Clean up keras_resnet_model and rename it to resnet56

official/resnet/keras/keras_cifar_main.py

@@ -28,7 +28,7 @@ import tensorflow as tf  # pylint: disable=g-bad-import-order
 from official.resnet import cifar10_main as cifar_main
 from official.resnet import resnet_run_loop
 from official.resnet.keras import keras_common
-from official.resnet.keras import keras_resnet_model
+from official.resnet.keras import resnet56
 from official.utils.flags import core as flags_core
 from official.utils.logs import logger
 from official.utils.misc import distribution_utils
@@ -167,10 +167,9 @@ def run_cifar_with_keras(flags_obj):
   opt, loss, accuracy = keras_common.get_optimizer_loss_and_metrics()
   strategy = keras_common.get_dist_strategy()
 
-  model = keras_resnet_model.ResNet56(input_shape=(32, 32, 3),
-                                      include_top=True,
-                                      classes=cifar_main._NUM_CLASSES,
-                                      weights=None)
+  model = resnet56.ResNet56(input_shape=(32, 32, 3),
+                                      classes=cifar_main._NUM_CLASSES)
+
   model.compile(loss=loss,
                 optimizer=opt,
                 metrics=[accuracy],