Taarak9/DL-Projects

Deep learning projects

Linear Classifier

Aeroplane Classification using Linear Classifier with two variants

• Perceptron learning rule; mode = online learning, loss fn = MSE
• Gradient descent; mode = batch learning, loss fn = Cross Entropy

Hopfield Neural Network

Distorted Character Recognition using Hopfield Neural Network

Handwritten-Digit-Recognition

• mnist_loader.py - splits MNIST data
• src/FNN/hdr_fnn.py - Feedforward Neural Network that recognize handwritten digits using mini-batch stochastic gradient descent with an accuracy of 97.72 %.

To-do list

• Extend it to EMNIST data
• Try different models