Vladimir

Current `QSeparableConv2D` layer follows Keras implementation and doesn't apply activation after the depthwise step. The readme refers to the old implementation based on MobileNet that does this. That implementation is...

@julesmuhizi Don't you also need to add the new layer to `bn_folding_utils.py`? @zhuangh Related to folding of dense+bn, `convert_to_folded_model` will [not include](https://github.com/google/qkeras/blob/master/qkeras/utils.py#L318) `Dense` layers. a bug?

Can we postpone this until after we merge the Intel backend and flesh out a strtegy for integrating Phil's CNN implementation? It's just one command, we can run it at...

Hi Manny, You were given a few more tips to try, did you try them? I don't see that in the code you shared. Also, you are profiling base model,...

That looks like an issue in handling of UpSampling2D layer. We'll investigate. Unrelated to that, a model with 60 million parameters won't work in hls4ml. You'll have to reduce your...

You'll need to make quite some changes to get it working with Vitis HLS. Good news is that the Vitis support is nearing completion, and may come as early as...

Add `io_type='io_stream'` to `convert_from_keras_model`

Then you'll also need to use the `Resource` strategy and increase the `ReuseFactor` significantly.

This one is harder to get around by tweaking the configuration and it will require changes to the model. You'll have to compress your model somewhat (less filters, less neurons),...

This comes from Tensorflow's `model.save()` call, and as the message implies, there's a duplicate name in the dataset (i.e., weights). I would go through the model and check that the...