YOLOV2Config
ModelConfigYOLOV2Config(num_classes: int = 80, in_channels: int = 3, num_anchors: int = 5, anchors: tuple[tuple[float, float], ...] = _COCO_ANCHORS, score_thresh: float = 0.5, nms_thresh: float = 0.5, lambda_coord: float = 5.0, lambda_noobj: float = 0.5, darknet_reorg: bool = False)Configuration for YOLOv2 / YOLO9000 (Redmon & Farhadi, CVPR 2017).
Args:
num_classes: Number of object classes C (COCO default: 80).
in_channels: Input image channels (default 3).
num_anchors: Number of anchor boxes per cell A (default 5).
anchors: Anchor (width, height) pairs relative to a 32-pixel cell.
Must have exactly num_anchors entries.
score_thresh: Minimum objectness × class-prob score for inference.
nms_thresh: IoU threshold for NMS at inference.
lambda_coord: Up-weighting for box regression loss.
lambda_noobj: Down-weighting for no-object confidence loss.
darknet_reorg: Reproduce the original flat-index passthrough layout.
Defaults to False for legacy user configurations; published
pretrained weights select True automatically.
Examples
>>> from lucid.models.vision.yolo import YOLOV2Config
>>> cfg = YOLOV2Config()
>>> cfg.num_classes
80
>>> cfg.model_type
'yolo_v2'