Reconstructed Slides: From 46% to 90%: Fine-Tuning Tiny LLMs for On-Device Agents — Cormac Brick, Google
Source Video
From 46% to 90%: Fine-Tuning Tiny LLMs for On-Device Agents — Cormac Brick, Google
Method
This deck is reconstructed from the existing video frame captures by detecting likely slide regions with OpenCV, cropping/upscaling those regions, deduplicating similar crops, and OCRing the cropped slide images locally. It is a cleaner companion to the full-stage frame deck.
Reconstructed Slides
- Source frame:
slide-001.jpg - Crop:
full[0, 0, 960, 540]score157.79
- Source frame:
slide-002.jpg - Crop:
contour[0, 0, 960, 540]score161.12
- Source frame:
slide-003.jpg - Crop:
contour[0, 0, 960, 540]score176.1
- Source frame:
slide-004.jpg - Crop:
contour[0, 0, 960, 540]score166.32
- Source frame:
slide-005.jpg - Crop:
contour[0, 0, 960, 540]score166.43
- Source frame:
slide-006.jpg - Crop:
contour[0, 0, 960, 540]score167.04
- Source frame:
slide-007.jpg - Crop:
contour[0, 0, 960, 540]score166.46
- Source frame:
slide-008.jpg - Crop:
contour[0, 0, 960, 540]score166.77
- Source frame:
slide-009.jpg - Crop:
contour[0, 0, 960, 540]score176.67