mlx-lm-omni 0.1.1

Add your description here

Inference for Speech Models in MLX

This repo implements some speech models in MLX for better performance on Mac devices.

Currently, it supports the following models:

• Qwen2.5-Omni (both original version and mlx-quant version), currently supported only speech model
• Ultravox-0.5

Performance

Tested on MacBook M4-Pro (48GB RAM):

Model	Prompt TPS	Generation TPS
Qwen/Qwen2.5-Omni-7B	259.5	17.8
Qwen/Qwen2.5-Omni-3B	468.4	38.8
giangndm/qwen2.5-omni-7b-mlx-8bit	253.4	31.7
giangndm/qwen2.5-omni-7b-mlx-4bit	259.2	57.6
giangndm/qwen2.5-omni-3b-mlx-8bit	456.2	67.0
fixie-ai/ultravox-v0_5-llama-3_1-8b and mlx-community/Llama-3.1-8B-Instruct-4bit	188.5	40.4

How to use

uv add mlx-lm-omni 
# or
uv add https://github.com/giangndm/mlx-lm-omni.git

from mlx_lm_omni import load, generate
import librosa
from io import BytesIO
from urllib.request import urlopen

# model, tokenizer = load("giangndm/qwen2.5-omni-7b-mlx-4bit")
model, tokenizer = load("fixie-ai/ultravox-v0_5-llama-3_1-8b", model_config={"text_model_id": "mlx-community/Llama-3.1-8B-Instruct-4bit"})

audio_path = "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen2-Audio/audio/1272-128104-0000.flac"
audio = librosa.load(BytesIO(urlopen(audio_path).read()), sr=16000)[0]

messages = [
    {"role": "system", "content": "You are a speech recognition model."},
    {"role": "user", "content": "Transcribe the English audio into text without any punctuation marks.", "audio": audio},
]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)