Add source separation recipe for audio domain

facebookresearch / recipes

Recipes are a standard, well supported set of blueprints for machine learning engineers to rapidly train models using the latest research techniques without significant engineering overhead.Specifically, recipes aims to provide- Consistent access to pre-trained SOTA models ready for production- Reference implementations for SOTA research reproducibility, and infrastructure to guarantee correctness, efficiency, and interoperability.

BSD 3-Clause "New" or "Revised" License

284 stars 35 forks source link

Add source separation recipe for audio domain #6

Closed nateanl closed 2 years ago

nateanl commented 2 years ago

The PR includes a training recipe for source separation that trains Conv-TasNet model on Libri2Mix dataset.

mthrok commented 2 years ago

This is most likely not about this PR, but it is not immediately clear where the entry point is. This makes it very difficult to understand the train logic from user perspective.

I think this might exhibit the same painpoints of existing training framework where source of training config value is not clear.

mthrok commented 2 years ago

How can I bring my own model if I want to replace the model with my own (with the same interface)?