Audio Classification
Transformers
ONNX
Safetensors
English
dualturn_endpointing
feature-extraction
turn-taking
endpointing
end-of-turn
voice-activity-detection
voice-agents
conversation
speech
audio
mimi
dualturn
real-time
custom_code
Instructions to use anyreach-ai/dualturn-endpointing with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use anyreach-ai/dualturn-endpointing with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("audio-classification", model="anyreach-ai/dualturn-endpointing", trust_remote_code=True)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("anyreach-ai/dualturn-endpointing", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Upload config.json with huggingface_hub
Browse files- config.json +24 -0
config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "dualturn",
|
| 3 |
+
"backbone_input_mode": "continuous",
|
| 4 |
+
"mimi_sample_rate": 24000,
|
| 5 |
+
"mimi_frame_rate": 12.5,
|
| 6 |
+
"mimi_frame_ms": 80.0,
|
| 7 |
+
"signal_keys": [
|
| 8 |
+
"vad_user", "vad_agent",
|
| 9 |
+
"eot_user", "eot_agent",
|
| 10 |
+
"bot_user", "bot_agent",
|
| 11 |
+
"fvad_user_short", "fvad_user_long",
|
| 12 |
+
"fvad_agent_short", "fvad_agent_long"
|
| 13 |
+
],
|
| 14 |
+
"st_threshold": 0.30,
|
| 15 |
+
"vad_edge_threshold": 0.50,
|
| 16 |
+
"agent_voice_min": 0.15,
|
| 17 |
+
"fvad_alpha_short": 0.3,
|
| 18 |
+
"fvad_alpha_long": 0.7,
|
| 19 |
+
"auto_map": {
|
| 20 |
+
"AutoConfig": "configuration_dualturn.DualTurnConfig",
|
| 21 |
+
"AutoModel": "modeling_dualturn.DualTurnModel"
|
| 22 |
+
},
|
| 23 |
+
"transformers_version": "4.0.0"
|
| 24 |
+
}
|