Skip to content

WIP implementation of "Box2Pix: Single-Shot Instance Segmentation by Assigning Pixels to Object Boxes"

License

Notifications You must be signed in to change notification settings

TheCodez/pytorch-box2pix

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

30 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

[WIP] pytorch-box2pix alt text

Inofficial PyTorch implementation of Box2Pix: Single-Shot Instance Segmentation by Assigning Pixels to Object Boxes (Uhrig et al., 2018).

TODO:

This is needed to get the project in a state where it can be trained:

  • mAP metric

Instance segmentation can be added later as it's just a post processing step.

Requirements

  • Install PyTorch (pytorch.org)
  • pip install -r requirements.txt (Currently requires torchvision master)
  • Download the Cityscapes dataset

Usage

Train model:

python train.py --dataset-dir 'data/cityscapes'

About

WIP implementation of "Box2Pix: Single-Shot Instance Segmentation by Assigning Pixels to Object Boxes"

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages