Gnnexplainer graph-explanation fix.

[wsad1]

Bug

explainer = GNNExplainer(model, epochs=200)
explainer.explain_graph(x, edge_index) 
explainer.visualize_subgraph(node_idx=-1,edge_index, edge_mask, y=torch.tensor(4))

Running above code throws IndexError: too many indices for tensor of dimension 0 at y = y[subset].to(torch.float) / y.max().item() inside visualize_subgraph.

Fix

y is set to None inside visualize_subgraph for graph explanation task. And all nodes will have the same color in this case.

Thanks @Luoyunsong for bringing this up.

[codecov-commenter]

Codecov Report

Merging #2615 (ca7b282) into master (e2abb69) will increase coverage by 0.28%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #2615      +/-   ##
==========================================
+ Coverage   83.62%   83.90%   +0.28%     
==========================================
  Files         205      205              
  Lines        9228     9229       +1     
==========================================
+ Hits         7717     7744      +27     
+ Misses       1511     1485      -26     

Impacted Files	Coverage Δ
torch_geometric/nn/models/gnn_explainer.py	94.33% <100.00%> (+16.49%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e2abb69...ca7b282. Read the comment docs.