John Boik

Great @dash14, thanks. I look forward to the changes. I'm sure you are familiar with vis.js, which writes to a HTML canvas rather than SVG. (In my case, an SVG...

I have the same issue as @kinow for the simple `UNWIND(RANGE(1, 10)) AS z RETURN z` test. I'm using Neo4j version 4.4.3. The hack solution I found was to alter...

Related to my previous post, and perhaps a topic deserving of documentation, suppose that I have a custom node for a function that looks like `z = f(x, y)`. In...

Thanks @bvdmitri for answering my questions. That helped quite a bit. But I think there is still a bit more to the story. I did not understand until I read...

Thanks @bvdmitri, that was very helpful. But I think a tweak is still needed. In the rules for calculating the backward messages, say, `x1`, the updated marginal for `x1` is...

My bad, @bvdmitri, I misunderstood the meaning of a pipeline with multiple messages, and my tweak was not necessary (although it does save a little computation). A working version using...

It looks like the problem is a bit worse. It appears that `circuit.ToJSON()` ignores both `xhl` and `x12` in DSS files for transformers, and instead uses the default value of...

@ChrisRackauckas, I wouldn't think so, but could this be related to the new DiffEqBase issue: https://github.com/SciML/DiffEqBase.jl/issues/1088?

@Vaibhavdixit02 Thanks for the reply. I use the extra variable(s) to send the current batch of training data to the callback function for plotting. It is very useful to me,...

With the new changes, perhaps the recommended way to pass data from the loss function to the callback is via a global variable. That seems to work fine. Is there...