Come As You Are and Learn to Code Your First AI
summary
A hands-on pipeline using a pretrained YOLOv8 (Ultralytics) model and a Roboflow dataset (~11,000 images, COCO-referenced) to do transfer learning for rock-paper-scissors detection. The workflow goes from notebook to Docker to GPU training (cost example ≈ €1.90/hour), fine-tuning a small model (from ~3.2M to ~12M parameters) for 5–10 epochs and validating with metrics (confusion matrix, F1). A live Streamlit webcam app is built for real-time inference demo.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 10:49AI basics
- 32:16Cloud dev environments
- 45:28Python for AI
- 1:02:25Demo with YOLO
- 1:20:37Model training
- 1:32:03App deployment
- 1:43:13AI careers
- 1:50:51AI ethics
- 1:54:23Q&A
related talks
45mWelcome to the AI jungle! Now what?
55mTutorial: Rock, Paper, Scissors! Build an AI-powered Interactive Game With Argo CD and Kubeflow
26mWelcome to the AI Jungle! Now What?
34mBuild an AI-powered Pet Detector in Visual Studio Code
24mDon't want to miss out on the AI wave?
56mCan I do AI?
40mMachines are Learning
37mHands on learning experience for non-technical people