NVIDIA In-Game Inferencing (NVIGI) Nemotron ASR Plugin Pack 1.0

Start with the pack overview, run a sample, and then open the guide for your integration path.

Getting Started

• Nemotron ASR Developer Pack Overview
  • 1. What’s Included
  • 2. Automatic Speech Recognition Models
  • 3. Accuracy and Performance
  • 4. Important Documentation in the Developer Pack
  • 5. Contents of the Developer Pack
  • 6. Shipping Checklist
  • Appendix A. Full Directory Listing
  • Appendix B. Migration from Earlier Documentation
• Getting Started with the Nemotron ASR Plugin Pack
  • 1. Prerequisites
  • 2. Run Your First ASR Sample
  • 3. Integrate into Your Application

Integration Guides

• Programming Guide: Automatic Speech Recognition with Nemotron ASR (GGML)
  • 1.0 Initialize and Shutdown
  • 2.0 Obtain the ASR Interface
  • 3.0 Create an ASR Instance
  • 4.0 Audio Input
  • 5.0 Receive Inferred Data
  • 6.0 Prepare the Execution Context
  • 7.0 Add ASR Inference to the Pipeline
  • 8.0 Destroy the Instance
  • 9.0 Unload the Interface
  • Appendix
• Silero VAD Speech-State Reporting
  • 1. Discover support
  • 2. Create the streaming instance
  • 3. Configure VAD at stream Start
  • 4. Read speech-state output
• Convert a NeMo ASR Model for the Nemotron ASR Plugin
  • 1. Compatibility contract
  • 2. Isolated Python environment
  • 3. Convert, quantize, and verify
  • 4. Standalone asr_quantize.exe
  • 5. Verify a model directory
  • 6. Install the converted model
  • 7. Troubleshooting

Samples

• Nemotron ASR Plugin Samples

Legal Notices

• 3rd PARTY SOFTWARE