For Cloud (Alveo U50LV/U55C Cards, Versal VCK5000 Card) - 2.0 English

Vitis AI Library User Guide (UG1354)

Document ID
UG1354
Release Date
2022-01-20
Version
2.0 English
To run an example for the Alveo U50LV and U55C data center accelerator cards, and the Versal VCK5000 development card use these steps:
  1. After downloading the Vitis AI Library, navigate to the Vitis-AI directory, and then start Docker.
  2. Enter the directory of the sample and then compile it. Take resnet50 as an example.
    cd /workspace/demo/Vitis-AI-Library/samples/classification
    bash -x build.sh
  3. Run the sample.
    ./test_jpeg_classification resnet50 sample_classification.jpg
    If you want to run the program in batch mode, which means that the DPU processes multiple images simultaneously, you have to compile the entire Vitis AI Library according to the instructions in the Setting Up the Host section. Then the batch program will be generated in the build_dir_default location. Enter build_dir_default, take facedetect as an example, and execute the following command.
    ./test_classification_batch resnet50 <img1_url> [<img2_url> ...]
  4. To run the video example, run the following command:
    ./test_video_classification resnet50 <video_input.mp4> -t 8

    Here, video_input.mp4 is the video file that you have to prepare for input and -t is the number of threads.

  5. To test the performance of the model, run the following command:
    ./test_performance_classification resnet50 test_performance_classification.list -t 8 -s 60 

    Here, -t is the number of threads and -s is the number of seconds.

    To view a list of parameter information, enter -h.

    Note: The performance test program is automatically run in the batch mode.