add

gcn/gcn.py

@@ -53,13 +53,11 @@ def build(self, input_shapes):
     def call(self, inputs, training=None, **kwargs):
         features, A = inputs
         features = self.dropout(features, training=training)
-        output = tf.matmul(tf.sparse_tensor_dense_matmul(
-            A, features), self.kernel)
-        if self.bias:
+        output = tf.matmul(tf.matmul(A, features), self.kernel)
+        if self.use_bias:
             output += self.bias
         act = self.activation(output)
 
-        act._uses_learning_phase = features._uses_learning_phase
         return act
 
     def get_config(self):
@@ -78,7 +76,7 @@ def get_config(self):
 
 def GCN(adj_dim, feature_dim, n_hidden, num_class, num_layers=2, activation=tf.nn.relu, dropout_rate=0.5, l2_reg=0,
         feature_less=True, ):
-    Adj = Input(shape=(None,), sparse=True)
+    Adj = Input(shape=(None,), sparse=False)
     if feature_less:
         X_in = Input(shape=(1,), )
 
@@ -100,4 +98,8 @@ def GCN(adj_dim, feature_dim, n_hidden, num_class, num_layers=2, activation=tf.n
     output = h
     model = Model(inputs=[X_in, Adj], outputs=output)
 
+    model.__setattr__("embedding", output)
+    model.__setattr__("adj_input", Adj)
+    model.__setattr__("feature_input", X_in)
+
     return model

gcn/main.py

@@ -0,0 +1,53 @@
+# -*- coding:utf-8 -*-
+# @Time : 2021/9/11 10:32 下午
+# @Author : huichuan LI
+# @File : main.py
+# @Software: PyCharm
+
+import scipy.sparse as sp
+import numpy as np
+from gcn import GCN
+from tensorflow.keras.callbacks import ModelCheckpoint
+from tensorflow.keras.layers import Lambda
+from tensorflow.keras.models import Model
+from tensorflow.keras.optimizers import Adam
+from utils import load_data
+import tensorflow
+
+if __name__ == "__main__":
+    # Read data
+
+    FEATURE_LESS = False
+
+    adj, features, y_train, y_val, y_test, train_mask, val_mask, test_mask = load_data(path="../Graph/data/cora/")
+
+    if FEATURE_LESS:
+        X = np.arange(adj.shape[-1])
+        feature_dim = adj.shape[-1]
+    else:
+        X = features
+        feature_dim = X.shape[-1]
+
+    model_input = [X.toarray(), adj.toarray()]
+
+    model = GCN(adj.shape[-1], feature_dim, 16, y_train.shape[1], dropout_rate=0.5, l2_reg=2.5e-4,
+                feature_less=FEATURE_LESS, )
+
+    model.compile(optimizer=Adam(0.01), loss='categorical_crossentropy',
+                  weighted_metrics=['categorical_crossentropy', 'acc', tensorflow.keras.metrics.AUC(name='auc'),
+                                    ])
+
+    NB_EPOCH = 200
+    PATIENCE = 200  # early stopping patience
+
+    val_data = (model_input, y_val, val_mask)
+    # train
+    print("start training")
+    model.fit(model_input, y_train, sample_weight=train_mask, validation_data=val_data,
+              batch_size=adj.shape[0], epochs=NB_EPOCH, shuffle=False, verbose=2)
+
+    model_input = [adj.toarray(), features.toarray()]
+    user_embedding_model = Model(inputs=[model.adj_input, model.feature_input], outputs=model.embedding)
+
+    user_embs = user_embedding_model.predict(model_input, batch_size=adj.shape[0])
+    print(user_embs.shape)