When I configured the hls4ml model, I used io_parallel as the io_type, and it can work as usual.
But if I set the io_type as io_stream, it gave an error about:
ERROR: [HLS 200-660] firmware/myproject_axi.cpp:14: the stream pragma must be specified with a constant stream depth.
My model is a large network with quite a lot calculations, so I think I need to use io_stream. Do you have any idea about this error? And I did not add #Pragma about stream in my C++ code. Should I add something related to stream?
Hi! We are using the extension API (referring to the kl_layer template) to create a custom layer. My the other question and details about our model ->#(https://github.com/fastmachinelearning/hls4ml/issues/787)
When I configured the hls4ml model, I used io_parallel as the io_type, and it can work as usual.
But if I set the io_type as io_stream, it gave an error about: ERROR: [HLS 200-660] firmware/myproject_axi.cpp:14: the stream pragma must be specified with a constant stream depth.
My model is a large network with quite a lot calculations, so I think I need to use io_stream. Do you have any idea about this error? And I did not add #Pragma about stream in my C++ code. Should I add something related to stream?
Thank you in advance. Yuming