Correctly setting the input to Train (#3853)

In the ResNetTrainActor class, the data are now exactly build using the Train flag for the cifar_input script.
2026-07-17 11:32:33 +08:00 · 2019-07-27 11:08:35 -07:00
parent a62c5f40f6
commit 10cbcced7e
1 changed files with 1 additions and 1 deletions
@@ -107,7 +107,7 @@ class ResNetTrainActor(object):
        with tf.device("/gpu:0" if num_gpus > 0 else "/cpu:0"):
            # Build the model.
            images, labels = cifar_input.build_input(data, hps.batch_size,
-                                                     dataset, False)
+                                                     dataset, True)
            self.model = resnet_model.ResNet(hps, images, labels, "train")
            self.model.build_graph()
            config = tf.ConfigProto(allow_soft_placement=True)