Pravega-TTS
Copyright 2026 Dheeyantra Research Labs

This repository redistributes components of CosyVoice 3, which are licensed
under the Apache License, Version 2.0:

  FunAudioLLM/Fun-CosyVoice3-0.5B-2512
  https://huggingface.co/FunAudioLLM/Fun-CosyVoice3-0.5B-2512
  Copyright FunAudioLLM / Alibaba Group

  Redistributed unchanged from that release:
    speech_tokenizer_v3.onnx     supervised semantic speech tokenizer
    campplus.onnx                CAM++ speaker embedding (3D-Speaker)
    hift.pt                      HiFTNet vocoder (not fine-tuned)
    CosyVoice-BlankEN/           text encoder and tokenizer
    cosyvoice3.yaml              model configuration

  You may obtain a copy of the Apache License at
  http://www.apache.org/licenses/LICENSE-2.0

Modified by Dheeyantra Research Labs and licensed under the Dhee Research-Only
Licence v1.0 (see LICENSE):

    llm.pt                       fine-tuned autoregressive text-to-token LM
    flow.pt                      fine-tuned flow-matching token-to-mel model

Reference voice clips in voice_bank/:
    or_*.wav, pa_*.wav           derived from Indic Parler-TTS (Apache-2.0)
    as_*.wav                     synthesised by Dheeyantra with IndicF5 (MIT)

Training data lineage, and why this release is non-commercial: the Indic
fine-tuning corpus draws on the SPRINGLab IndicTTS corpora, which derive from
the IIT-Madras Indic TTS database, whose terms are non-commercial. It also uses
ai4bharat/indicvoices_r (CC-BY-4.0), google/fleurs (CC-BY-4.0) and
facebook/voxpopuli (CC0-1.0).