Interested in implemented AI at the edge with the Avnet platforms ?
Check out my new designs for Vitis-AI 1.3, the latest edge AI solution for Xilinx based platforms.
Vitis-AI 1.3 flow for Avnet Vitis platforms
This project provides detailed instructions for targeting the Xilinx Vitis-AI 1.3 flow to the following Avnet Vitis 2020.2 platforms:
- Ultra96-V2 Development Board
- UltraZed-EV SOM (7EV) + FMC Carrier Card
- UltraZed-EG SOM (3EG) + IO Carrier Card
Vitis-AI 1.3 includes a Model Zoo with over 90 pre-built models, ready for deployment
The model zoo contains various types of models such as:
- classification
- detection
- segmentation
which target a variety of applications
- face applications
- pose estimation
- automotive (ADAS/AD)
- medical
- surveillance
- etc…
Check out these projects which describe how to leverage these pre-built models to build cool applications
Head Pose Estimation on Ultra96-V2
License Plate Recognition with Vitis-AI
Vitis-AI 1.3 supports many industry standard framewroks, allowing the deployment of custom models to the edge.
- Caffe
- PyTorch
- TensorFlow & TensorFlow 2
- TVM
Check out these projects which describes how to use TensorFlow framework to train and deploy a custom deep neural network to the Ultra96-V2 development board
Deploying the Dobble Challenge on the Ultra96-V2
Kudos to zebular13 and kevinkeryk for their collaboration on these dobble projects 
Do you have a project that you want to share ?
Publish your project on hackster.io, and enroll in Xilinx's Big Xcellent Adventure.
Every month, the Xilinx team will be browsing the Hackster community for impressive projects showcasing their technology. Winners will receive $500 and get their project featured on developer.xilinx.com !
Regards,
Mario Bergeron.