Shenghui Li / Teaching
Machine Learning for IoT
Course tools and gesture data for the Arduino Nano 33 BLE Sense Rev2 labs.
Teaching design
Explore the Lab 0–3 teaching design — a password-protected overview for teaching collaborators. Sign in with the access details provided separately.
Edgelet
Install or update the Lab 2 package:
python -m pip install -U "edgelet[lab2]" -f https://lishenghui.github.io/teaching/mliot/edgelet-releases/Package index and available versions
Gesture dataset
119 provided recordings covering idle, circle, leftright, updown, snake and twist. The Lab 2 ZIP preserves the Lab 1 split: 95 training recordings and 24 fixed validation recordings, stored in the testing/ folder.
| Download | Use |
|---|---|
| Lab 2 dataset (JSON ZIP) | Extract into dataset/ next to the notebook. Includes training/, testing/ and manifest.csv. |
| Original Edgelet dataset (CBOR ZIP) | The base dataset shipped with Edgelet. |
| File sizes and SHA-256 checksums | Verify a download. |
To include your own Lab 1 recordings, export from your Edgelet workspace instead:
edgelet export-data --include-base --out lab1_data.zip