mlx-community/Irodori-TTS-v4.1-Small-8bit

This model was converted to MLX format from Aratako/Irodori-TTS-v4.1-Small using mlx-audio version 0.4.7.

Use with mlx-audio

pip install -U mlx-audio

Command line

mlx_audio.tts.generate --model mlx-community/Irodori-TTS-v4.1-Small-8bit --text "こんにちは、Irodori TTSのMLX版です。" --ref_audio reference.wav

Python

from mlx_audio.tts.generate import generate_audio

generate_audio(
    text="こんにちは、Irodori TTSのMLX版です。",
    model="mlx-community/Irodori-TTS-v4.1-Small-8bit",
    ref_audio="reference.wav",
    file_prefix="output",
)

Notes

Unified v4.1 model: voice cloning, VoiceDesign (caption) and automatic duration prediction in one checkpoint.
v4.1 differs from v4 only in the duration predictor, which upstream retrained separately with every other parameter frozen. The remaining 683 of 714 tensors are bit-identical to v4.
The ModernBERT-ja-310m text encoder weights and its tokenizer are bundled, and the Semantic-DACVAE-Japanese-32dim codec ships in dacvae/, so inference needs no additional downloads.
Reference audio: up to 120s. Passing a list of clips encodes each separately and concatenates them, which matches training better than one long recording.
Short caption-only prompts are shorter than in v4, but still over-predicted. Measured with the same reference clip and caption:

Text Tokens v4 caption only v4.1 caption only

こんにちは。 3 3.64s 2.88s

今日はいい天気ですね。 5 5.60s 4.72s

MLXへの移植が完了しました。 7 4.08s 3.84s

Under about seven tokens the model can still fill the surplus by reading the sentence a second time. This matches the reference PyTorch implementation, which predicts 117.41 frames against MLX's 117.56 for the second row, and produces the same repeat. Pass reference audio, or set duration_scale / seconds.
Quantized to 8 bits with mlx.nn.quantize (group size 64). Note that this covers more layers than the upstream torchao checkpoints, which keep AdaLN, the projectors and the duration predictor in BF16. Duration predictions stay within one frame (40 ms) of the fp16 model.