clinfo / kgcn Goto Github PK

Hello,

I was trying to run these nodes through VirtualBox but it does not work at all,
So there would a possibility to for you to create a specific workflows for windows in the future?

Thanks

Hi,
I am curious about the GCNlearner node,
in model.py variable default is "sample_chem.singletask.solubility.model"
i think that is only for example model not fit with other situation problem?
does it has any other model i can change on model.py variable?

thank you very much

ありがとう!

As a suggestion and for others potentially falling into this trap:

The requirements list tensorflow (>1.12). However it should also be added <2.0 as tf2 is incompatible.
1.13 seems to work (example notebook) also on Windows using anaconda tensorflow-gpu=1.13.

This feature aims to provide neural network optimization especially for multimodal graph networks.

Reference

• https://arxiv.org/abs/1502.05700
• https://arxiv.org/abs/1905.13403
• https://arxiv.org/abs/1905.06159

Hello, I work for KNIME (https://www.linkedin.com/in/victorapalacios/) and I'm interested in learning if the workflow used here can be uploaded to the KNIME Hub. We at KNIME are currently working on a PyTorch implementation of GCNs, so we would be very interested in seeing your framework. I tried to use your links for the workflows, but both seem to no longer work.

dependencies

• tensorflow-gpu == 1.15.0
• tensorflow_federated == 0.4.0

Ref: https://github.com/tensorflow/federated/
Examples: https://github.com/tensorflow/federated/blob/v0.4.0/docs/tutorials/federated_learning_for_image_classification.ipynb