VibeVoice How to use ncoder-ai/VibeVoice-Large-AWQ with VibeVoice:
import torch, soundfile as sf, librosa, numpy as np
from vibevoice.processor.vibevoice_processor import VibeVoiceProcessor
from vibevoice.modular.modeling_vibevoice_inference import VibeVoiceForConditionalGenerationInference
# Load voice sample (should be 24kHz mono)
voice, sr = sf.read("path/to/voice_sample.wav")
if voice.ndim > 1: voice = voice.mean(axis=1)
if sr != 24000: voice = librosa.resample(voice, sr, 24000)
processor = VibeVoiceProcessor.from_pretrained("ncoder-ai/VibeVoice-Large-AWQ")
model = VibeVoiceForConditionalGenerationInference.from_pretrained(
"ncoder-ai/VibeVoice-Large-AWQ", torch_dtype=torch.bfloat16
).to("cuda").eval()
model.set_ddpm_inference_steps(5)
inputs = processor(text=["Speaker 0: Hello!\nSpeaker 1: Hi there!"],
voice_samples=[[voice]], return_tensors="pt")
audio = model.generate(**inputs, cfg_scale=1.3,
tokenizer=processor.tokenizer).speech_outputs[0]
sf.write("output.wav", audio.cpu().numpy().squeeze(), 24000) 
