Adding QKeras to conda forge

Is it possible to add QKeras to conda-forge? We need this to have a working hls4ml conda-forge recipe (See
https://github.com/fastmachinelearning/hls4ml/issues/790). I used Grayskull to try to build a recipe (see: https://conda-forge.org/docs/maintainer/adding_pkgs.html)

The main issue I think is also needing to add tensorflow-model-optimization to conda-forge first (but there is an example here: https://anaconda.org/HCC/tensorflow-model-optimization). It may also be possible to simply use that?

Start with the conda-forge guide for adding packages and the Grayskull recipe mentioned in the issue. Compare the tensorflow-model-optimization example and determine whether that dependency must be added first or can be used directly. Done means QKeras has a conda-forge recipe that enables a working hls4ml conda-forge recipe.

Written by the indexing model from the issue text.