Vladimir

In the same setup, do the previous steps (which are only related to the IP, not the accelerator backend) work if you add `export=True` to `build()`? My guess is that...

``` Total params: 482849 (1.84 MB) ``` Not gonna work. A much, much smaller model may work with io_stream. And a much, much, much smaller model may work with io_parallel....

Because it generates huge source files in io_parallel (you didn't pass the option, so it defaults to that), and the compiler simply fails. Looks at the memory spike on your...

Partly, it is the difference in algorithm behind this. To achieve the best performance, in io_parallel the im2col transformation of the convolution is manually unrolled with specific instructions for each...

Can you try with 2020.1, or switch to Vitis HLS 2023.2 or newer? Not sure but maybe the 2019.2 doesn't work anymore. I see that the documentation says we support...

None of the developers use Windows, so we don't know if it works now or not. We know that the the python integration (calling `compile()`) won't work. Installing Vitis in...

My experience with this is somewhat opposite: - Pooling with `II=CONFIG_T::reuse_factor` results in worse latency. Especially if the input is small, regardless of network architecture. For larger inputs I've seen...

From casually browsing sympy code it seems to me that `strict` has existed since 1.12, just the default value changed in 1.13? That particular line got there between 1.12 and...

WOW 😳 . That level of amateurism in the release process trumps everything we did for hls4ml releases 😄 . Thanks for looking into it. We should then fix the...

Fix this one then open another issue for the problems you encounter in pysr_luts test and we'll deal with those there after this one (maybe not immediately).