Vladimir

Set `IOType` to `io_stream`, for example `hls4ml.converters.convert_from_keras_model(model, ..., io_type='io_stream')`

There is no perfect solution since we cannot know the mangled name of the templated C++ function, but perhaps a less hacky solution can be explored. Instead the heavy hammer...

Depends on your config. Assuming you use `io_stream`, the limit will be related to the strategy used, since that affects the algorithm used for the CNN kernel. If you use...

This looks like an issue with your protobuf installation. Did you already have the onnx package installed and you upgraded it?It is possible that protobuf dependency wasn't upgraded, this happens...

With `-DNO_VIVADO` we will get different results from python simulation (the `compile()` feature) and the C simulation/synthesis from Vivado. In other layers we use the lookup tables for expensive math...

IMO we should move away from hardcoded `__version__` attribute and instead use something like `setuptools_scm`. That way the git tag becomes the version, and this information is available in the...

I think we should, though I would advocate for the more recent version than 2.4. Since version [2.6](https://github.com/tensorflow/tensorflow/releases/tag/v2.6.0) keras has been split into a separate package and this causes some...

Hi Lukas (LordScarface), Thanks for your interest in hls4ml. I see two problems here, first is that `QBatchNormalization` is created without any quantizers, which I could say is the issue...

I fixed the warnings. The current CNN implementations for io_parallel are going to be replaced with #600 so I did't test them thoroughly. The warning in dense_stream was due to...

I would say this is ready for review/merge. Far from complete as it is missing the resource strategy and some warnings about the interfaces still need tweaks, but it mostly...