Merge pull request #8720 from helinwang/vgg16

Fix vgg16 fc layer size: 4096 rather than 512

benchmark/cluster/vgg16/vgg16_fluid.py

@@ -106,10 +106,10 @@ def vgg16_bn_drop(input):
     conv5 = conv_block(conv4, 512, 3, [0.4, 0.4, 0])
 
     drop = fluid.layers.dropout(x=conv5, dropout_prob=0.5)
-    fc1 = fluid.layers.fc(input=drop, size=512, act=None)
+    fc1 = fluid.layers.fc(input=drop, size=4096, act=None)
     bn = fluid.layers.batch_norm(input=fc1, act='relu')
     drop2 = fluid.layers.dropout(x=bn, dropout_prob=0.5)
-    fc2 = fluid.layers.fc(input=drop2, size=512, act=None)
+    fc2 = fluid.layers.fc(input=drop2, size=4096, act=None)
     return fc2
 
 

python/paddle/fluid/tests/book/test_image_classification.py

@@ -86,10 +86,10 @@ def vgg16_bn_drop(input):
     conv5 = conv_block(conv4, 512, 3, [0.4, 0.4, 0])
 
     drop = fluid.layers.dropout(x=conv5, dropout_prob=0.5)
-    fc1 = fluid.layers.fc(input=drop, size=512, act=None)
+    fc1 = fluid.layers.fc(input=drop, size=4096, act=None)
     bn = fluid.layers.batch_norm(input=fc1, act='relu')
     drop2 = fluid.layers.dropout(x=bn, dropout_prob=0.5)
-    fc2 = fluid.layers.fc(input=drop2, size=512, act=None)
+    fc2 = fluid.layers.fc(input=drop2, size=4096, act=None)
     return fc2
 
 

python/paddle/fluid/tests/book_memory_optimization/test_memopt_image_classification_train.py

@@ -89,10 +89,10 @@ def vgg16_bn_drop(input):
     conv5 = conv_block(conv4, 512, 3, [0.4, 0.4, 0])
 
     drop = fluid.layers.dropout(x=conv5, dropout_prob=0.5)
-    fc1 = fluid.layers.fc(input=drop, size=512, act=None)
+    fc1 = fluid.layers.fc(input=drop, size=4096, act=None)
     bn = fluid.layers.batch_norm(input=fc1, act='relu')
     drop2 = fluid.layers.dropout(x=bn, dropout_prob=0.5)
-    fc2 = fluid.layers.fc(input=drop2, size=512, act=None)
+    fc2 = fluid.layers.fc(input=drop2, size=4096, act=None)
     return fc2