Developing and Deploying Vision AI with Dell and NVIDIA Metropolis

• 1

  Overview
• 2

  Revisions
• 3

  We value your feedback

• 1

  Overview

• 1

  Overview
• 2

  What is transfer learning?
• 3

  Overview of NVIDIA TAO pre-trained models for object detection
• 4

  NVIDIA DeepStream SDK overview
• 5

  NVIDIA Triton Inference Server overview
• 6

  NVIDIA DeepStream and Triton integration

• 1

  Overview
• 2

  Training the RetinaNet or ResNet18 model with TAO Toolkit and deploying it with DeepStream and Triton
• 3

  Software stack
• 4

  Dataset
• 5

  Terminology

• 1

  Overview
• 2

  Low-level tracker library comparison

• 1

  Overview

• 1

  Overview

• 1

  Appendix A: Backbones and networks matrix combination

• 1

  Appendix B: NGC registry pre-trained models

• 1

  Step 1: Train and optimize the model with TAO Toolkit
• 2

  Step 2: Generate TensorRT engines with TAO converter (from .etlt format to .engine format)
• 3

  Step 3: Configure the model repository
• 4

  Step 4: Update the config files
• 5

  Step 5: Run the Inference TensorRT DS-Triton

• 1

  Appendix D: Training specification files

• 1

  Appendix E: Gst-nvtracker Tracker library comparison

• 1

  Dell Technologies documentation
• 2

  Citations