct intervention -- Ony makes modern art at the moment

Reinforcement learning, custom loss functions, BenchMarking against examples

separate git https://github.com/cagneymoreau/ReinforcmentLearning

Convolution Layer Viewers

Use the two different style of networks and view the activations maps or the filters maximizations

image segmentation

Use the Oxford-IIIT Pets dataset to perform segmentation and classification opening tar.gz too U-Net

word object embedding

Attention, Transformer, pervasive attention and NLP

vectorization and basic concept of dimensional tunneling vs static structure https://www.youtube.com/watch?v=XrZ_Y4koV5A https://github.com/treo/dl4j_attention

sparse training rigl

Simplification/ Quantization

prado pqrnn realm quantization fp to int

transfering models

object detection ILR/ efficientdet/ affordance

pointcloud keypoint

About

dl4j roadmap

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages