/BeamBench

Beam Prediction using multi-modal fusion

Primary LanguagePython

Part of the BeamBench Framework

BeamBench Framework by Maximilian Arnold for evaluating multi-modal beam prediction.

Installation

NOTE: This is for Ubunutu 18.04 with CUDA 11.4!

Documentation

See report and Slides in /docs/

Build docker

docker build -t beambench .

Run docker

chmod +x run_docker.sh
./run_docker

CD to your code folder

cd /workspace/home/.../challenge_net/

Download the models from

cd /workspace/home/.../challenge_net/

and put them into

results/models

Evaluate the models

Run the function

python3 challenge.py --gpu_id 0  --data_folder ./raw_data/test/ --csv ml_challenge_test_multi_modal.csv

Please update the raw test folder directory and the name if you changed the csv(normally not)

Support

Please let me know if the code doesnt work for you under:

maximilian.wolfgang.arnold@gmail.com