mobilenetv2

MobileNetV2

Input

Ailia input shape: (224, 224, 3)
Range: [0, 255] 8-bit unsigned integer

Output

class_count=3
+ idx=0
  category=409[analog clock ]
  prob=114
+ idx=1
  category=892[wall clock ]
  prob=56
+ idx=2
  category=826[stopwatch, stop watch ]
  prob=8

Usage

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

For the sample image,

$ python3 mobilenetv2.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 mobilenetv2.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 mobilenetv2.py --video VIDEO_PATH --savepath SAVE_VIDEO_PATH

Reference

MobileNetV2: Inverted Residuals and Linear Bottlenecks

Framework

TensorFlow 2.4.1, 1.15

Netron

• mobilenetv2_quant_recalib.tflite
• mobilenetv2_quant.tflite
• mobilenetv2_float.tflite