Open SaeedFarahani opened 6 years ago
The repo that you mentioned is a fork. Why shall installed a fork repository and not the original one?
We use "Deep_GCN" model that implement in this version.
We added gcn to the backend.
thanks a lot.
We added gcn to the backend.
Still, as described in #5, we should make sure we keep track of the software stack installed networkx
, joblib
, tensorflow
on top of the gcn
code, so others participants don't have to request it.
We can see the following error while using GCN:
absl.flags._exceptions.UnrecognizedFlagError: Unknown command line flag 'save-y-preds'
In case that you are passing flags be aware that the test submission is called on the backend with an extra argument:
ramp_test_submission --submission my_submission --save-y-preds
So if you are catching all arguments be aware that save-y-preds
is also there and you need to filter it.
add gcn package to standard environment
gcn https://github.com/parisots/gcn
and dependencies of this package
tensorflow (>0.12) networkx joblib