Namo Turn Detector v1

---

Namo-v1 is a collection of open-source turn detection models that solve one of the most challenging problems in conversational AI: knowing when a user has finished speaking. Namo uses advanced Natural Language Understanding to analyze semantic context, creating more natural conversations with reduced interruptions and latency.

Namo's Solution: Uses Natural Language Understanding to analyze the semantic meaning and context of speech, distinguishing between:

• ✅ Complete utterances (user is done speaking)
• 🔄 Incomplete utterances (user will continue speaking)

🚀 Key Features

• Semantic Understanding: Analyzes meaning and context, not just silence
• Ultra-Fast Inference: <19ms for specialized models, <29ms for multilingual
• Lightweight: ~135MB (specialized) / ~295MB (multilingual)
• High Accuracy: Up to 97.3% for specialized models, 90.25% average for multilingual
• Production-Ready: ONNX-optimized for real-time, enterprise-grade applications
• Easy Integration: Standalone usage or plug-and-play with VideoSDK Agents SDK

Available Models

Namo offers both specialized single-language models and a unified multilingual model:

Multilingual Model (Recommended)

• Model: Namo-Turn-Detector-v1-Multilingual
• Base: mmBERT
• Languages: All 23 supported languages
• Inference: <29ms
• Size: ~295MB
• Average Accuracy: 90.25%

📊 Performance Benchmarks for Multilingual Model

Evaluated on 25,000+ diverse utterances across all supported languages.

Average Accuracy: 90.25% across all languages

For detailed performance metrics of individual specialized models, visit their respective model pages.

Specialized Single-Language Models

• Architecture: DistilBERT-based
• Inference: <19ms
• Size: ~135MB each

📊 View Full Collection: All Namo Models on HuggingFace

ONNX Quantization Benefits

All Namo models are optimized using ONNX quantization, which reduces model complexity while maintaining high accuracy:

• 2.19x relative speedup compared to unquantized models
• Inference time: Reduced from 61.3ms → 28.0ms
• Throughput: More than doubled
• Accuracy Impact: Negligible
• Latency: Virtually zero perceptible delay in conversations

🚀 Try It Yourself!

We’ve provided an inference script to help you quickly test these models. Just plug it in and start experimenting!

Integration with VideoSDK Agents

For seamless integration into your voice agent pipeline:

from videosdk_agents import NamoTurnDetectorV1, pre_download_namo_turn_v1_model

# Download model files (one-time setup)
# For multilingual (default):
pre_download_namo_turn_v1_model()

# For specific language:
# pre_download_namo_turn_v1_model(language="en")

# Initialize turn detector
turn_detector = NamoTurnDetectorV1()  # Multilingual
# turn_detector = NamoTurnDetectorV1(language="en")  # English-specific

# Add to your agent pipeline
from videosdk_agents import CascadingPipeline

pipeline = CascadingPipeline(
    stt=your_stt_service,
    llm=your_llm_service,
    tts=your_tts_service,
    turn_detector=turn_detector  # Namo integration
)

📚 Complete Integration Guide: VideoSDK Agents Documentation

🔧 Training & Testing

Each model includes Colab notebooks for training and testing:

• Training Notebooks: Fine-tune models on your own datasets
• Testing Notebooks: Evaluate model performance on custom data

Visit individual model pages for notebook links:

• Multilingual Training Notebook
• Specialised Model Training Notebook
• Testing Notebook

License

This project is licensed under the Apache License 2.0.

Citation

If you use Namo-v1 in your research or project, please cite:

@software{namo2025,
  title = {Namo Turn Detector v1: Semantic Turn Detection for Conversational AI},
  author = {VideoSDK Team},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/collections/videosdk-live/namo-turn-detector-v1-68d52c0564d2164e9d17ca97}
}

---