This model requires you to accept the CC BY-NC-SA 4.0 license terms before download.
Description
OpenAudio S1-Mini — Fish Audio compact multilingual (13-language) text-to-speech model offering expressive, emotion-controllable voice synthesis and zero-shot voice cloning. Non-commercial license (CC-BY-NC-SA-4.0).
Tags
At a glance
| Kind | text-to-speech model |
|---|---|
| Precision | bfloat16 |
| Container size | 2.6 GiB |
| Weights inside | 3.3 GiB |
| Components | 5 |
| Licence | CC BY-NC-SA 4.0 |
| Origin | fishaudio/openaudio-s1-mini @ f4b445029346 |
What is inside the container
A .nbx container packs every part of s1-mini — 5 components with weights, graphs and runtime settings — so one file is all the runtime needs.
| Component | Blocks | Layers | Weights |
|---|---|---|---|
| codec.encoder | 0 | — | 551 MiB |
| codec.quantizer | 0 | — | 1.0 GiB |
| codec.decoder | 0 | — | 206 MiB |
| model | 0 | 28 | 1.5 GiB |
| model.fast | 0 | — | 112 MiB |
Default settings
The values the container ships with. They apply when you do not pass your own.
- Temperature
- 0.7
- Top-p
- 0.8
Memory needed
The weights of s1-mini take 3.3 GiB in bfloat16. That is the floor for the memory the runtime has to place, spread over the devices it finds; activations come on top and grow with the size and length you ask for. A measured peak per device class is not published for this model yet.
Verify what you downloaded
Every download carries its SHA-256, and the file endpoint sends it in the X-NBX-SHA256 header. The container is 2.6 GiB (2,811,042,520 bytes).
$ sha256sum s1-mini.nbx
ea8f46deb5cdb49df365e59906824c51f661b16ea23c562cb0ed882ec847ed8e
Questions
What is s1-mini?
A text-to-speech model by fishaudio, distributed as a bfloat16 .nbx container and licensed CC BY-NC-SA 4.0.
How big is the download?
2.6 GiB, of which 3.3 GiB are weights.
How do I run it?
Install the runtime from the download page, then use the commands under “CLI Usage” below: neurobrix import fishaudio/s1-mini, then neurobrix serve --model s1-mini.
CLI Usage
$ neurobrix import fishaudio/s1-mini
$ neurobrix serve --model s1-mini
$ neurobrix run --prompt "your prompt here"
$ neurobrix stop