Audio-Text-to-Text
Transformers
Safetensors
vibevoice_asr
automatic-speech-recognition
ASR
Diarization
Speech-to-Text
Transcription
Instructions to use bezzam/VibeVoice-ASR-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use bezzam/VibeVoice-ASR-7B with Transformers:
# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("bezzam/VibeVoice-ASR-7B") model = AutoModelForMultimodalLM.from_pretrained("bezzam/VibeVoice-ASR-7B", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Upload processor
Browse files- processor_config.json +4 -0
processor_config.json
CHANGED
|
@@ -1,4 +1,8 @@
|
|
| 1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
"feature_extractor": {
|
| 3 |
"eps": 1e-06,
|
| 4 |
"feature_extractor_type": "VibeVoiceAcousticTokenizerFeatureExtractor",
|
|
|
|
| 1 |
{
|
| 2 |
+
"audio_bos_token": "<|object_ref_start|>",
|
| 3 |
+
"audio_duration_token": "<|AUDIO_DURATION|>",
|
| 4 |
+
"audio_eos_token": "<|object_ref_end|>",
|
| 5 |
+
"audio_token": "<|box_start|>",
|
| 6 |
"feature_extractor": {
|
| 7 |
"eps": 1e-06,
|
| 8 |
"feature_extractor_type": "VibeVoiceAcousticTokenizerFeatureExtractor",
|