Right data at the right time: Tools demo
summary
Demo of a tool called Cider that manages deep‑learning image labels, stores pixelwise semantic segmentations, and exposes a metadata query UI that returns exact frames (example query motorbike + Germany + raining yields ~200 frames). Labels undergo review cycles and accepted snapshots are pushed as artifacts into an artifactory in MongoDB where datasets become ODP documents with stable IDs for training. A six‑person team develops and maintains the system and nightly artifact creation is asynchronous.
AI-written summary · about
chapters (6)
- 0:00Start
- 5:16Semantic Segmentation Labels
- 8:39Stabilization Phase
- 9:06Use Case
- 11:09Snapshots
- 13:50Fine Tuning Phase
related talks
01h 00mData, visualization, and designing AI
28mData Curation: Transforming Bytes into AI Gold
21mDataEngBytes 2023 - SYD-T2-04 - Suneeta Mall
29mLabeling Tools are Great, but What About Quality Checks?
30mStarting deep learning projects without enough labelled data
31mFrom Radar Echoes To Real-Time Recognition: Deep Learning on Another Planet -An Open Source Workflow
20mPragmatic Deep Learning for image labelling
01h 08mGoing multi-modal