This is a simple pytorch implementation for SSD for object detection.
The model was trained using the African Wildlife Dataset. The annotation.txt file is created as:
python annotation.py
which stores the image and its metadata file addresses:
data/AfricanWildlife/rhino/347.jpg data/AfricanWildlife/rhino/347.txt
data/AfricanWildlife/rhino/066.jpg data/AfricanWildlife/rhino/066.txt
data/AfricanWildlife/rhino/146.jpg data/AfricanWildlife/rhino/146.txt
data/AfricanWildlife/rhino/233.jpg data/AfricanWildlife/rhino/233.txt
data/AfricanWildlife/rhino/345.jpg data/AfricanWildlife/rhino/345.txt
data/AfricanWildlife/rhino/278.jpg data/AfricanWildlife/rhino/278.txt
data/AfricanWildlife/rhino/342.jpg data/AfricanWildlife/rhino/342.txt
data/AfricanWildlife/rhino/242.jpg data/AfricanWildlife/rhino/242.txt
data/AfricanWildlife/rhino/043.jpg data/AfricanWildlife/rhino/043.txt
data/AfricanWildlife/rhino/126.jpg data/AfricanWildlife/rhino/126.txt
python train.py
python predict_wildlife.py
