Lesson ml-17 · ML Portfolio

Recycle vs Trash

Demo student F · 15–16 · ML · Low-code image classifier

Capstone: train two classes on your own photos, run live predict with confidence bars, and document data ethics — the same loop as Teachable Machine inside the ML track.

Class A · Recycle

♻️

📦

🥤

📰

🧴

＋

18 samples · webcam captures

Class B · Trash

🍌

🥡

☕

🧻

🍕

＋

16 samples · balanced set

Recycle

78%

Trash

22%

Predict: Plastic bottle → likely Recycle (rinse first — check local rules).

Data ethics note (ml-17): Only my own desk items · no faces · no photos of classmates · v1 worked on bottles only — v2 added food waste after mislabeling a banana peel.