vita-epfl/detection-attributes-fields

PyTorch implementation of "Detecting 32 Pedestrian Attributes for Autonomous Vehicles"

36
/ 100
Emerging

Implements a Multi-Task Learning framework using composite fields for joint pedestrian detection and 32-attribute recognition (appearance, behavior, crossing intent) optimized for low-resolution autonomous driving scenarios. Introduces fork-normalization to stabilize gradient flow across numerous tasks during backpropagation, addressing scale imbalances in multi-task learning. Operates as an OpenPifPaf plugin, leveraging spatial field representations and instance-wise decoding on the JAAD dataset.

No commits in the last 6 months.

Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 13 / 25

How are scores calculated?

Stars

33

Forks

5

Language

Python

License

Last pushed

Oct 16, 2021

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/vita-epfl/detection-attributes-fields"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.