Thank you! :)
@avans06 Hey bro!
Just wanted to say thank you for enhancing my project and for giving me credit! :) Really appreciate it :)
I wanted to suggest adding BS-Former/Mel-Former for nice audio separation prior to transcription. Here is the link I use:
https://huggingface.co/spaces/JacobLinCool/vocal-separation
I also wanted to ask you to take a look at MT3 for multi-instrumental stuff:
https://huggingface.co/spaces/Hmjz100/MT3
https://huggingface.co/spaces/mimbres/YourMT3
And also make sure to update my TMIDIX and midi_to_colab_audio to latest version. I made some improvements/enhancements to the code.
Anyway, let me know.
Sincerely,
Alex
@asigalov61 Hello, thank you for your suggestions and message.
At first glance, BS-Former looks like an excellent vocal separation model, and I’ll give it a try when I have time.
I’ll also update TMIDIX and midi_to_colab_audio to the latest version when I get a chance. Thank you!
@avans06 Hey, bro!
Just checking in with you again! :)
Great job so far on all the improvements to the space! Looks very nice and useful! :)
I wanted to direct your attention to the MuScriptor, which is a SOTA Audio to MIDI model for multi-instrumental music transcription. It can also be used for music stems separation.
It would be cool if you would add it too. Either directly or via bootstrapping to HF space.
Here are the links:
https://muscriptor.kyutai.org/
https://huggingface.co/spaces/hugging-apps/muscriptor-music-transcription
Also, pull the latest TMIDIX from my tegridy-tools since I've done some minor updates to some functions.
https://github.com/asigalov61/tegridy-tools/blob/main/tegridy-tools/TMIDIX.py
Sincerely,
Alex.