Dive into the world of image classification model training with PyTorch on Jetson Nano in this comprehensive 33-minute video. Learn transfer learning techniques, starting with re-training a ResNet-18 model on a cat/dog dataset. Follow step-by-step instructions for downloading data, re-training the model, and exporting it to ONNX format. Explore testing on live data streams and creating custom datasets for personalized models. Gain hands-on experience in applying trained models to live camera feeds, equipping you with practical skills for real-world AI applications on Jetson devices.
Jetson AI Fundamentals - Training Image Classification Models
Overview
Syllabus
- Introduction - Training Image Classification Models
- Getting Started - Transfer Learning with PyTorch
- Example 1 - Re-training on Cat/Dog Dataset
- Example 1 - Step 1: Downloading the data
- Re-training ResNet-18 Model
- Exporting Pytorch model to Onyx and testing the model
- Testing on a live data stream
- Training on a custom dataset
- Testing the trained model on a live camera feed
- Conclusion
Taught by
NVIDIA Developer
