Pocket Tts Pt Onnx
Flat mirror of the portuguese int8 ONNX bundle for Pocket TTS (Kyutai CALM zero-shot voice-cloning TTS), repackaged for the Sokuji Local Native provider's standard model-repo download path. Source bundle: the...
Model source
- Kind
- Other
- File size
- 9.50–72.8 MB
- Version
- vFlow lm flow int8
- Updated
- Oct 19, 2026
- Source
- Hugging Face
Source description
Pocket TTS ONNX — Portuguese
Flat mirror of the portuguese int8 ONNX bundle for Pocket TTS
(Kyutai CALM zero-shot voice-cloning TTS), repackaged for the
Sokuji Local Native provider's standard
model-repo download path. Source bundle: the
KevinAHM/pocket-tts-web Space
(onnx/portuguese/), whose export scripts and licensing apply.
Contents: five int8 ONNX graphs (mimi encoder/decoder, flow-LM main/flow, text
conditioner), tokenizer.model, bundle.json, bos_before_voice.npy, voices.bin
(eight predefined voices as flow-LM KV-cache prefixes), and voices/manifest.json
(the voice list Sokuji's UI reads).
Model weights: CC-BY-4.0 (attribution: Kyutai's Pocket TTS; ONNX export by KevinAHM). Do not use these files outside the terms of that license.
Sources
1 source
Sources
1 sourceVerified Oct 19
Download files
Model artifacts
6 artifacts
Model artifacts
6 artifactsSource excerpts
2 excerpts
Source excerpts
2 excerpts--- license: cc-by-4.0 language: - pt tags: - text-to-speech - voice-cloning - onnx - pocket-tts --- # Pocket TTS ONNX — Portuguese Flat mirror of the `portuguese` int8 ONNX bundle for [Pocket TTS](https://huggingface.co/kyutai/pocket-tts) (Kyutai CALM zero-shot voice-cloning TTS), repackaged for the [Sokuji](https://github.com/kizuna-ai-lab/sokuji) Local Native provider's standard model-repo download path. Source bundle: the [KevinAHM/pocket-tts-web](https://huggingface.co/spaces/KevinAHM/pocket-tts-web) Space (`onnx/portuguese/`), whose export scripts and licensing apply. Contents: five int8 ONNX graphs (mimi encoder/decoder, flow-LM main/flow, text conditioner), `tokenizer.model`, `bundle.json`, `bos_before_voice.npy`, `voices.bin` (eight predefined voices as flow-LM KV-cache prefixes), and `voices/manifest.json` (the voice list Sokuji's UI reads). Model weights: CC-BY-4.0 (attribution: Kyutai's Pocket TTS; ONNX export by KevinAHM). Do not use these files outside the terms of that license.
Source context: 0 downloads · 0 likes · Pipeline text-to-speech · Library pocket-tts · Repo jiangzhuo9357/pocket-tts-pt-onnx
