How can I migrate my single-speaker Piper (version 1.3.0) model to a multi-speaker model using my own data?

[King-colab]

Hello,

I previously trained a Piper single-speaker model of my own voice (about 2000 epochs) using the original piper1 setup. Now I want to create a multi-speaker model, where each "speaker" is myself speaking different languages.

Here’s what I tried so far:

I created a CSV with a speaker ID column and edited the JSON speaker mapping.
I tried training with the ckpt_path set to my single-speaker checkpoint, hoping Piper would pick up the extra speaker(s) and transition to multi-speaker, but it only produces a single-speaker model.
I also saw that the resume_from_single_speaker_checkpoint parameter no longer exists.

My questions are:

Is there a recommended, supported way to migrate a single-speaker model (with training already invested) to a multi-speaker model in Piper1?
If direct migration isn’t possible, is there a workflow or script to transfer or reuse my learned weights or voice quality for a new multi-speaker training run, or do I need to start all over from scratch?
What is the best practice for this scenario? Should I start from a blank multi-speaker checkpoint and just add all my old data as a new “speaker”?

Any guidance, documentation links, or scripts that could help me avoid wasting my original training effort would be greatly appreciated!

Thank you!