Streaming

PyTorch Serving

X86 + L4T

Serving PyTorch inference code within Savant pipelines, demonstrating GPU memory interaction and torch hub integration.

PyTorch Serving demo

Key Features

  • PyTorch inference in Savant
  • Torch hub model integration
  • GPU memory management
  • Real-time processing

Shows how to integrate PyTorch inference directly into Savant pipelines using Torch hub models with proper GPU memory management for real-time processing.

View on GitHub ← All Cases