Name		Name	Last commit message	Last commit date
parent directory ..
.vscode		.vscode
checkpoints		checkpoints
core		core
my_code		my_code
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
demo.sh		demo.sh
evaluate.py		evaluate.py
evaluate.sh		evaluate.sh
evaluate_single.py		evaluate_single.py
my_demo.sh		my_demo.sh
my_eval_and_save.py		my_eval_and_save.py
my_eval_and_save_bidirect.py		my_eval_and_save_bidirect.py
my_eval_and_save_bidirect_bdd.py		my_eval_and_save_bidirect_bdd.py
my_eval_and_save_bidirect_bdd_parallel.py		my_eval_and_save_bidirect_bdd_parallel.py
my_eval_and_save_bidirect_ytvis_2019.py		my_eval_and_save_bidirect_ytvis_2019.py
my_install.md		my_install.md
my_script.sh		my_script.sh
my_script_parallel.sh		my_script_parallel.sh
my_script_tesla_server.sh		my_script_tesla_server.sh
my_script_ytvis_2019_tesla_server.sh		my_script_ytvis_2019_tesla_server.sh
things_val_test_set.txt		things_val_test_set.txt
train.py		train.py
train.sh		train.sh

README.md

GMA Usage

Please check the my_install doc for installation details.

Download the optical flow model gma-sintel.pth from the official repo.
Run my_script.sh to extract optical flow information of the KITTI/BDD dataset. Remember to change the optical flow model and dataset path to yours. Both the optical flow visualization images and the optical flow data will be generated.

The generated optical flow data is large (KITTI: 36 GB/BDD: 431 GB). Ensure your hard disk has enough space.