imagenet21k

Imagenet21K

Input

(Image from https://github.com/pytorch/hub/raw/master/images/dog.jpg)

Ailia input shape: (1,3,299,299) Range: [0.0, 255.0]

Output

class_count=7
+ idx=0
  category=1531[spitz ]
  prob=84.49806213378906
+ idx=1
  category=1386[dog ]
  prob=81.74344635009766
+ idx=2
  category=155[domestic_animal ]
  prob=81.34649658203125
+ idx=3
  category=3[animal ]
  prob=81.01296997070312
+ idx=4
  category=1532[Samoyed ]
  prob=59.17914581298828
+ idx=5
  category=1385[bitch ]
  prob=38.642704010009766
+ idx=6
  category=1384[canine ]
  prob=37.53316879272461

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 imagenet21k.py

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

$ python3 imagenet21k.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 imagenet21k.py --video VIDEO_PATH

You can select a pretrained model by specifying -a mixer, resnet50 ,mobilenet or vit(default).

Reference

ImageNet21K

Model Format

ONNX opset = 14

Framework

pytorch

Netron

mobilenetv3_large_100.onnx.prototxt

mixer_b16_224_miil_in21k.onnx.prototxt

resnet50.onnx.prototxt

vit_base_patch16_224_miil_in21k