this is an implementation of YOLO object detection algorithme. the implementation is a mix between YOLO v1 and v2. I used pytorch models trained on ImageNet as backbone model then added final conv layer for the output THERE IS NO FULLY CONNECTED LAYERS
- torch==1.3.1
- torchvision==0.4.2
- opencv-python==4.1.2.30
- albumentations==0.4.3
- numba==0.46.0
clone and install the requirements
git clone https://github.com/AhmedGhazale/cifar100-classifier.git
cd cifar100-classifier
pip3 install -r requirements.txt
to run an image
python3 predict.py path/to/image
to run video
python3 video_demo.py path/to/video
the output will be a video named output.avi in the same directory
- download voc dataset
- edit the dataset path in config.py
- run
python3 train.py
TODO