Giter VIP home page Giter VIP logo

deep-learning---automatic-music-transcription's Introduction

Deep-learning-Automatic-Music-Transcription

This repository contains the research done and the codes written for the semester project - Automatic Music Transcription, offered by Electronics Club IIT Kanpur.

  1. We have explored various models, for example, CNN, RNN and models with the combination of two, with a sole motive of achieveing maximum accuracy.
  2. For preprocessing of the music files, we have tried various ways, for example, constant Q- transform, Short- time fourier transform, and mel-frequency spectrum, and have compared the results with respect to the model used.
  3. Each folder in this repository represents a different model trained for automatic music transcription.
  4. We will keep on updating the repository until we arrive at the end dates of the projects.

Till then, stay tuned.

deep-learning---automatic-music-transcription's People

Contributors

agarwalapurb avatar ananya0102 avatar ankurkraj avatar athulkj avatar harsharya3107 avatar himanshujindal20 avatar janhvi-rochwani avatar naiza2000 avatar rrustagi20 avatar shivangj20 avatar shreyabhatta avatar shrutinisal avatar vineet-the-git avatar

Stargazers

avatar avatar avatar avatar

Watchers

avatar avatar

Forkers

shreyabhatta janhvi-rochwani ankurkraj ananya0102 vineet-the-git nitya8747 shrutinisal athulkj shivangj20 himanshujindal20 rrustagi20 harsharya3107 agarwalapurb xxyyh

deep-learning---automatic-music-transcription's Issues

RNN : Model 2

  • If any changes in preprocessing have been made in comparison to the previous model, add the code
  • Update README.md telling how it is better than the previous model
  • Upload all the necessary codes. Unlike the previous model, there is no need to form sub-folders
  • Show results for the same test file used in the previous model.
  • Calculate error

Preprocess MAPS dataset for Model-1 (CNN)

Create three notebooks using STFT, Mel transforms and Constant Q-transform separately for processing the audio files. Write comments properly. Should be done by a team of 2 people. A comparison of the results will be performed in the end.
After this, update the README.md file with the parameters that you used and the reasons for using the same. Proper documentation should be done side by side. Consult the team involved in making the model for the shape and representation that they want for their model.
Should be completed by 12:00 PM, 14th July 2021.

Implement basic Model-1 (CNN)

  • Submit .ipynb file
  • Write comments properly
  • Train on the entire dataset
  • Test the model using a test file
  • Update the readme.MD file

Preprocess MAPS dataset for model-1(RNN)

Create three notebooks using STFT, Mel transforms and Constant Q-transform separately for processing the audio files. Write comments properly. Should be done by a team of 2 people. A comparison of the results will be performed in the end.
After this, update the README.md file with the parameters that you used and the reasons for using the same. Proper documentation should be done side by side. Consult the team involved in making the model for the shape and representation that they want for their model.
Should be completed by 12:00 PM, 14th July 2021

Implement basic model-1 (RNN)

  • Submit .ipynb file
  • Write comments properly
  • Train on the entire dataset
  • Test the model using a test file
  • Update the readme.MD file
    Deadline : 14th July, 2021 EOD

CNN : Model 1

Preprocessing folder:

  • upload codes for all three methods used
  • Update the README.md file
  • Mention time taken for each preprocessing method
    Model:
  • Upload saved models in respective directories
  • Upload codes
  • Update README.md
  • Time taken for training
  • Write code for calculating error by predicting output for 60 test files.
  • Use one of those files for comparison of models
    Post-processing :
  • Upload codes
  • Do a comparison of the three preprocessed datasets
  • Upload generated MIDI files from the model.
  • Update MIDI file

Also, keep your forked repos in sync with the main branch before sending the PR.
Whenever you send the PR, specify the issue number that you are addressing by mentioning the "#"

RNN : Model 1

Preprocessing folder:

  • upload codes for all three methods used
  • Update the README.md file
  • Mention time taken for each preprocessing method
    Model:
  • Upload saved models in respective directories
  • Upload codes
  • Update README.md
  • Time taken for training
  • Write code for calculating error by predicting output for 60 test files.
  • Use one of those files for comparison of models
    Post-processing :
  • Upload codes
  • Do a comparison of the three preprocessed datasets
  • Upload generated MIDI files from the model.
  • Update MIDI file

Also, keep your forked repos in sync with the main branch before sending the PR.
Whenever you send the PR, specify the issue number that you are addressing by mentioning the "#"

Generate Piano Roll Model-1(CNN)

  • Discuss with preprocessing team and backtrack the steps.
  • Convert model output to Piano Roll
  • Convert piano roll to MIDI
  • Update README.md
    Deadline : EOD, 14th July 2021

Generate piano roll Model-1 (RNN)

  • Discuss with preprocessing team and backtrack the steps.
  • Convert model output to Piano Roll
  • Convert piano roll to MIDI
  • Update README.md
    Deadline : EOD, 14th July 2021

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.