MLX Speech Models
Collection
Speech AI models for Apple Silicon via MLX. ASR, TTS, VAD, diarization, speaker embedding. • 96 items • Updated • 6
How to use aufklarer/Chatterbox-Multilingual-hi-MLX-fp16 with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir Chatterbox-Multilingual-hi-MLX-fp16 aufklarer/Chatterbox-Multilingual-hi-MLX-fp16
How to use aufklarer/Chatterbox-Multilingual-hi-MLX-fp16 with Chatterbox:
# pip install chatterbox-tts
import torchaudio as ta
from chatterbox.tts import ChatterboxTTS
model = ChatterboxTTS.from_pretrained(device="cuda")
text = "Ezreal and Jinx teamed up with Ahri, Yasuo, and Teemo to take down the enemy's Nexus in an epic late-game pentakill."
wav = model.generate(text)
ta.save("test-1.wav", wav, model.sr)
# If you want to synthesize with a different voice, specify the audio prompt
AUDIO_PROMPT_PATH="YOUR_FILE.wav"
wav = model.generate(text, audio_prompt_path=AUDIO_PROMPT_PATH)
ta.save("test-2.wav", wav, model.sr)Hindi-first MLX fp16 export of Resemble AI's Chatterbox multilingual Hindi checkpoint for speech-swift and Speech Studio.
model.safetensors: combined T3 Hindi checkpoint and S3Gen/vocoder weights.conformer.safetensors: voice encoder weights from the upstream Chatterbox
release.s3_tokenizer.safetensors: S3 tokenizer weights required by the Chatterbox
decoder path.tokenizer.json: Hindi grapheme tokenizer exported from
grapheme_mtl_merged_expanded_v1.json.config.json: bundle metadata, default language, and T3 tokenizer vocabulary
size.This bundle is intended for ChatterboxTTSModel.fromPretrained(...) in
speech-swift. Speech Studio loads it through the chatterbox-hi engine.
Converted from the MIT-licensed Resemble AI Chatterbox checkpoints:
ResembleAI/chatterboxResembleAI/Chatterbox-Multilingual-hiQuantized