Key Features And Enhancements
This Vitis AI Library release includes the following key features and enhancements:
- New Board Support
- Versal® ACAP: VEK280 is supported in this release (early access only).
- New Runtime Support
- ONNX Runtime is supported in this release. The quantized model in the ONNX format can be directly run on the target by ONNX Runtime.
- New Model Libraries
- Following are the new model libraries supported in this release.
- New Model Support
- Added 14 new PyTorch models
- Added one new TensorFlow2 model
- New CPU Ops Supported
- Added three CPU Ops.
Models based on the Caffe framework are no longer maintained and released.
The Vitis AI Library 3.0 is tested with the following images.
The following platforms and evaluation boards (EVB) are supported by the Vitis AI Library 3.0.
|Zynq® UltraScale+™ MPSoC ZU9EG||Xilinx ZCU102||1.1|
|Zynq UltraScale+ MPSoC ZU7EV||Xilinx ZCU104||1.0|
|Zynq UltraScale+ MPSoC||Xilinx Kria™ KV260||1.0|
|Versal AI Core series VC1902||Xilinx VCK190||Production|
|Versal AI Core series VE2802||Xilinx VEK280||ES1|
|Versal AI Core series VCK5000 Data Center development kit|
- For VCK5000, only production cards are supported. Use Vitis AI 1.4.x if using VCK5000-ES1 cards.
- Alveo U50 and U280 cards are no longer supported. Use Vitis AI 1.4.x if using these cards.
- Due to limitations of the Docker environment, MultiTask demos cannot run in DRM mode for Data Center targets.