How to use the deep learning IP core generated by deep learning hdl toolbox

Hi,

I have successfully generated a deep learning IP core and integrated it into the official reference design, where it works very well using the hw.predict command. Now, I would like to integrate this IP core into my own real-time processing design. I have implemented my own AXI Lite master block, but I need guidance on how to configure the IP core properly. The ip core has much registers, I do not know the sequence and which register I should write.

I wonder if there is any guidence, or how can I know how the Matlab function hW.predicet control the deep learning processor.

Looking forward to your reply!

Answer 1

[Anumeha]

Hi KH,

You can check out this documentation page for more information on how the deep learning IP core works in MATLAB:

https://www.mathworks.com/help/deep-learning-hdl/ug/deep-learning-processor-ip-core.html

Hope this helps!

[KH]

Thanks for your reply, I have read the document.

However, the documents are not clear, and I dont know which register should i read or write.

So I have to use wireshark to see how matlab communicate with the deep learning core!