Build an AI-powered Pet Detector in Visual Studio Code

on YouTube ↗31m

" summary

An AI pet detector gets built inside Visual Studio Code using TensorFlow, Jupyter support and Azure Machine Learning; training uses MobileNet transfer learning on the Oxford Pets dataset and runs on a remote GPU via SSH. Initial transfer learning yields 80% in ~25s, distributed hyperparameter search across a four‑node GPU cluster with bandit early stopping pushes accuracy to ~90% after tuning, and the model is deployed as a web service from VS Code.

AI-written summary · about

" chapters (12)

1. 0:00AI pet detection intro
2. 7:52Oxford pet dataset
3. 10:16Train model in Jupyter
4. 12:05Transfer learning for accuracy
5. 13:30Hyperparameter tuning
6. 16:18Operationalize in VS Code
7. 22:21Notebook to Python module
8. 23:23Deploy as web service
9. 26:00Test pet detector
10. 27:03End-to-end workflow
11. 28:06Next steps
12. 29:10Q&A

" related talks

• 40mServerless Machine Learning 101Tania Allard · PyCon India 2020
• 35mAI is for the birdsGreg Blomquist · DevConf.CZ 2025
• 01h 00m5 Levels of AI in Microsoft AzureAlan Smith · CloudBurst 2022
• 24mDon't want to miss out on the AI wave?Alyssa Ong · GitHub Constellation Singapore: Explore 2018
• 01h 02mAI on the EdgeNigel Parker, Dan Glover · NDC Sydney 2018
• 33mBuilding end-to-end computer vision solutions from deep learning modelsVanja Paunic · Data Science Conference 4.0
• 32mYou Can Do Deep Learning!William Horton · PyOhio 2018
• 01h 00mAI for Every DeveloperGian Paolo Santopaolo · .NET Conf 2018