Skip to content
This is a baseline framework for vehicle re-id on the VeRi dataset, implementated by PyTorch.
Python Shell
Branch: master
Clone or download
Latest commit 6eb9163 Apr 24, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
dataset Initial commit Apr 24, 2019
list Initial commit Apr 24, 2019
networks Initial commit Apr 24, 2019
.gitattributes
README.md
eval.py Initial commit Apr 24, 2019
eval.sh Initial commit Apr 24, 2019
eval_image.py
eval_reid.py Initial commit Apr 24, 2019
eval_track.py
reid_metric.py Initial commit Apr 24, 2019
train.py Initial commit Apr 24, 2019
train.sh Initial commit Apr 24, 2019

README.md

Vehiclereid Baseline

This is a baseline framework for vehicle re-id on the VeRi dataset, implementated by PyTorch.

The backbone is ResNet-50 or ResNet-101 which is trained by cross entropy loss.

This is a very simple baseline for image-based vehicle re-id without any tricks, augmentation, metric learning methods, and use of license plates and spatio-temporal information. Therefore, the results may be very poor.

Dependencies

  • PyTorch 1.0
  • Python3
You can’t perform that action at this time.