Kursplan

Introduction to Edge AI and Nano Banana

  • Key characteristics of edge-AI workloads
  • Nano Banana architecture and capabilities
  • Comparing edge vs cloud deployment strategies

Preparing Models for Edge Deployment

  • Model selection and baseline evaluation
  • Dependency and compatibility considerations
  • Exporting models for further optimization

Model Compression Techniques

  • Pruning strategies and structural sparsity
  • Weight sharing and parameter reduction
  • Evaluating compression impacts

Quantization for Edge Performance

  • Post-training quantization methods
  • Quantization-aware training workflows
  • INT8, FP16, and mixed-precision approaches

Acceleration with Nano Banana

  • Using Nano Banana accelerators
  • Integrating ONNX and hardware backends
  • Benchmarking accelerated inference

Deployment to Edge Devices

  • Integrating models into embedded or mobile applications
  • Runtime configuration and monitoring
  • Troubleshooting deployment issues

Performance Profiling and Trade-off Analysis

  • Latency, throughput, and thermal constraints
  • Accuracy vs performance trade-offs
  • Iterative optimization strategies

Best Practices for Maintaining Edge-AI Systems

  • Versioning and continuous updates
  • Model rollback and compatibility management
  • Security and integrity considerations

Summary and Next Steps

Krav

  • An understanding of machine learning workflows
  • Experience with Python-based model development
  • Familiarity with neural network architectures

Audience

  • ML engineers
  • Data scientists
  • MLOps practitioners
 14 timer

Antall deltakere


Pris per deltaker

Referanser (1)

Kommende kurs

Relaterte kategorier