Skip to content

Latest commit

 

History

History
25 lines (24 loc) · 1.51 KB

INSTALL.md

File metadata and controls

25 lines (24 loc) · 1.51 KB
Raw

Installation

  1. Create and activate an environment with all required packages:
conda env create -f environment.yml
conda activate motsynth
  1. Install the additional modules: Tracktor, Torchreid and Trackeval:
pip install https://github.com/phil-bergmann/tracking_wo_bnw/archive/master.zip
pip install https://github.com/KaiyangZhou/deep-person-reid/archive/master.zip
pip install https://github.com/JonathonLuiten/TrackEval/archive/master.zip
  1. Optionally modify the path OUTPUT_DIR in configs/path_cfg.py, download TrackEval's ground truth data from here and place it in at MOTCHA_ROOT. E.g.:
MOTCHA_ROOT=$(python -c "from configs.path_cfg import MOTCHA_ROOT; print(MOTCHA_ROOT);")
wget -P $MOTCHA_ROOT https://omnomnom.vision.rwth-aachen.de/data/TrackEval/data.zip
  1. Optionally modify the path OUTPUT_DIR in configs/path_cfg.py, and download our trained models:
OUTPUT_DIR=$(python -c "from configs.path_cfg import OUTPUT_DIR; print(OUTPUT_DIR);")
mkdir ${OUTPUT_DIR}/models
wget -P ${OUTPUT_DIR}/models https://vision.in.tum.de/webshare/u/brasoand/motsynth/resnet50_fc512_reid_epoch_19.pth
wget -P ${OUTPUT_DIR}/models https://vision.in.tum.de/webshare/u/brasoand/motsynth/maskrcnn_resnet50_fpn_epoch_10.pth