_forward_unimplemented() got an unexpected keyword argument 'x' while using GNNExplainer

[LeavesLi1015]

I was planning to use GNNExplainer to explain GCN model training on MIND_small dataset.

# GNNExplainer
from torch_geometric.nn import GNNExplainer
explainer = GNNExplainer(model, epochs=500, return_type='log_prob')
x, edge_index = data.x, data.edge_index
node_idx = 10
node_feat_mask, edge_mask = explainer.explain_node(node_idx, x, edge_index)
ax, G = explainer.visualize_subgraph(node_idx, edge_index, edge_mask, y=data.y)
plt.show()

The model was well trained.
But it returned TypeError: _forward_unimplemented() got an unexpected keyword argument 'x'.
By the way, my data looks like this: Data(x=[101278, 100], edge_index=[2, 468930], weighted=[468930]).
How do I fix this? Thx!

[rusty1s]

How is your model's forward function defined? It currently requires that it looks as follows:

def forward(self, x, edge_index, any_other_kwargs, ...)
    pass

[LeavesLi1015]

I figured it out! I rewrote forward function and it worked! Thx a lot!