YOLOv2 (Darknet-19)
Redmon & Farhadi, 2017 — YOLO9000: Better, Faster, Stronger (arXiv:1612.08242)
Lucid port of darknet/yolov2.weights,
converted to Lucid-native safetensors.
Available weights
| Tag | mAP@0.5 | Params | GFLOPs | Size | Source |
|---|---|---|---|---|---|
COCO_2014 (default) | 48.1 | 51.0M | — | 194.5 MB | darknet |
Usage
import lucid.models as models
from lucid.models.weights import YOLOV2Weights
# default tag
model = models.yolo_v2(pretrained=True)
# explicit tag (enum or string)
model = models.yolo_v2(weights=YOLOV2Weights.COCO_2014)
model = models.yolo_v2(pretrained="COCO_2014")
# preprocessing travels with the weights
weights = YOLOV2Weights.COCO_2014
preprocess = weights.transforms()
out = model(preprocess(image)[None])
# ObjectDetectionOutput: per-query/proposal class logits + boxes
logits, boxes = out.logits, out.pred_boxes
Conversion
Converted from darknet/yolov2.weights via
python -m tools.convert_weights yolo_v2 --tag COCO_2014.
Key mapping + numerical parity verified against the source.
License
other — inherited from the original weights.
Citation
@inproceedings{redmon2017yolo9000,
title={YOLO9000: Better, Faster, Stronger},
author={Redmon, Joseph and Farhadi, Ali},
booktitle={IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
year={2017}
}