Major Updates in Elixir YOLO v0.2.0: YOLOX and Enhanced Custom Model Support

Elixir YOLO v0.2.0 is a significant update that brings three major enhancements: support for YOLOX, model-agnostic postprocessing, and a performance boost for postprocessing tasks. The addition of YOLOX support allows for greater licensing freedom compared to Ultralytics models, facilitating the use of custom datasets. The library has also removed previous constraints on model shapes, enabling it to work with various model outputs, which opens the door for custom-trained models. Performance improvements have resulted in a 100x speedup in postprocessing times, making real-time applications more feasible. Future updates will focus on training custom models, object tracking, and extending support to additional frameworks.