Jovan Mitrevski

This PR has not been very active recently, but I think we should make some decisions soon and incorporate something along this line.

This looks very useful to me. I have a few questions: * Do you understand why the test is failing? * What about the transient time, after we merge this,...

Hi, The batch size is often stripped. The issue may be that for pytorch (as for ONNX) input_shape[0] is generally not None but 1. I think line 4 needs to...

Just to make sure, you ran the following steps: ``` % git clone https://github.com/fastmachinelearning/example-models.git % cd example-models/ % hls4ml convert -c keras-config.yml ``` and then this command fails: ``` %...

I think lots of these types of errors are due to not including proper header files. They maybe worked before because the header files were included by other includes, but...

> For example, the `propagate_dense_precision` and `propagate_conv_precision` passes are obviously really nice - and could be expanded to other layers - but they look for attributes `quant_precision`, `weight_precision`, `bias_precision` which...

This pull request was presented at an hls4ml meeting: https://indico.cern.ch/event/1184299/contributions/4975803/attachments/2484362/4265432/QONNX%20Ingestion.pdf

Concerning the propagation of precisions down, my plan was that once the QKeras parsing was improved, you would enable it there, too. This was meant to be an optimization that...

I think I need to copy the type from the `input_var` in the `InplaceTensorVariable.__init__`. Let me test it a bit more extensively, though.

Some quartus-specific updates are in the QONNX PR #591. Should they be cherry-picked and added here, or is it fine to have them as part of that?