NeuTTS Nano GGUF

GGUF conversions of Neuphonic's NeuTTS Nano — a compact (~228 M params, ~117 M active) on-device TTS LM with instant voice cloning.

The stack is split so each runtime owns the part it's good at:

• LLM-part — text-and-reference-codes to speech token IDs, stock llama arch, runs in llama.cpp.
• Codec-part — speech token IDs to/from 24 kHz PCM via NeuCodec, runs in codec.cpp.

Files

LLM-part (llama arch, vocab 194256, RoPE θ=500000 + linear scaling x32)

neutts-nano-<quant>.gguf

Codec-part (NeuCodec, neucodec arch, 65536 codebook, 24 kHz, 80 Hz token rate)

codec[-<quant>].gguf

Prompt format

The reference flow (from neuphonic/neutts):

user: Convert the text to speech:<|TEXT_PROMPT_START|>{ref_phones} {input_phones}<|TEXT_PROMPT_END|>
assistant:<|SPEECH_GENERATION_START|>{ref_codes_as_speech_tokens}

• {ref_phones} / {input_phones} — IPA-phonemized text (espeak-ng / phonemizer).
• {ref_codes_as_speech_tokens} — concatenation of <|speech_N|> strings for each NeuCodec token of a reference clip.
• Generation continues with more <|speech_N|> tokens until <|SPEECH_GENERATION_END|>.

Speech-token vocab range: <|speech_0|> = id 128262 … <|speech_65535|> = id 193797. Stop token: <|SPEECH_GENERATION_END|> = id 128261.

Notes

• Reference encoding (audio → speech tokens) needs the codec encode path.
• Phonemization isn't part of the GGUF — apply espeak-ng (or equivalent) on the host before tokenizing prompts; raw English text will be off-distribution.
• Source weights: neuphonic/neutts-nano (LLM) and neuphonic/neucodec (codec).