Darknet is an open source neural network framework. It is a fast and highly accurate framework for real time object detection (also can be used for images). In the provided notebook, a fast and simple approach is shown to detect one or more objects within an image.
- GPU was enabled
- Cloned into the darknet repository
- Built darknet into the chosen folder
- Downloaded weights in order to use those weights on unseen images
- Made predictions on several images
- Darknet used here is pretrained on COCO dataset. Thus, making predictions is faster and simpler.