Skip to content

Latest commit

 

History

History
148 lines (122 loc) · 6.38 KB

README.md

File metadata and controls

148 lines (122 loc) · 6.38 KB

Vitis AI Runtime v1.3

Introduction

Vitis AI Run time enables applications to use the unified high-level runtime API for both cloud and edge. Therefore, making cloud-to-edge deployments seamless and efficient. The Vitis AI Runtime API features are:

  • Asynchronous submission of jobs to the accelerator
  • Asynchronous collection of jobs from the accelerator
  • C++ and Python implementations
  • Support for multi-threading and multi-process execution

For edge users, click Quick Start For Edge to get started quickly.

For cloud users, click Quick Start For Alveo to get started quickly.

Vitis AI Runtime directory structure introduction

VART
├── README.md
├── adas_detection
│   ├── build.sh
│   └── src
├── common
│   ├── common.cpp
│   └── common.h
├── inception_v1_mt_py
│   ├── inception_v1.py
│   └── words.txt
├── pose_detection
│   ├── build.sh
│   └── src
├── resnet50
│   ├── build.sh
│   ├── src
│   └── words.txt
├── resnet50_mt_py
│   ├── resnet50.py
│   └── words.txt
├── segmentation
│   ├── build.sh
│   └── src
├── squeezenet_pytorch
│   ├── build.sh
│   ├── src
│   └── words.txt
└── video_analysis
	├── build.sh
	└── src

Quick Start For Edge

Setting Up the Host

Follow Setting Up the Host to set up the host for edge.

Setting Up the Target

Follow Setting Up the Target to set up the target.

Running Vitis AI Examples

Follow Running Vitis AI Examples to run Vitis AI examples.

Quick Start For Alveo

Setting Up the Host

  1. Click Setup Alveo Accelerator Card with HBM for DPUCAHX8H/L to set up the Alveo Card.

  2. Download the xclbin files from here. Untar it, choose the Alveo card and install it. Take U50 as an example.

cd /workspace
wget https://www.xilinx.com/bin/public/openDownload?filename=alveo_xclbin-1.3.1.tar.gz -O alveo_xclbin-1.3.1.tar.gz
tar -xzvf alveo_xclbin-1.3.1.tar.gz
cd alveo_xclbin-1.3.1/U50/6E300M
sudo cp dpu.xclbin hbm_address_assignment.txt /usr/lib

This step is also described in DPUCAHX8H/L Overlay Usage.

Running Vitis AI Examples

Suppose you have downloaded Vitis-AI, entered Vitis-AI directory, and then started Docker. Thus, VART is located in the path of /workspace/demo/VART/ in the docker system.

/workspace/demo/VART/ is the path for the following example.

If you encounter any path errors in running examples, check to see if you follow the steps above.

  1. Download the vitis_ai_runtime_r1.3.0_image_video.tar.gz package and unzip it.

    cd /workspace/demo
wget https://www.xilinx.com/bin/public/openDownload?filename=vitis_ai_runtime_r1.3.0_image_video.tar.gz -O vitis_ai_runtime_r1.3.0_image_video.tar.gz
tar -xzvf vitis_ai_runtime_r*1.3*_image_video.tar.gz -C VART

  2. Download the model. For each model, there will be a yaml file which is used for describe all the details about the model. In the yaml, you will find the model's download links for different platforms. Please choose the corresponding model and download it. Click Xilinx AI Model Zoo to view all the models.

    • Take resnet50 of U50 as an example.
      cd /workspace
  wget https://www.xilinx.com/bin/public/openDownload?filename=resnet50-u50-r1.3.1.tar.gz -O resnet50-u50-r1.3.1.tar.gz
    • Install the model package.
      If the /usr/share/vitis_ai_library/models folder does not exist, create it first.
      sudo mkdir /usr/share/vitis_ai_library/models

    Then install the model package.

      tar -xzvf resnet50-u50-r1.3.1.tar.gz
  sudo cp resnet50 /usr/share/vitis_ai_library/models -r

  3. Compile the sample, take resnet50 as an example.

    cd /workspace/demo/VART/resnet50
bash -x build.sh

  4. Run the example, take U50 platform as an example.

    ./resnet50 /usr/share/vitis_ai_library/models/resnet50/resnet50.xmodel

    Note that different alveo cards correspond to different model files, which cannot be used alternately.

Launching Commands for VART Samples on U50/U50lv/U280
No. Example Name Command
1 resnet50 ./resnet50 /usr/share/vitis_ai_library/models/resnet50/resnet50.xmodel
2 resnet50_mt_py /usr/bin/python3 resnet50.py 1 /usr/share/vitis_ai_library/models/resnet50/resnet50.xmodel
3 inception_v1_mt_py /usr/bin/python3 inception_v1.py 1 /usr/share/vitis_ai_library/models/inception_v1_tf/inception_v1_tf.xmodel
4 pose_detection ./pose_detection video/pose.webm /usr/share/vitis_ai_library/models/sp_net/sp_net.xmodel /usr/share/vitis_ai_library/models/ssd_pedestrian_pruned_0_97/ssd_pedestrian_pruned_0_97.xmodel
5 video_analysis ./video_analysis video/structure.webm /usr/share/vitis_ai_library/models/ssd_traffic_pruned_0_9/ssd_traffic_pruned_0_9.xmodel
6 adas_detection ./adas_detection video/adas.webm /usr/share/vitis_ai_library/models/yolov3_adas_pruned_0_9/yolov3_adas_pruned_0_9.xmodel
7 segmentation ./segmentation video/traffic.webm /usr/share/vitis_ai_library/models/fpn/fpn.xmodel
8 squeezenet_pytorch ./squeezenet_pytorch /usr/share/vitis_ai_library/models/squeezenet_pt/squeezenet_pt.xmodel