Add tfhub finetuning tutorial.

Will upload these tfhub models to official tfhub soon.

Author: mingxingtan

efficientnetv2/README.md

@@ -1,7 +1,8 @@
 # EfficientNetV2
 
-[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.sandbox.google.com/github/google/automl/blob/master/efficientnetv2/tutorial.ipynb)
-[![Paper](http://img.shields.io/badge/Paper-arXiv.2104.00298-B3181B?logo=arXiv)](https://arxiv.org/abs/2104.00298)
+[Paper![Paper](http://img.shields.io/badge/Paper-arXiv.2104.00298-B3181B?logo=arXiv)](https://arxiv.org/abs/2104.00298)
+[Tutorial![Tutorial In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.sandbox.google.com/github/google/automl/blob/master/efficientnetv2/tutorial.ipynb)
+[TF-hub![TF-Hub In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.sandbox.google.com/github/google/automl/blob/master/efficientnetv2/tfhub.ipynb)
 
 
    - May13/2021: Initial code release for [EfficientNetV2 models](https://arxiv.org/abs/2104.00298): accepted to ICML'21.

efficientnetv2/effnetv2_model.py

@@ -579,13 +579,14 @@ def summary(self, input_shape=(224, 224, 3), **kargs):
     model = tf.keras.Model(inputs=[x], outputs=self.call(x, training=True))
     return model.summary()
 
-  def call(self, inputs, training, features_only=None):
+  def call(self, inputs, training, features_only=None, single_out=None):
     """Implementation of call().
 
     Args:
       inputs: input tensors.
       training: boolean, whether the model is constructed for training.
       features_only: build the base feature network only.
+      single_out: If true, only return the single output.
 
     Returns:
       output tensors.
@@ -628,6 +629,9 @@ def call(self, inputs, training, features_only=None):
       outputs = self._head(outputs, training)
       self.endpoints.update(self._head.endpoints)
 
+    if single_out:  # Use for building sequential models.
+      return outputs
+
     return [outputs] + list(
         filter(lambda endpoint: endpoint is not None, [
             self.endpoints.get('reduction_1'),