Model Samples - 1.1 English

Vitis AI Library User Guide (UG1354)

Document ID
UG1354
Release Date
2020-03-23
Version
1.1 English

There are up to 14 model samples that are loacated in ~/Vitis-AI/Vitis-AI-Library/overview/samples. Each sample has the following four kinds of test samples.

  • test_jpeg_[model type]
  • test_video_[model type]
  • test_performance_[model type]
  • test_accuracy_[model type]

Take yolov3 as an example.

  1. Before you run the yolov3 detection example, you can choose one of the following yolov3 model to run.
    1. yolov3_bdd
    2. yolov3_voc
    3. yolov3_voc_tf
  2. Ensure that the following test program exists.
    1. test_jpeg_yolov3
    2. test_video_yolov3
    3. test_performance_yolov3
    4. test_accuracy_yolov3
    If the executable program does not exist, it can be compiled and generated as follows:
    	#sh -x build.sh
  3. To test the image data, execute the following command.
    	#./test_jpeg_yolov3 yolov3_bdd sample_yolov3.jpg

    You will see the printing result on the terminal. Also, you can view the output image: sample_yolov3_result.jpg.

  4. To test the video data, execute the following command.
    	#./test_video_yolov3 yolov3_bdd video_input.mp4 -t 8
  5. To test the model performance, execute the following command.
    	#./test_performance_yolov3 yolov3_bdd test_performance_yolov3.list -t 8
    You will see the printing result on the terminal.
  6. To test the model accurary, prepare your own image dataset, image list file and the ground truth of the images. Then execute the following command.
    	#./test_accuracy_yolov3 yolov3_bdd [image_list_file] [output_file]

After the output_file is generate, a script file is needed to automatically compare the results. Finally, the accuracy result can be obtained.