readme

Author: dragen1860

README.MD

@@ -1 +1,19 @@
 # Graph Convolution Network for TF2
+
+GCN implementation for paper: [Semi-Supervised Classification with Graph Convolutional Networks](https://arxiv.org/pdf/1609.02907.pdf)
+
+# Benchmark
+
+| dataset       | Citeseea | Cora | Pubmed | NELL |
+|---------------|----------|------|--------|------|
+| GCN(official) | 70.3     | 81.5 | 79.0   | 66.0 |
+| This repo.    |          | 81.8 |        |      |
+
+# HOWTO
+```
+python train.py
+```
+
+# Screenshot
+
+![](res/scree.png)

layers.py

@@ -124,9 +124,9 @@ def call(self, inputs, training=None):
         x, support_ = inputs
 
         # dropout
-        if training and self.is_sparse_inputs:
+        if training is not False and self.is_sparse_inputs:
             x = sparse_dropout(x, self.dropout, self.num_features_nonzero)
-        elif training:
+        elif training is not False:
             x = tf.nn.dropout(x, self.dropout)
 
 

models.py

@@ -96,7 +96,7 @@ def call(self, inputs, training=None):
         outputs = [x]
 
         for layer in self.layers:
-            hidden = layer((outputs[-1], support))
+            hidden = layer((outputs[-1], support), training)
             outputs.append(hidden)
         output = outputs[-1]
 

res/screen.png

@@ -63,6 +63,8 @@
 train_mask = tf.convert_to_tensor(train_mask)
 val_label = tf.convert_to_tensor(y_val)
 val_mask = tf.convert_to_tensor(val_mask)
+test_label = tf.convert_to_tensor(y_test)
+test_mask = tf.convert_to_tensor(test_mask)
 features = tf.SparseTensor(*features)
 support = [tf.cast(tf.SparseTensor(*support[0]), dtype=tf.float32)]
 num_features_nonzero = features.values.shape
@@ -71,7 +73,6 @@
 
 optimizer = optimizers.Adam(lr=1e-2)
 
-cost_val = []
 
 
 for epoch in range(args.epochs):
@@ -85,4 +86,11 @@
     _, val_acc = model((features, val_label, val_mask, support), training=False)
 
 
-    print(epoch, float(loss), float(acc), '\tval:', float(val_acc))
+    print(epoch, float(loss), float(acc), '\tval:', float(val_acc))
+
+
+
+test_loss, test_acc = model((features, test_label, test_mask, support), training=False)
+
+
+print('\ttest:', float(test_loss), float(test_acc))