Neural-Nets-Pytorch

This repo consists of deep learning models using PyTorch.

Installations for the repo

Command to install pytorch in your environment

1. For GPU:

conda install pytorch torchvision cudatoolkit=10.0 -c pytorch

2. For CPU:

conda install pytorch torchvision cpuonly -c pytorch

For all the packages used for the project, kindly install them using the command

pip install requirements.txt

Implementation

The implementation is divide into 2 stages

1. Stage 1:

This stage consists of code for learnig pytorch, implementing perceptron using numpy and our first pytorch model for KMNIST dataset.

2. Stage 2:

Stage 2 consists of code relating to RNNs where implementation of LSTM, BiGRU is done on a sample of the whole IMDB dataset.

All the datasets should be stored in the data directory for the codes to run successfully.

Acknowledgement

I would like to thank Alan Blair and his team for supporting us with the materials and the templates required to learn pytorch in this repository as part of the course.