AutoML Vision: the PaaS of Machine Learning
summary
Shows how to build a "Shazam for TV series" using Google AutoML Vision: extract frames from episodes, hand-label a few hundred images, train AutoML Vision and call its REST API for predictions. Deploys the demo on Cloud Run, compares cloud-hosted and edge model exports (examples: ~2.5 MB vs ~5 MB), inspects misclassifications to iterate, and reports pricing and limits (free 1 training hour/month, about $20/hr training, ~$3 per 1,000 predictions with first 1,000 free).
AI-written summary · about
chapters (10)
- 0:00Speaker Introduction
- 1:00What is ML?
- 2:15ML Overview
- 4:44The Different Steps of ML
- 5:40Image Classification
- 7:35The Different Levels of PAAS
- 9:47The Series
- 12:41Results
- 18:17Application Pricing
- 20:08Conclusion
related talks
24mAutoML: No Code No Problem
57mBuilding smarter solutions with Machine Learning, from magic to reality
54mBuilding smarter apps with Machine Learning, from magic to reality
47mBuilding smarter solutions without machine learning expertise
47mBuilding smarter apps with Machine Learning, from magic to reality
42mUse Machine Learning in your code, without being a ML expert
22mTrain Custom Machine Learning Models with No Data Science Expertise
44mBuilding smarter solutions with no expertise in machine learning