Zynq Zcu102 custom model implementation to generate RTL code
I have developed 2D CNN VGG network custom model based on tensorflow and AI application for DPU and deployed the application on ZCU102 board. Now I would like to generate a RTL code from the Python based custom AI Application . Could anyone help me how to generate RTL from custom AI application based Python.
Re: Zynq Zcu102 custom model implementation to generate RTL code
Certainly something the DNNDK is not capable of. With the DNNDK the RTL which runs your network will always be the DPU, which you can implement as an IP core. There are other frameworks out there that generate RTL based on your python model, check out DNNWeaver for example.