Dockerized VOTS 2024 code

Download the data

To download the data, you need to install the vot-toolkit in any environment. This step only needs to be done ONCE.

pip install vot-toolkit

Then download the data, run the following command (it will download 26 GiB of data)

vot initialize vot2024/main

This will download the data. The main thing is that you should have a sequences folder with all the challenge sequences from VOTS2023/2024.

Create a directory to store results and logs (for debugging):

mkdir results logs

Run the tracker

1. Build the Docker Image:

docker build -t tracker-image .

2. Run the Docker Container: Make sure the results and sequences directories exist on your host machine before running this command. This will also ensure that the container can write outputs to these directories.

docker run --gpus all -v $(pwd)/results:/app/results -v $(pwd)/sequences:/app/sequences -v $(pwd)/logs:/app/logs tracker-image

This command mounts the results and sequences directories from your current directory to the container and runs the default command specified in the Dockerfile.