Hi everyone, I’m working on a project on adaptive noise cancellation using an FPGA. I’ve tried different methods and architectures. Most recently, I trained a U-Net model, exported the weights to Vivado, implemented the RTL, and ran synthesis, but the neural network is not producing denoised output.
I also tried some other models, but they didn’t work due to the large number of parameters, and they ended up overutilizing the PYNQ’s LUTs and BRAMs.
Is anyone facing a similar issue, or can someone provide proper guidance?