numpy-neural-network

Сustom torch style machine learning framework with automatic differentiation implemented on numpy, allows build GANs, VAEs, etc.

Implementation of Adam Optimization algorithm using Numpy

deep learning from scratch. uses numpy/cupy, trains in GPU, follows pytorch API

training models with different optimizers using NumPy only. Featuring SGD, Adam, Adagrad, NAG, RMSProp, and Momentum. This repo also includes a benchmark against Pytorch developed optims.

Nice place to start seeing matrices shapes ! Great place for beginners to understand neural networks computation

This repository contains an implementation of a neural network from scratch using only NumPy, a fundamental library for numerical computing in Python. The neural network is designed to perform tasks such as classification, regression, or any other supervised learning problem.

𝗔𝗡𝗡 𝗮𝗻𝗱 𝗖𝗡𝗡 𝗳𝗿𝗼𝗺 𝘀𝗰𝗿𝗮𝘁𝗰𝗵 | 𝗡𝗼 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸𝘀, 𝗷𝘂𝘀𝘁 𝗽𝘂𝗿𝗲 𝗡𝘂𝗺𝗽𝘆

Implementation of artificial neural networks

Identify Handwritten Digits Using Numpy and MNIST dataset Using Neural Network

A simple fully connected feed forward neural network written in python from scratch using numpy & optimized using numba. It is possible to have multiple hidden layers, change amount of neurons per layer & have a different activation function per layer.

Python3 implementation of the Unsupervised Deep Learning Algorithm, Restricted Boltzmann Machine.

NumPy (short for Numerical Python) is a powerful Python library used for working with arrays, matrices, and numerical computations.

I made LeNet5 (one of the first convolution neural networks) in plain NumPy to classify digits from the MNIST dataset! Accuracy reaches 91.5% after one epoch :-)

Computer vision project that utilized openCV to detect a soccer ball and players in a livestream of a soccer game.

I made fully connected neural network in plain NumPy to classify digits from the MNIST dataset! It achieves 95% accuracy :-)

This project leverages the high-performance computational power of NumPy alongside the scalable machine learning framework TensorFlow to build a robust end-to-end data pipeline and predictive model.

A basic neural network from scratch with numpy

A proof of concept of a recursion doing stochastic gradient descent for a simple neural network. Done in Python3 with numpy

Implemented a two-layer neural network from scratch using NumPy to classify handwritten digits from the MNIST dataset. The project covers data preprocessing, one-hot encoding, forward propagation, ReLU and Softmax activation functions, backpropagation, and gradient descent for training.

Small NeuralNet-Framework implemented with NumPy (Convolution|TransposeConv|Linear)