/crowd-counting

Primary LanguageJupyter Notebook

Crowd Counting

Create virtual environment

python3 -m venv venv
source ./venv/bin/activate
pip install -r requirements.txt

Download JHU Crowd++ dataset and extract into ./jhu_crowd_v2.0 path

For starting the training run

python -m crowdnet.train

For inference run

python -m crowdnet.predict --model-path MODEL_PATH --img-path IMAGE_PATH

Download pretrained weights