Christabella Irwanto

This is probably too late but an infer.net-specific workaround could be to use `Invoker.InvokeStatic(Variable, "Constant", args)` e.g. https://github.com/dotnet/infer/blob/0c75b5ef4190c5eee55e663741e645cb997b7236/test/TestPython/test_tutorials.py#L217

I guess the zero padding is like the pre-processing in order to obtain input x, and from the paper "M is the dimensionality of x", which would be 32x32? While...

Just to clarify, the purpose of the constant "scaling penalty" `c` is just to ensure accurate likelihood computations? Since the minimum would be the same with or without `c`. Comparison...

Thank you for the explanation!

Thank you very much for the reply! I just tried visualizing the weights in the provided notebook `Celeba48x48_22steps.ipynb` without changing anything else, and some gradients are also exploding even in...

Thanks a lot @geosada for pointing out the importance of the warm-up strategy! Although I have heard of starting with a small learning rate that [goes up and down again](https://bella.cc/zettels/learning_rate_scheduling_warm_up/),...

I also noticed that the CelebA notebook uses a per-pixel loss ``` loss_per_pixel = loss / image_size / image_size total_loss = l2_loss + loss_per_pixel ``` while for MNIST, it's the...

@crockpotveggies yeah a brief section in the README.md on where the user can find FunMessages.scala and modify it to their heart's content would be good :) (I must say I'm...

Hi @loongmxbt, thank you for your reply! I was just wondering what happened to your `/build` folder (containing `dev-server.js`, `webpack.base.conf.js`, `webpack.dev.conf.js`... etc.) that got generated with the `vue init webpack...

Same issue here, coming from my categorical string-valued variables. We got around this with a hack, to temporarily integer-encode these string-valued categories (with `sklearn.preprocessing.OrdinalEncoder`) just so the `x` could pass...