Incorporating MultiTrack Data

[Red-Giuliano]

Hi,

First off let me say thanks so much for this amazing package.

I'm interested in running the pytorch implementation using the data from the original Tensorflow MuseGAN project that links to this project (https://github.com/salu133445/musegan). The data I am trying to use can be found at:

https://docs.google.com/uc?export=download&id=14rrC5bSQkB9VYWrvt2IhsCjOKYrguk3S

The issue I am running into is that the input data is of a different shape than the model currently supports. How should I go about changing the code to accept this input data?

Thanks so much!

[akanametov]

Hi,

First off let me say thanks so much for this amazing package.

I'm interested in running the pytorch implementation using the data from the original Tensorflow MuseGAN project that links to this project (https://github.com/salu133445/musegan). The data I am trying to use can be found at:

https://docs.google.com/uc?export=download&id=14rrC5bSQkB9VYWrvt2IhsCjOKYrguk3S

The issue I am running into is that the input data is of a different shape than the model currently supports. How should I go about changing the code to accept this input data?

Thanks so much!

Thanks for interest!

Unfortunately I don`t have enough time currently to train it on that dataset. But as I understood, you should preprocess that dataset according to the script: https://github.com/salu133445/musegan/blob/main/src/process_data.py
Today that script was added to this project with modifications: https://github.com/akanametov/musegan/blob/dev/data/lpd/prepare_data.py
To prepare data, put .npz file in directory: data/lpd/ and run the scipt. Prepared dataset requires around 7Gb of memory.

Model hyperparameters and dataset loader can be found in demo notebook: https://github.com/akanametov/musegan/blob/dev/demo/demo_lpd.ipynb
That's all I can do for now, hope you understand))

Good luck!

[Red-Giuliano]

Hi, Thanks so much for your timely response! I was able to get the code to run but I believe there are some further modifications that need to be made to the postprocessing function and the trainer. Also I believe based on this issue salu133445/musegan#63 that the .npz file they have saved is in the following format (n_samples, n_bars, n_timesteps, n_pitchs, n_tracks) so I'm not sure if the current loading function would work properly... I can send over my working code as soon as I have the time. Thanks again! Best!

…

On Sat, Aug 27, 2022 at 6:17 AM Azamat Kanametov ***@***.***> wrote: Hi, First off let me say thanks so much for this amazing package. I'm interested in running the pytorch implementation using the data from the original Tensorflow MuseGAN project that links to this project ( https://github.com/salu133445/musegan). The data I am trying to use can be found at: https://docs.google.com/uc?export=download&id=14rrC5bSQkB9VYWrvt2IhsCjOKYrguk3S The issue I am running into is that the input data is of a different shape than the model currently supports. How should I go about changing the code to accept this input data? Thanks so much! Thanks for interest! Unfortunately I don`t have enough time currently to train it on that dataset. But as I understood, you should preprocess that dataset according to the script: https://github.com/salu133445/musegan/blob/main/src/process_data.py Today that script was added to this project with modifications: https://github.com/akanametov/musegan/blob/dev/data/lpd/prepare_data.py To prepare data, put .npz file in directory: *data/lpd/* and run the scipt. Prepared dataset requires around 7Gb of memory. Model hyperparameters and dataset loader can be found in demo notebook: https://github.com/akanametov/musegan/blob/dev/demo/demo_lpd.ipynb That all I can do for now, hope you understand)) Good luck! — Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKTYWPT7A4OQW7MQBJKPCXDV3HTJZANCNFSM57XTT75Q> . You are receiving this because you authored the thread.Message ID: ***@***.***>