Bundled Darknet model-definition (.cfg) files
=============================================

The .cfg files in this directory are the model definitions for YOLOv1, YOLOv2,
YOLOv3, and YOLOv4 (and their tiny/spp variants). They originate from the
Darknet project and are public domain:

    pjreddie/darknet  (YOLOv1/v2/v3)  https://github.com/pjreddie/darknet
    AlexeyAB/darknet  (YOLOv4)        https://github.com/AlexeyAB/darknet

The YOLOv1 cfgs (yolov1.cfg, yolov1-tiny.cfg) reproduce pjreddie's public-domain
yolo.cfg / tiny-yolo.cfg with the ``[connected] output`` and ``[detection] num``
set to the values of the released yolov1.weights / tiny-yolo.weights (the classic
7x7x30, 2-boxes-per-cell head). The LibreYOLO weight reader asserts byte-exact
consumption of those .weights files against these cfgs.

Darknet is released under the "YOLO LICENSE":

    "0. Darknet is public domain.
     1. Do whatever you want with it.
     2. Stop emailing me about it!"

LibreYOLO bundles these public-domain cfgs as the single source of truth for the
Darknet-lineage architectures: the runtime graph builder
(``libreyolo/models/darknet/net.py``) and the weight converter
(``weights/convert_darknet_weights.py``) both read them, so the architecture
definition never drifts between model and converter.

Only the .cfg text format and the numerical behaviour of the Darknet layers are
reproduced. No Darknet C source is copied into LibreYOLO.
