Skip to content

Panoptic segmentation baseline implemented based on the MinkowskiEngine library

Notifications You must be signed in to change notification settings

PRBonn/MinkowskiPanoptic

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 

Repository files navigation

Panoptic segmentation network using MinkowskiNet as backbone and predicting offset and clustering

Installation

Install requirements

Install this repo as a package via

pip3 install -U -e .

how to run

Training

python3 scripts/train_model.py


Options:
  --w TEXT            weights to load
  --ckpt [ckpt_path]  checkpoint to resume training
  --nuscenes
  --mini              use mini split for nuscenes
  --seq [seq_num]     use a single sequence for train and val
  --id [exp_name]    set id of the experiment
  --help             Show this message and exit.

Evaluation

python3 scripts/evaluate_model.py --w path_to_weights


Options:
  --w [path_to_weights]   weights to load
  --save                  save ply predictions
  --save_testset          save predictions for test set
  --nuscenes              use nuscenes dataset
  --mini                  use mini split for nuscenes
  --seq [seq_num]         use a single sequence for train and val
  --help                  Show this message and exit.

About

Panoptic segmentation baseline implemented based on the MinkowskiEngine library

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages