darknet
darknet copied to clipboard
YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
`thresh` is always set to 0. If we don't need to use thresh, better to erase this variable.
Realsense2 camera (RGB) can be used.
## What is the purpose of the change Provide darknet the ability to do RTMP video streaming via FFMPEG. Now user can stream the detector result video stream to a...
Hi @AlexeyAB @cenit @WongKinYiu Following the pull request #6668 , It is a good practice to move the `free_image()` call inside `detect_image()` function in `darknet.py`. This change is suggested as...
I have added rotation support in detection mode, if anyone wants to have rotation then you can add `angle_detector` in cfg file, for example I want to have rotation between...
In `make_convolutional_layer` function, `l.output` is initialized to 0 by `xcalloc`. Then it is copied to `l.output_gpu`, and conditionally copied to `l.x_gpu` and `l.x_norm_gpu`. We can use `cuda_memset` to set these...
I have added a python file video_detector.py. It can be used to detect objects in a video. If a user wants to use this file for custom object detection than...
I have added a folder named custom_data which has required folders and files for custom object detection in a specific required format. It has a python and jupyter notebook file...
Fix for extreme verbose while training yolo models. "v3 ...", "total_bbox ..." verbose is supressed
I've implemented the script that generates text file with dataset images relative paths. It is needed for training.