{"id":20847160,"url":"https://github.com/vanajmoorthy/cs4099-polytab","last_synced_at":"2025-03-12T11:43:08.918Z","repository":{"id":233129699,"uuid":"762271305","full_name":"vanajmoorthy/CS4099-PolyTab","owner":"vanajmoorthy","description":"A convolutional neural network to transcribe solo acoustic guitar recordings to tabs 🎸 ","archived":false,"fork":false,"pushed_at":"2024-04-14T22:17:41.000Z","size":5771,"stargazers_count":1,"open_issues_count":1,"forks_count":0,"subscribers_count":3,"default_branch":"main","last_synced_at":"2024-05-28T20:47:08.548Z","etag":null,"topics":["cnn","guitar-tabs","machine-learning"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/vanajmoorthy.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-02-23T12:39:03.000Z","updated_at":"2024-06-24T16:29:51.925Z","dependencies_parsed_at":null,"dependency_job_id":"35156774-a21b-4dc7-a159-781ac6765295","html_url":"https://github.com/vanajmoorthy/CS4099-PolyTab","commit_stats":null,"previous_names":["vanajmoorthy/cs4099-polytab"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/vanajmoorthy%2FCS4099-PolyTab","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/vanajmoorthy%2FCS4099-PolyTab/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/vanajmoorthy%2FCS4099-PolyTab/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/vanajmoorthy%2FCS4099-PolyTab/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/vanajmoorthy","download_url":"https://codeload.github.com/vanajmoorthy/CS4099-PolyTab/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":243213867,"owners_count":20254879,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["cnn","guitar-tabs","machine-learning"],"created_at":"2024-11-18T02:19:08.241Z","updated_at":"2025-03-12T11:43:08.902Z","avatar_url":"https://github.com/vanajmoorthy.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# PolyTab\n\nA machine-learning model to automatically transcribe audio recordings of solo acoustic guitars to readable guitar tablature built using TensorFlow. Undertaken for my Senior-Honours dissertation at the University of St Andrews, this model builds on academic work in the nascent field of using machine learning for Music Information Retrieval (MIR), particularly the [TabCNN](https://archives.ismir.net/ismir2019/paper/000033.pdf) model and utilises a convolutional neural network to learn features from spectograms of guitar audio recordings to automatically produce usable guitar tablature\n\nSupervised by [Dr Oggie Arandelovic](https://www.st-andrews.ac.uk/computer-science/people/oa7/) who awarded me a grade of a first and called this project \"very good work with some genuine contributions to the state of the art\".\n\nTo run the project and train the model yourself a few step must be taken.\n* The dataset can be downloaded [here](https://zenodo.org/record/1422265/files/GuitarSet_audio_and_annotation.zip?download=1). Make sure to unzip this folder and place it in the root directory of the project.\n* After that, you must activate a virtual environment to install the dependencies. This can be done with `python3 -m venv \u003cenv\u003e`\n* Then you can activate the virtual environment by running `source \u003cenv\u003e/bin/activate`\n* Next, please install the dependencies with `pip install -r requirements.txt`\n* After this, you have to generate the CQT representations for the audio files using `python3 ParallelGenerateCQTs.py`\n* And finally you can train the model with `python3 PolyTab.py`\n* Once the model has trained, you can run `python3 PolyTabPredictor.py --weights \"path/to/weights.h5\" --audio \"path/to/audio/file.wav\"`\nwith the trained weights and the audio you want to predict for. The saved predictions can be found in the /predictions folder.\n* You can then run `python3 PolyTabPredictor.py --weights \"saved/c 2024-03-21 171741/5/weights.h5\" --audio \"path/to/audio/file.wav\"` to predict using the model which was trained with the learnable weighted loss and AdamW optimiser.\n\nThe accompanying paper for this project can be found [here](https://github.com/vanajmoorthy/CS4099-PolyTab/blob/main/Automatic%20Polyphonic%20Guitar%20Transcription.pdf)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fvanajmoorthy%2Fcs4099-polytab","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fvanajmoorthy%2Fcs4099-polytab","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fvanajmoorthy%2Fcs4099-polytab/lists"}