ttspro-voces

Llicklair/ttspro-voces

ttspro-voces is a text-to-speech model from Llicklair. Use it when you need text read aloud. It is set up for piper. The card lists the license as mit.

Downloads · 30 days

26% of all-time downloads

All-time downloads

257

Public

Repo size

713 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.onnx713 MB · 100%

At a glance

Task: Text-to-Speech
Library: piper
License: mit
Access: Public
Created: Sep 5, 2026
Updated: Sep 5, 2026
SHA: f72457b4

Model card

From the Hugging Face model README

ttspro · voice packs

Base voices for ttspro, a Spanish text-to-speech with voice cloning that runs entirely in the browser with ONNX Runtime Web. Each pack is a Piper voice (MIT) ported onto ttspro's modules and exported within ONNX Runtime Web's WebGPU operator set, with its own contract (symbol table, graph signature) and its base-voice vector for the tone-colour converter.

Files: indice.json lists the packs; <clave>.json is meta + contract + vector; <clave>.tts.onnx (fp32) and <clave>.tts.fp16.onnx are the graphs.

Use: open the ttspro page with ?voces=https://huggingface.co/Llicklair/ttspro-voces/resolve/main/, or TTS.cargar({ vocesBase: "https://huggingface.co/Llicklair/ttspro-voces/resolve/main/" }) in the library. Built by uv run python -m ttspro.export.paquete --todas es.

Pack	fp32 / fp16	ms per sentence (CPU)
es_AR-daniela-high	114.0 / 57.6 MB	210
es_ES-carlfm-x_low	20.7 / 11.0 MB	33
es_ES-davefx-medium	63.3 / 32.3 MB	43
es_ES-mls_10246-low	63.2 / 32.2 MB	129
es_ES-mls_9972-low	63.2 / 32.2 MB	62
es_MX-ald-medium	63.3 / 32.3 MB	63
es_MX-ald-x_low	20.8 / 11.0 MB	66
es_MX-claude-high	63.3 / 32.3 MB	60

Task: Text-to-Speech
Library: piper
License: mit
Languages: es
Created: Sep 5, 2026
Updated: Sep 5, 2026