vgg16

VGG16

Input

Shape : (1,3,224,224)

Output

class_count=3
+ idx=0
  category=963[pizza, pizza pie ]
  prob=14.844779968261719
  value=47
+ idx=1
  category=927[trifle ]
  prob=12.971166610717773
  value=34
+ idx=2
  category=926[hot pot, hotpot ]
  prob=10.52105712890625
  value=17
Script finished successfully.

Usage

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

For the sample image,

$ python3 vgg16.py

If you want to specify the input image, put the image path after the --input option.

$ python3 vgg16.py --input 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 vgg16.py --video VIDEO_PATH

Reference

Very Deep Convolutional Networks for Large-Scale Image Recognition

VGG16 - Torchvision

Keras Applications : VGG16

keras2caffe

Framework

TensorFlow 2.12.0

Netron

vgg16_pytorch_quant_recalib.tflite

vgg16_pytorch_float32.tflite