Skip to content

Model Deployment Edge

model-deployment-edge

Deploying ML models on edge-computing devices

Methods:

  1. Leverage hardware-accelerator
  2. Use mobile-optimized architecture
  3. Apply quantization
  4. Apply pruning

Formats: tensorflow-lite, json-format + bin

Platforms: Android, Raspberry Pi, google-coral

High-level APIs: ml-kit Customizable: tensorflow-lite, Google Cloud


*References

On this page