Automatic Speech Recognition
Transformers
Safetensors
VibeVoice
multilingual
vibevoice_asr
bitsandbytes
4-bit precision
quantized
diarization
Instructions to use Dubedo/VibeVoice-ASR-HF-NF4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Dubedo/VibeVoice-ASR-HF-NF4 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="Dubedo/VibeVoice-ASR-HF-NF4")# Load model directly from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq processor = AutoProcessor.from_pretrained("Dubedo/VibeVoice-ASR-HF-NF4") model = AutoModelForSpeechSeq2Seq.from_pretrained("Dubedo/VibeVoice-ASR-HF-NF4") - VibeVoice
How to use Dubedo/VibeVoice-ASR-HF-NF4 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("Dubedo/VibeVoice-ASR-HF-NF4") model = VibeVoiceForConditionalGenerationInference.from_pretrained( "Dubedo/VibeVoice-ASR-HF-NF4", 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) - Notebooks
- Google Colab
- Kaggle
Upload quantization_metadata.json with huggingface_hub
Browse files- quantization_metadata.json +18 -0
quantization_metadata.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"quantization_method": "bitsandbytes",
|
| 3 |
+
"bits": 4,
|
| 4 |
+
"quant_type": "nf4",
|
| 5 |
+
"double_quant": true,
|
| 6 |
+
"modules_not_quantized": [
|
| 7 |
+
"acoustic_tokenizer_encoder",
|
| 8 |
+
"semantic_tokenizer_encoder",
|
| 9 |
+
"acoustic_projection",
|
| 10 |
+
"semantic_projection",
|
| 11 |
+
"lm_head"
|
| 12 |
+
],
|
| 13 |
+
"modules_quantized": [
|
| 14 |
+
"language_model (excluding lm_head)"
|
| 15 |
+
],
|
| 16 |
+
"base_model": "microsoft/VibeVoice-ASR-HF",
|
| 17 |
+
"transformers_version": "5.3.0"
|
| 18 |
+
}
|