Instructions to use openensemble/pocket-tts with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Pocket-TTS
How to use openensemble/pocket-tts with Pocket-TTS:
from pocket_tts import TTSModel import scipy.io.wavfile tts_model = TTSModel.load_model("openensemble/pocket-tts") voice_state = tts_model.get_state_for_audio_prompt( "hf://kyutai/tts-voices/alba-mackenna/casual.wav" ) audio = tts_model.generate_audio(voice_state, "Hello world, this is a test.") # Audio is a 1D torch tensor containing PCM data. scipy.io.wavfile.write("output.wav", tts_model.sample_rate, audio.numpy()) - Notebooks
- Google Colab
- Kaggle
license: cc-by-4.0
language: en
pipeline_tag: text-to-speech
tags:
- tts
- voice-cloning
- pocket-tts
- mirror
pocket-tts — OpenEnsemble mirror
Non-gated mirror of the English voice-cloning weights from kyutai/pocket-tts, redistributed by OpenEnsemble under the model's CC-BY-4.0 license so OpenEnsemble users can install Pocket TTS without the upstream access gate.
Original model © Kyutai Labs — https://github.com/kyutai-labs/pocket-tts
This mirror contains only languages/english/model.safetensors (the voice-cloning
weights). The tokenizer and non-cloning weights are pulled at install time from the
upstream non-gated kyutai/pocket-tts-without-voice-cloning repo.
License & acceptable use (CC-BY-4.0)
Licensed CC-BY-4.0. You must attribute Kyutai Labs and comply with the upstream acceptable-use terms — in particular: no voice cloning without the explicit, lawful consent of the person whose voice is cloned, and no impersonation, misinformation, or other unlawful/harmful use.