Yolo v2 implementation in Chainer
YOLO9000: Better, Faster, Stronger link
See the project webpage for more details.
original code: https://github.com/pjreddie/darknet
・COCO
http://cocodataset.org/#download
・VOC
http://host.robots.ox.ac.uk/pascal/VOC/voc2012/#devkitpython to_chainer_converter.py --orig /path/to/original_model --name yolo_v2_chainerpython demo.py experiments/yolov2_update_608_test.yml --img_path ./data/dog.jpg --thresh 0.20 --nms_thresh 0.3 --save dog
nms: by class
nms_thresh = 0.3
img_thresh = ?python train.py experiments/yolov2_update_voc_416.yml
nms: by classpython evaluate.py experiments/yolov2_update_416_eval.yml --batchsize 32 --gpu 0
# fps of evaluation includes the time of loading images.
nms: by class
nms_thresh = 0.5
img_thresh = 0.001| Model | Dataset | darknet map | chainer map | cfg | darknet weight | Chainer | orig fps | chainer fps |
|---|---|---|---|---|---|---|---|---|
| YOLOv2 | VOC2007+20012 | 76.8 | 75.2 | link | link | link | **** | **** |
| YOLOv2 544×544 | VOC2007+20012 | 78.6 | 77.9 | link | link | link | ||
| Tiny YOLO | VOC2007+20012 | link | link | |||||
| YOLOv2 608×608 | COCO | link | link | link | 58.8 | 66.6 | ||
| Tiny YOLO | COCO | link | link | |||||
| prior version(YOLOv2) | COCO | link | link | link |