yolov5processor 1.2

Wrapper package for yolov5 inference:v3.0

Yolo v5 Inference Engine

Basically a wrapper for the detection program for ease of use.

How to install:

pip install yolov5processor

How to use:

from yolov5processor.infer import ExecuteInference 

model = ExecuteInference(weight="path-to-weight.pt", confidence=0.4, \
            img_size=640, agnostic_nms=False, gpu=False, iou=0.5)

image = cv2.imread("imagepath.jpg")
pred = model.predict(image)

Analyse the Output:

print(pred)

{
  "points": [100, 120, 400, 500],
  "conf": 0.7,
  "class": "apple"
}

points key represents the bounding box coordinates in the order x1, y1, x2, y2 respectively. conf key represents the confidence. class key represents the object class.

Update:

• Added compatibility with torch 1.9 and torchvision 0.10