Configuration Parsing Warning:Config file tokenizer_config.json cannot be fetched (too big)
supertonic-neutts
Pinned, public mirror of the model files the Supertonic Reader iOS app downloads for its
on-device NeuTTS Nano neural speech engine. Redistribution is permitted by the
NeuTTS Open License v1.0 (see NEUTTS_LICENSE.txt); these files are unmodified copies.
| file | source | purpose |
|---|---|---|
neutts-nano-q8_0.gguf |
BricksDisplay/NeuTTS-Nano-GGUF |
Q8_0 backbone (phoneme text + reference codes -> speech token IDs), run with llama.cpp |
neucodec_decoder.int8.onnx |
neuphonic/neucodec-onnx-decoder-int8 |
int8 NeuCodec decoder (speech token IDs -> 24 kHz PCM), run with ONNX Runtime |
reference_voices.json |
neuphonic/neutts samples (dave/emily/greta/jo) |
bundled speaker references: transcript + pre-encoded NeuCodec codes |
tokenizer.json, config.json |
neuphonic/neutts-nano |
reference only; llama.cpp reads the vocab from the GGUF |
Prompt format (phonemes), speech-token id range <|speech_0|> = 128262 .. <|speech_65535|> = 193797,
stop <|SPEECH_GENERATION_END|> = 128261. See the app's docs/NEUTTS_ENGINE.md.
- Downloads last month
- 54
Hardware compatibility
Log In to add your hardware
8-bit
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for grimshaw/supertonic-neutts
Base model
neuphonic/neucodec