yolov3-tiny

yolov3-tiny

Input

• Shape : (1, 3, 416, 416)
• Range : [0.0, 1.0]

Output

• category : [0,79]
• probablity : [0.0,1.0]
• position : x, y, w, h [0,1]

Usage

Automatically downloads the onnx and prototxt files on the first run. It is necessary to be connected to the Internet while downloading.

For the sample image,

$ python3 yolov3-tiny.py

If you want to specify the input image, put the image path after the --input option.
You can use --savepath option to change the name of the output file to save.

$ python3 yolov3-tiny.py --input IMAGE_PATH --savepath SAVE_IMAGE_PATH

By adding the --video option, you can input the video.
If you pass 0 as an argument to VIDEO_PATH, you can use the webcam input instead of the video file.

$ python3 yolov3-tiny.py --video VIDEO_PATH

Reference

• YOLO: Real-Time Object Detection
• A Keras implementation of YOLOv3 (Tensorflow backend)
• keras-onnx

Framework

Keras 2.2.4

Model Format

ONNX opset=10

Netron

yolov3-tiny.opt.onnx.prototxt