This repository contains the unofficial implementation of the DeepSRGM paper by Sathwik Tejaswi Madhusudhan and Girish Chowdhary.
This was done as a part of the course project of IE 643 - Deep Learning: Theory and Practice, under Prof. P. Balamurugan.
The details about the project, previous work, our contributions, and our results can be found in the report. The detailed summaries and trained weights of our models can be found here.
Repository Structure:
automatic-raga-recognition/
├── README.md
├── data
│ └── README.md
├── models
│ ├── gru_30_checkpoint.pth
│ └── lstm_25_checkpoint.pth
├── report.pdf
└── src
├── __init__.py
├── attention_layer.py
├── dataset_preprocessing.ipynb
├── deepSRGM.py
├── evaluate.ipynb
├── original-project-implementation
│ ├── dataset_preprocessing-40.ipynb
│ ├── deepSRGM-40rag.ipynb
│ └── deepSRGM.ipynb
├── test_utils.py
├── train.ipynb
└── train_utils.py
Note: The original-project-implementation directory contains the code for the original implementation during the IE643 course-project. This code has since been cleaned up and refactored. It has been included only to provide credibility to the claims in report.pdf.
Created with ❤️ by Shubham Lohiya & Swarada Bharadwaj