Release 2.5.1

This is a minor release with updated .xmodel links for the ZUBoard_1CG, and small yolov3 notebook update for reproducibility.

Release v2.5.0

This release supports the latest PYNQ 3.0.1 image and Vitis AI 2.5 compilers and runtime and delivers the DPU-PYNQ overlay to 21 boards. Please see readme for list of newly supported boards. We also ship an additional model and example notebook for object detection.

Release v1.4.0

This release supports the latest PYNQ 2.7.0 image and Vitis AI 1.4 compilers and runtime. The host scripts now have added support for PyTorch and Tensorflow2, and hardware build tools have been upgraded to 2021.1. This release will support ZCU104, Ultra96 and KV260 boards.

Release v1.3.2

This release supports the latest PYNQ 2.6.0 image and Vitis-AI 1.3.2 compilers and runtime.

Release v1.3.0

This release supports the latest PYNQ 2.6.0 image. From this version going forward, DNNDK runtime is deprecated; we only support VART. This release is compatible with VAI 1.3.0. Due to lack of support for Vitis AI on Zynq 7000, DPU-PYNQ still only supports Zynq Ultrascale devices for now.

Release v1.2.0

This release supports the latest PYNQ 2.6.0 image. One extra VART example has been added. This release is compatible with VAI 1.2.1.

Release v1.1.2

This release has minor updates to the python code for a few DNNDK runtime. VART support has been added with an example notebook and a new DPU model.

Release v1.1.1

This is the 1st release of DPU-PYNQ. It supports PYNQ image v2.5 for Ultra96, ZCU104, and ZCU111 boards officially.