Training Trains Internship
Student login
← Computer Vision learning path
Computer Vision · Module 10 of 10

Deploying vision applications

Ship CV models to apps, edge devices and the cloud.

📖 Key ideas

  1. Export models to ONNX, TensorRT, TFLite or Core ML.
  2. Edge devices run models locally for privacy and speed.
  3. Cloud APIs (Google Vision, AWS Rekognition) offer ready features.
  4. Build demos with Streamlit or Gradio.
  5. Monitor accuracy in real conditions (lighting, cameras).
💡 Real-life example: Deploying CV is installing your trained eye into a real camera that works day and night.
TensorRTNVIDIA inference optimiser
Edge AIRunning AI on local devices
GradioQuick ML demo UI library
RekognitionAWS vision API