Shivam Khandelwal

On running again, with ```CUDA_LAUNCH_BLOCKING=1``` I get this in the error stack trace: ``` /pytorch/aten/src/ATen/native/cuda/MultinomialKernel.cu:256: void at::native::::sampleMultinomialOnce(long *, long, int, scalar_t *, scalar_t *, int, int) [with scalar_t = float,...

I had to install `ipywidgets` and `widgetsnbextension`. I saw their documentations but found nothing about them being either JS or server extensions. @takluyver , is there any way to know...

Kindly provide the configurations!!

How to embed a new subgraph? For example, how to create the embeddings of a number of new nodes that are being added to the graph? BY providing all of...

Hi @sam-lev , if by disjoint graph, you mean vertex disjoint graph, which means that any nodes in this new graph are not same as any node in the previous...

> Unfortunately, doing this I still run into issues of placehodlers and variables learned during training being uninitialized or inapplicable. This is exactly the problem I am facing, with the...

How did you solve this? By using some fake nodes while training and then filling their feature vectors with relevant data and their adjacency with the relevant nodes?

Following, if ever there is a reply to this.