Calculate mAP

Save dataset for mAP

Downloads

• mAP: https://github.com/Cartucho/mAP
• val2017 images: https://cocodataset.org/#download
• coco.names, val2017.txt: https://github.com/hhk7734/tensorflow-yolov4/tree/master/test/dataset

Script

v3

• 2021-02-21
• OS: Ubuntu 20.04
• TF: v2.4.1
• yolov4: v3.1.0

from yolov4.tf import YOLOv4, YOLODataset, create_mAP_input_files


yolo = YOLOv4()
yolo.config.parse_names("coco.names")
yolo.config.parse_cfg("yolov4-tiny.cfg")

yolo.make_model()
yolo.load_weights("/home/hhk7734/NN/yolov4-tiny.weights", weights_type="yolo")

dataset = YOLODataset(
    config=yolo.config,
    dataset_list="/home/hhk7734/NN/val2017.txt",
    image_path_prefix="/home/hhk7734/NN/val2017",
    training=False,
)

create_mAP_input_files(yolo, mAP_path="/home/hhk7734/mAP", dataset=dataset)

v2

• 2020-10-15
• OS: Ubuntu 20.04
• TF: v2.3.1
• yolov4: v1.2.1

from yolov4.tf import YOLOv4

yolo = YOLOv4(tiny=True)

yolo.classes = "coco.names"

yolo.make_model()
yolo.load_weights("yolov4-tiny.weights", weights_type="yolo")

dataset = yolo.load_dataset(
    "val2017.txt",
    training=False,
    image_path_prefix="/home/hhk7734/val2017",
)

yolo.save_dataset_for_mAP("/home/hhk7734/mAP", dataset)

• @param yolo
• @parma dataset
• @param mAP_path
• @parma images_optional: If True, images are copied to the mAP_path.
• @param num_sample: Number of images for mAP. If None, all images in dataset are used.

Calculate mAP

cd <mAP directory>

python3 -m pip install numpy matplotlib

python3 main.py

• --no-animation
• --no-plot
• --quiet