Multispeaker support and some changes

[ShiromiyaG]

Description

This PR adds support for training multispeaker models and some changes

Motivation and Context

How has this been tested?

Screenshots (if appropriate):

Types of changes

• [ ] Bug fix (non-breaking change which fixes an issue)
• [X] New feature (non-breaking change which adds functionality)
• [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)

Checklist:

• [X] My code follows the code style of this project.
• [ ] My change requires a change to the documentation.
• [ ] I have updated the documentation accordingly.
• [ ] I have added tests to cover my changes.
• [X] All new and existing tests passed.

[blaisewf]

inference works well?

[ShiromiyaG]

@blaisewf Yes

[blaisewf]

ok, core.py is missing things

[ShiromiyaG]

Ow, okay, I'll add

[ShiromiyaG]

@blaisewf I think I've added everything that was missing

[blaisewf]

extract or preprocess doesn't need something on the core?

[ShiromiyaG]

extract or preprocess doesn't need something on the core?

No

[blaisewf]

extract or preprocess doesn't need something on the core?

No

so how do you preprocess a multispeaker dataset

[AznamirWoW]

extract or preprocess doesn't need something on the core?

No

so how do you preprocess a multispeaker dataset

user provides a location of the training set, such as C:\training all acceptable files in C:\training are used for sid 0 as compatibility all child folders with numeric values (we actually need to check that they are int) are used as speaker IDs and content of those folders is processed normally. the file index is written with these SID values + 2 mute files per speaker.

[blaisewf]

well so this needs to be well explained on the docs