Prediction with a trained model

[serawork]

Got input shape error when trying to do prediction after the model is trained. What should be the input shape? I use model.predict([X, A]).

[va957]

Hello ,
So input shape should be list of Adjacency Matrix and X_input Features.
This is an example for predict function for model.

"# Load the new graph
g = nx.from_pandas_edgelist(train_df, source='question_body', target=targ)

Check if the new graph has fewer nodes

if g.number_of_nodes() < graph.number_of_nodes():
# Add dummy nodes to the new graph
num_dummy_nodes = graph.number_of_nodes() - g.number_of_nodes()
g.add_nodes_from(range(g.number_of_nodes(), g.number_of_nodes() + num_dummy_nodes))

elif g.number_of_nodes() > graph.number_of_nodes():
# Trim the new graph
g = g.subgraph(range(graph.number_of_nodes()))

Convert the graph to an adjacency matrix and preprocess it

Ad = nx.adjacency_matrix(g, nodelist=range(g.number_of_nodes()))
Ad = preprocess_adj(Ad)

Create input features for the model

feature_dim1 = Ad.shape[-1]
X1 = np.arange(Ad.shape[-1])
X_n1 = []
for i in range(feature_dim1):
X_n1.append(X1)
X1 = np.asarray(X_n1)

model_input1 = [X1, Ad]
pred = model.predict(model_input1)

"