kirushyk / le Goto Github PK

Machine Learning Framework

License: MIT License

C 85.35% Meson 5.13% C++ 5.24% Python 0.40% Metal 0.16% Objective-C 3.26% Cuda 0.46%

machine-learning deep-learning svm linear-regression sequential-minimal-optimization gradient-descent meson c11 python sgd regression polynomial-regression mnist computer-vision

le's Introduction

Le - Machine Learning Framework.

Le is Machine Learning Framework designed so that programs using it will be easy to read. Library is written in pure C but in object-oriented way. Bindings to other languages are provided so Le can be used by C++, Rust and Python programs.

Le is now under heavy development. Please come back soon.

At this moment following ML models are implemented:

Polynomial Regression.
Support Vector Machines (SVM).
Sequential Feed-forward Neural Network (Multiple Layer Perceptron, MLP).
k-Nearest Neighbors Algorithm (k-NN).

Optimization algorithms supported:

Batch Gradient Descent (BGD).
Stochastic Gradient Descent (SGD) with momentum.
Sequential Minimal Optimization (SMO).

Supported backends:

NVIDIA CUDA.
Apple Metal.

Installation

From Source

Examples

Tools

License

Le is released under the MIT License.

le's People

Contributors

Stargazers

Watchers

Forkers

renguifeng neonkore onurgezgin2

le's Issues

Add simple RNN demo

Implement Neural Style Transfer Examples

We need it written in C, Python and Rust.

Add Windows support

Some ML Researches use Microsoft Windows.

Add more benchmarks

Implement Back Propagation Algorithm

We want Neural Networks to report gradients for Gradient Descent optimizers to use.

Sequential Neural Network: Add Output Dimensionality Check

Implement Autoencoder Demo

Implement feature mapping

Implement Python Interface

Port tools to GTK+ 4

Implement SMO

https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/tr-98-14.pdf

Make tests work in Meson way

Add Rust Examples

We need to provide demo code for those ones who will try our framework:

Polynomial Regression;
Regularization;
SVM;
MNIST-related stuff;
Dense Neural Network.

Playground: Separate training and test data sets

We need combobox element added to left part of UI to specify training and test sets size.

Add Python Examples

We need to provide demo code for those ones who will try our framework:

Polynomial Regression;
Regularization;
SVM;
MNIST-related stuff;
Dense Neural Network.

Implement Convolutional Layers

Implement Long Short-Term Memory (LSTM)

Implement Rust Bindings

Fix SGD on SNN with MNIST demo

At this moment, all gradients and weights going to NAN soon. First fixed error was stride not being taken into account when copying signal in Sequential model.

Implement F1 score calculation

Add logistic regression tests

Add automated tests for IDX reader

Fix MNIST test in case datafiles were not installed

Implement Gated Recurrent Unit (GRU)

Implement SVM

Implement Restricted Boltzmann Machine Demo

We need examples on all the languages supported.

Fix possible buffer overflow when printing tensors and shapes

Implement Pooling Layers

Add byte order check for IDX reader

Implement gradient checking

Sequential model needs gradient checking test.

Implement OpenBLAS support

We need some acceleration on Linux.

Implement L2 regularization

Rework Log Categories

Also, maybe use secure_getenv .

Could not compile on Ubuntu 20, could not run on MacOS :)

I fetched the project in case it can help me downsize my ML and still keep it portable. I was pleasantly surprise to see it build on MacOS after fetching bits and pieces from "brew", but the latest Ubuntu (tried both gcc9 and gcc10) gave me

/home/ubuntu/le/bin/../le/tensors/letensor.c:38: undefined reference to LE_ERROR'
`

On the Mac, I had no luck with the MNIST examples, it might be something trivial but I don't know at this stage

./mnist-snn Segmentation fault: 11

I also don't know if this is expected behavior

./polynomial-logistic-regression Train set: x = [1.000 2.000 3.000 4.000; 4.000 3.000 2.000 1.000] y = [0.000 0.000 1.000 1.000; 4.000 3.000 2.000 1.000] Assertion failed: (le_shape_equal(h->shape, y->shape)), function le_logistic_loss, file ../le/leloss.c, line 17. Iteration 0. Abort trap: 6