Part 2/3 of the update of tf.keras to the Keras 2.2.4 API.

PiperOrigin-RevId: 216442569

14 files changed, 296 insertions, 65 deletions

diff --git a/tensorflow/python/keras/backend.py b/tensorflow/python/keras/backend.py
index 13f52fbae7..7509ef9c59 100644
--- a/tensorflow/python/keras/backend.py
+++ b/tensorflow/python/keras/backend.py
@@ -2338,7 +2338,8 @@ def permute_dimensions(x, pattern):
 
 
 @tf_export('keras.backend.resize_images')
-def resize_images(x, height_factor, width_factor, data_format):
+def resize_images(x, height_factor, width_factor, data_format,
+                  interpolation='nearest'):
   """Resizes the images contained in a 4D tensor.
 
   Arguments:
@@ -2346,40 +2347,55 @@ def resize_images(x, height_factor, width_factor, data_format):
       height_factor: Positive integer.
       width_factor: Positive integer.
       data_format: One of `"channels_first"`, `"channels_last"`.
+      interpolation: A string, one of `nearest` or `bilinear`.
 
   Returns:
       A tensor.
 
   Raises:
-      ValueError: if `data_format` is neither
-          `channels_last` or `channels_first`.
+      ValueError: in case of incorrect value for
+        `data_format` or `interpolation`.
   """
   if data_format == 'channels_first':
-    original_shape = int_shape(x)
-    new_shape = array_ops.shape(x)[2:]
-    new_shape *= constant_op.constant(
-        np.array([height_factor, width_factor]).astype('int32'))
+    rows, cols = 2, 3
+  elif data_format == 'channels_last':
+    rows, cols = 1, 2
+  else:
+    raise ValueError('Invalid `data_format` argument: %s' % (data_format,))
+
+  original_shape = int_shape(x)
+  new_shape = array_ops.shape(x)[rows:cols + 1]
+  new_shape *= constant_op.constant(
+      np.array([height_factor, width_factor], dtype='int32'))
+
+  if data_format == 'channels_first':
     x = permute_dimensions(x, [0, 2, 3, 1])
+  if interpolation == 'nearest':
     x = image_ops.resize_nearest_neighbor(x, new_shape)
+  elif interpolation == 'bilinear':
+    x = image_ops.resize_bilinear(x, new_shape)
+  else:
+    raise ValueError('interpolation should be one '
+                     'of "nearest" or "bilinear".')
+  if data_format == 'channels_first':
     x = permute_dimensions(x, [0, 3, 1, 2])
-    x.set_shape((None, None, original_shape[2] * height_factor
-                 if original_shape[2] is not None else None,
-                 original_shape[3] * width_factor
-                 if original_shape[3] is not None else None))
-    return x
-  elif data_format == 'channels_last':
-    original_shape = int_shape(x)
-    new_shape = array_ops.shape(x)[1:3]
-    new_shape *= constant_op.constant(
-        np.array([height_factor, width_factor]).astype('int32'))
-    x = image_ops.resize_nearest_neighbor(x, new_shape)
-    x.set_shape((None, original_shape[1] * height_factor
-                 if original_shape[1] is not None else None,
-                 original_shape[2] * width_factor
-                 if original_shape[2] is not None else None, None))
-    return x
+
+  if original_shape[rows] is None:
+    new_height = None
   else:
-    raise ValueError('Invalid data_format: ' + str(data_format))
+    new_height = original_shape[rows] * height_factor
+
+  if original_shape[cols] is None:
+    new_width = None
+  else:
+    new_width = original_shape[cols] * width_factor
+
+  if data_format == 'channels_first':
+    output_shape = (None, None, new_height, new_width)
+  else:
+    output_shape = (None, new_height, new_width, None)
+  x.set_shape(output_shape)
+  return x
 
 
 @tf_export('keras.backend.resize_volumes')
diff --git a/tensorflow/python/keras/callbacks.py b/tensorflow/python/keras/callbacks.py
index 3d6000f223..4c12c83a4c 100644
--- a/tensorflow/python/keras/callbacks.py
+++ b/tensorflow/python/keras/callbacks.py
@@ -24,6 +24,7 @@ from collections import Iterable
 from collections import OrderedDict
 import copy
 import csv
+import io
 import json
 import math
 import os
@@ -606,24 +607,28 @@ class EarlyStopping(Callback):
   """Stop training when a monitored quantity has stopped improving.
 
   Arguments:
-      monitor: quantity to be monitored.
-      min_delta: minimum change in the monitored quantity
+      monitor: Quantity to be monitored.
+      min_delta: Minimum change in the monitored quantity
           to qualify as an improvement, i.e. an absolute
           change of less than min_delta, will count as no
           improvement.
-      patience: number of epochs with no improvement
+      patience: Number of epochs with no improvement
           after which training will be stopped.
       verbose: verbosity mode.
-      mode: one of {auto, min, max}. In `min` mode,
+      mode: One of `{"auto", "min", "max"}`. In `min` mode,
           training will stop when the quantity
           monitored has stopped decreasing; in `max`
           mode it will stop when the quantity
           monitored has stopped increasing; in `auto`
           mode, the direction is automatically inferred
           from the name of the monitored quantity.
-      baseline: baseline value for the monitored quantity.
+      baseline: Baseline value for the monitored quantity.
           Training will stop if the model doesn't show improvement over the
           baseline.
+      restore_best_weights: Whether to restore model weights from
+          the epoch with the best value of the monitored quantity.
+          If False, the model weights obtained at the last step of
+          training are used.
   """
 
   def __init__(self,
@@ -632,7 +637,8 @@ class EarlyStopping(Callback):
                patience=0,
                verbose=0,
                mode='auto',
-               baseline=None):
+               baseline=None,
+               restore_best_weights=False):
     super(EarlyStopping, self).__init__()
 
     self.monitor = monitor
@@ -642,6 +648,8 @@ class EarlyStopping(Callback):
     self.min_delta = abs(min_delta)
     self.wait = 0
     self.stopped_epoch = 0
+    self.restore_best_weights = restore_best_weights
+    self.best_weights = None
 
     if mode not in ['auto', 'min', 'max']:
       logging.warning('EarlyStopping mode %s is unknown, '
@@ -673,25 +681,37 @@ class EarlyStopping(Callback):
       self.best = np.Inf if self.monitor_op == np.less else -np.Inf
 
   def on_epoch_end(self, epoch, logs=None):
-    current = logs.get(self.monitor)
+    current = self.get_monitor_value(logs)
     if current is None:
-      logging.warning('Early stopping conditioned on metric `%s` '
-                      'which is not available. Available metrics are: %s',
-                      self.monitor, ','.join(list(logs.keys())))
       return
     if self.monitor_op(current - self.min_delta, self.best):
       self.best = current
       self.wait = 0
+      if self.restore_best_weights:
+        self.best_weights = self.model.get_weights()
     else:
       self.wait += 1
       if self.wait >= self.patience:
         self.stopped_epoch = epoch
         self.model.stop_training = True
+        if self.restore_best_weights:
+          if self.verbose > 0:
+            print('Restoring model weights from the end of the best epoch.')
+          self.model.set_weights(self.best_weights)
 
   def on_train_end(self, logs=None):
     if self.stopped_epoch > 0 and self.verbose > 0:
       print('Epoch %05d: early stopping' % (self.stopped_epoch + 1))
 
+  def get_monitor_value(self, logs):
+    logs = logs or {}
+    monitor_value = logs.get(self.monitor)
+    if monitor_value is None:
+      logging.warning('Early stopping conditioned on metric `%s` '
+                      'which is not available. Available metrics are: %s',
+                      self.monitor, ','.join(list(logs.keys())))
+    return monitor_value
+
 
 @tf_export('keras.callbacks.RemoteMonitor')
 class RemoteMonitor(Callback):
@@ -839,6 +859,12 @@ class TensorBoard(Callback):
           `embeddings_layer_names`. Numpy array (if the model has a single
           input) or list of Numpy arrays (if the model has multiple inputs).
           Learn [more about embeddings](https://www.tensorflow.org/programmers_guide/embedding)
+      update_freq: `'batch'` or `'epoch'` or integer. When using `'batch'`,
+          writes the losses and metrics to TensorBoard after each batch.
+          The same applies for `'epoch'`. If using an integer, let's say `1000`,
+          the callback will write the metrics and losses to TensorBoard every
+          1000 samples. Note that writing too frequently to TensorBoard
+          can slow down your training.
 
   Raises:
       ValueError: If histogram_freq is set and no validation data is provided.
@@ -862,7 +888,8 @@ class TensorBoard(Callback):
                embeddings_freq=0,
                embeddings_layer_names=None,
                embeddings_metadata=None,
-               embeddings_data=None):
+               embeddings_data=None,
+               update_freq='epoch'):
     super(TensorBoard, self).__init__()
     self.log_dir = log_dir
     self.histogram_freq = histogram_freq
@@ -882,6 +909,12 @@ class TensorBoard(Callback):
     self.embeddings_layer_names = embeddings_layer_names
     self.embeddings_metadata = embeddings_metadata
     self.embeddings_data = embeddings_data
+    if update_freq == 'batch':
+      self.update_freq = 1
+    else:
+      self.update_freq = update_freq
+    self._samples_seen = 0
+    self._samples_seen_at_last_write = 0
 
   def _init_writer(self):
     """Sets file writer."""
@@ -1045,13 +1078,17 @@ class TensorBoard(Callback):
       # use v2 summary ops
       with self.writer.as_default(), summary_ops_v2.always_record_summaries():
         for name, value in logs.items():
-          summary_ops_v2.scalar(name, value.item(), step=step)
+          if isinstance(value, np.ndarray):
+            value = value.item()
+          summary_ops_v2.scalar(name, value, step=step)
     else:
       # use FileWriter from v1 summary
       for name, value in logs.items():
+        if isinstance(value, np.ndarray):
+          value = value.item()
         summary = tf_summary.Summary()
         summary_value = summary.value.add()
-        summary_value.simple_value = value.item()
+        summary_value.simple_value = value
         summary_value.tag = name
         self.writer.add_summary(summary, step)
     self.writer.flush()
@@ -1076,10 +1113,14 @@ class TensorBoard(Callback):
     """Writes scalar summaries for metrics on every training batch."""
     # Don't output batch_size and batch number as Tensorboard summaries
     logs = logs or {}
-    batch_logs = {('batch_' + k): v
-                  for k, v in logs.items()
-                  if k not in ['batch', 'size', 'num_steps']}
-    self._write_custom_summaries(self._total_batches_seen, batch_logs)
+    self._samples_seen += logs.get('size', 1)
+    samples_seen_since = self._samples_seen - self._samples_seen_at_last_write
+    if self.update_freq != 'epoch' and samples_seen_since >= self.update_freq:
+      batch_logs = {('batch_' + k): v
+                    for k, v in logs.items()
+                    if k not in ['batch', 'size', 'num_steps']}
+      self._write_custom_summaries(self._total_batches_seen, batch_logs)
+      self._samples_seen_at_last_write = self._samples_seen
     self._total_batches_seen += 1
 
   def on_epoch_begin(self, epoch, logs=None):
@@ -1103,7 +1144,11 @@ class TensorBoard(Callback):
     logs = {('epoch_' + k): v
             for k, v in logs.items()
             if k not in ['batch', 'size', 'num_steps']}
-    self._write_custom_summaries(epoch, logs)
+    if self.update_freq == 'epoch':
+      step = epoch
+    else:
+      step = self._samples_seen
+    self._write_custom_summaries(step, logs)
 
     # pop the histogram summary op after each epoch
     if self.histogram_freq:
@@ -1309,7 +1354,12 @@ class CSVLogger(Callback):
     self.writer = None
     self.keys = None
     self.append_header = True
-    self.file_flags = 'b' if six.PY2 and os.name == 'nt' else ''
+    if six.PY2:
+      self.file_flags = 'b'
+      self._open_args = {}
+    else:
+      self.file_flags = ''
+      self._open_args = {'newline': '\n'}
     super(CSVLogger, self).__init__()
 
   def on_train_begin(self, logs=None):
@@ -1317,9 +1367,12 @@ class CSVLogger(Callback):
       if os.path.exists(self.filename):
         with open(self.filename, 'r' + self.file_flags) as f:
           self.append_header = not bool(len(f.readline()))
-      self.csv_file = open(self.filename, 'a' + self.file_flags)
+      mode = 'a'
     else:
-      self.csv_file = open(self.filename, 'w' + self.file_flags)
+      mode = 'w'
+    self.csv_file = io.open(self.filename,
+                            mode + self.file_flags,
+                            **self._open_args)
 
   def on_epoch_end(self, epoch, logs=None):
     logs = logs or {}
@@ -1345,9 +1398,13 @@ class CSVLogger(Callback):
       class CustomDialect(csv.excel):
         delimiter = self.sep
 
+      fieldnames = ['epoch'] + self.keys
+      if six.PY2:
+        fieldnames = [unicode(x) for x in fieldnames]
+
       self.writer = csv.DictWriter(
           self.csv_file,
-          fieldnames=['epoch'] + self.keys,
+          fieldnames=fieldnames,
           dialect=CustomDialect)
       if self.append_header:
         self.writer.writeheader()
diff --git a/tensorflow/python/keras/callbacks_test.py b/tensorflow/python/keras/callbacks_test.py
index 467bc4cdc4..bb85347033 100644
--- a/tensorflow/python/keras/callbacks_test.py
+++ b/tensorflow/python/keras/callbacks_test.py
@@ -313,6 +313,42 @@ class KerasCallbacksTest(test.TestCase):
       hist = model.fit(data, labels, callbacks=[stopper], verbose=0, epochs=20)
       assert len(hist.epoch) >= patience
 
+  def test_EarlyStopping_final_weights_when_restoring_model_weights(self):
+
+    class DummyModel(object):
+
+      def __init__(self):
+        self.stop_training = False
+        self.weights = -1
+
+      def get_weights(self):
+        return self.weights
+
+      def set_weights(self, weights):
+        self.weights = weights
+
+      def set_weight_to_epoch(self, epoch):
+        self.weights = epoch
+
+    early_stop = keras.callbacks.EarlyStopping(monitor='val_loss',
+                                               patience=2,
+                                               restore_best_weights=True)
+    early_stop.model = DummyModel()
+    losses = [0.2, 0.15, 0.1, 0.11, 0.12]
+    # The best configuration is in the epoch 2 (loss = 0.1000).
+    epochs_trained = 0
+    early_stop.on_train_begin()
+    for epoch in range(len(losses)):
+      epochs_trained += 1
+      early_stop.model.set_weight_to_epoch(epoch=epoch)
+      early_stop.on_epoch_end(epoch, logs={'val_loss': losses[epoch]})
+      if early_stop.model.stop_training:
+        break
+    # The best configuration is in epoch 2 (loss = 0.1000),
+    # and while patience = 2, we're restoring the best weights,
+    # so we end up at the epoch with the best weights, i.e. epoch 2
+    self.assertEqual(early_stop.model.get_weights(), 2)
+
   def test_RemoteMonitor(self):
     if requests is None:
       return
@@ -534,11 +570,15 @@ class KerasCallbacksTest(test.TestCase):
           batch_size=BATCH_SIZE,
           validation_data=(x_test, y_test),
           callbacks=cbks,
-          epochs=1,
+          epochs=2,
           verbose=0)
 
       with open(filepath) as csvfile:
-        output = ' '.join(csvfile.readlines())
+        list_lines = csvfile.readlines()
+        for line in list_lines:
+          assert line.count(sep) == 4
+        assert len(list_lines) == 5
+        output = ' '.join(list_lines)
         assert len(re.findall('epoch', output)) == 1
 
       os.remove(filepath)
@@ -1115,11 +1155,11 @@ class KerasCallbacksTest(test.TestCase):
     temp_dir = self.get_temp_dir()
     self.addCleanup(shutil.rmtree, temp_dir, ignore_errors=True)
 
-    tb_cbk = keras.callbacks.TensorBoard(temp_dir)
+    tb_cbk = keras.callbacks.TensorBoard(temp_dir, update_freq='batch')
     tb_cbk.writer = FileWriterStub(temp_dir)
 
     for batch in range(5):
-      tb_cbk.on_batch_end(batch, {'acc': np.float32(batch)})
+      tb_cbk.on_batch_end(batch, {'acc': batch})
     self.assertEqual(tb_cbk.writer.batches_logged, [0, 1, 2, 3, 4])
     self.assertEqual(tb_cbk.writer.summary_values, [0., 1., 2., 3., 4.])
     self.assertEqual(tb_cbk.writer.summary_tags, ['batch_acc'] * 5)
@@ -1147,14 +1187,17 @@ class KerasCallbacksTest(test.TestCase):
     temp_dir = self.get_temp_dir()
     self.addCleanup(shutil.rmtree, temp_dir, ignore_errors=True)
 
-    tb_cbk = keras.callbacks.TensorBoard(temp_dir)
+    tb_cbk = keras.callbacks.TensorBoard(temp_dir, update_freq='batch')
     tb_cbk.writer = FileWriterStub(temp_dir)
 
-    tb_cbk.on_batch_end(0, {'acc': np.float32(5.0)})
-    tb_cbk.on_epoch_end(0, {'acc': np.float32(10.0)})
+    tb_cbk.on_batch_end(0, {'acc': 5.0})
     batch_step, batch_summary = tb_cbk.writer.batch_summary
     self.assertEqual(batch_step, 0)
     self.assertEqual(batch_summary.value[0].simple_value, 5.0)
+
+    tb_cbk = keras.callbacks.TensorBoard(temp_dir, update_freq='epoch')
+    tb_cbk.writer = FileWriterStub(temp_dir)
+    tb_cbk.on_epoch_end(0, {'acc': 10.0})
     epoch_step, epoch_summary = tb_cbk.writer.epoch_summary
     self.assertEqual(epoch_step, 0)
     self.assertEqual(epoch_summary.value[0].simple_value, 10.0)
@@ -1192,6 +1235,66 @@ class KerasCallbacksTest(test.TestCase):
 
     self.assertTrue(os.path.exists(temp_dir))
 
+  def test_TensorBoard_update_freq(self):
+
+    class FileWriterStub(object):
+
+      def __init__(self, logdir, graph=None):
+        self.logdir = logdir
+        self.graph = graph
+        self.batch_summaries = []
+        self.epoch_summaries = []
+
+      def add_summary(self, summary, step):
+        if 'batch_' in summary.value[0].tag:
+          self.batch_summaries.append((step, summary))
+        elif 'epoch_' in summary.value[0].tag:
+          self.epoch_summaries.append((step, summary))
+
+      def flush(self):
+        pass
+
+      def close(self):
+        pass
+
+    temp_dir = self.get_temp_dir()
+    self.addCleanup(shutil.rmtree, temp_dir, ignore_errors=True)
+
+    # Epoch mode
+    tb_cbk = keras.callbacks.TensorBoard(temp_dir, update_freq='epoch')
+    tb_cbk.writer = FileWriterStub(temp_dir)
+
+    tb_cbk.on_batch_end(0, {'acc': 5.0, 'size': 1})
+    self.assertEqual(tb_cbk.writer.batch_summaries, [])
+    tb_cbk.on_epoch_end(0, {'acc': 10.0, 'size': 1})
+    self.assertEqual(len(tb_cbk.writer.epoch_summaries), 1)
+
+    # Batch mode
+    tb_cbk = keras.callbacks.TensorBoard(temp_dir, update_freq='batch')
+    tb_cbk.writer = FileWriterStub(temp_dir)
+
+    tb_cbk.on_batch_end(0, {'acc': 5.0, 'size': 1})
+    self.assertEqual(len(tb_cbk.writer.batch_summaries), 1)
+    tb_cbk.on_batch_end(0, {'acc': 5.0, 'size': 1})
+    self.assertEqual(len(tb_cbk.writer.batch_summaries), 2)
+    self.assertFalse(tb_cbk.writer.epoch_summaries)
+
+    # Integer mode
+    tb_cbk = keras.callbacks.TensorBoard(temp_dir, update_freq=20)
+    tb_cbk.writer = FileWriterStub(temp_dir)
+
+    tb_cbk.on_batch_end(0, {'acc': 5.0, 'size': 10})
+    self.assertFalse(tb_cbk.writer.batch_summaries)
+    tb_cbk.on_batch_end(0, {'acc': 5.0, 'size': 10})
+    self.assertEqual(len(tb_cbk.writer.batch_summaries), 1)
+    tb_cbk.on_batch_end(0, {'acc': 5.0, 'size': 10})
+    self.assertEqual(len(tb_cbk.writer.batch_summaries), 1)
+    tb_cbk.on_batch_end(0, {'acc': 5.0, 'size': 10})
+    self.assertEqual(len(tb_cbk.writer.batch_summaries), 2)
+    tb_cbk.on_batch_end(0, {'acc': 10.0, 'size': 10})
+    self.assertEqual(len(tb_cbk.writer.batch_summaries), 2)
+    self.assertFalse(tb_cbk.writer.epoch_summaries)
+
   def test_RemoteMonitorWithJsonPayload(self):
     if requests is None:
       self.skipTest('`requests` required to run this test')
@@ -1226,6 +1329,7 @@ class KerasCallbacksTest(test.TestCase):
   def test_fit_generator_with_callback(self):
 
     class TestCallback(keras.callbacks.Callback):
+
       def set_model(self, model):
         # Check the model operations for the optimizer operations that
         # the _make_train_function adds under a named scope for the
diff --git a/tensorflow/python/keras/layers/convolutional.py b/tensorflow/python/keras/layers/convolutional.py
index 8f5872385c..58024677ee 100644
--- a/tensorflow/python/keras/layers/convolutional.py
+++ b/tensorflow/python/keras/layers/convolutional.py
@@ -1951,6 +1951,7 @@ class UpSampling2D(Layer):
           It defaults to the `image_data_format` value found in your
           Keras config file at `~/.keras/keras.json`.
           If you never set it, then it will be "channels_last".
+      interpolation: A string, one of `nearest` or `bilinear`.
 
   Input shape:
       4D tensor with shape:
@@ -1967,10 +1968,18 @@ class UpSampling2D(Layer):
           `(batch, channels, upsampled_rows, upsampled_cols)`
   """
 
-  def __init__(self, size=(2, 2), data_format=None, **kwargs):
+  def __init__(self,
+               size=(2, 2),
+               data_format=None,
+               interpolation='nearest',
+               **kwargs):
     super(UpSampling2D, self).__init__(**kwargs)
     self.data_format = conv_utils.normalize_data_format(data_format)
     self.size = conv_utils.normalize_tuple(size, 2, 'size')
+    if interpolation not in {'nearest', 'bilinear'}:
+      raise ValueError('`interpolation` argument should be one of `"nearest"` '
+                       'or `"bilinear"`.')
+    self.interpolation = interpolation
     self.input_spec = InputSpec(ndim=4)
 
   def compute_output_shape(self, input_shape):
@@ -1992,7 +2001,8 @@ class UpSampling2D(Layer):
 
   def call(self, inputs):
     return backend.resize_images(
-        inputs, self.size[0], self.size[1], self.data_format)
+        inputs, self.size[0], self.size[1], self.data_format,
+        interpolation=self.interpolation)
 
   def get_config(self):
     config = {'size': self.size, 'data_format': self.data_format}
diff --git a/tensorflow/python/keras/layers/convolutional_test.py b/tensorflow/python/keras/layers/convolutional_test.py
index f88d632ab5..bdc175b8b9 100644
--- a/tensorflow/python/keras/layers/convolutional_test.py
+++ b/tensorflow/python/keras/layers/convolutional_test.py
@@ -790,6 +790,42 @@ class UpSamplingTest(test.TestCase):
             np.testing.assert_allclose(np_output, expected_out)
 
   @tf_test_util.run_in_graph_and_eager_modes
+  def test_upsampling_2d_bilinear(self):
+    num_samples = 2
+    stack_size = 2
+    input_num_row = 11
+    input_num_col = 12
+    for data_format in ['channels_first', 'channels_last']:
+      if data_format == 'channels_first':
+        inputs = np.random.rand(num_samples, stack_size, input_num_row,
+                                input_num_col)
+      else:
+        inputs = np.random.rand(num_samples, input_num_row, input_num_col,
+                                stack_size)
+
+      testing_utils.layer_test(keras.layers.UpSampling2D,
+                               kwargs={'size': (2, 2),
+                                       'data_format': data_format,
+                                       'interpolation': 'bilinear'},
+                               input_shape=inputs.shape)
+
+      if not context.executing_eagerly():
+        for length_row in [2]:
+          for length_col in [2, 3]:
+            layer = keras.layers.UpSampling2D(
+                size=(length_row, length_col),
+                data_format=data_format)
+            layer.build(inputs.shape)
+            outputs = layer(keras.backend.variable(inputs))
+            np_output = keras.backend.eval(outputs)
+            if data_format == 'channels_first':
+              self.assertEqual(np_output.shape[2], length_row * input_num_row)
+              self.assertEqual(np_output.shape[3], length_col * input_num_col)
+            else:
+              self.assertEqual(np_output.shape[1], length_row * input_num_row)
+              self.assertEqual(np_output.shape[2], length_col * input_num_col)
+
+  @tf_test_util.run_in_graph_and_eager_modes
   def test_upsampling_3d(self):
     num_samples = 2
     stack_size = 2
diff --git a/tensorflow/python/kernel_tests/rnn_test.py b/tensorflow/python/kernel_tests/rnn_test.py
index 2f6963f6b8..907e1277a9 100644
--- a/tensorflow/python/kernel_tests/rnn_test.py
+++ b/tensorflow/python/kernel_tests/rnn_test.py
@@ -571,8 +571,8 @@ class RNNTest(test.TestCase):
       cell.set_weights(tf_weights)
       [tf_out, tf_state] = sess.run([tf_out, tf_state], {inputs: x_train})
 
-    self.assertAllClose(tf_out, k_out)
-    self.assertAllClose(tf_state, k_state)
+    self.assertAllClose(tf_out, k_out, atol=1e-5)
+    self.assertAllClose(tf_state, k_state, atol=1e-5)
 
   def testBasicLSTMCellInterchangeWithLSTMCell(self):
     with self.session(graph=ops_lib.Graph()) as sess:
diff --git a/tensorflow/tools/api/golden/v1/tensorflow.keras.backend.pbtxt b/tensorflow/tools/api/golden/v1/tensorflow.keras.backend.pbtxt
index 9feb7c09b8..5f0dfd7ae7 100644
--- a/tensorflow/tools/api/golden/v1/tensorflow.keras.backend.pbtxt
+++ b/tensorflow/tools/api/golden/v1/tensorflow.keras.backend.pbtxt
@@ -386,7 +386,7 @@ tf_module {
   }
   member_method {
     name: "resize_images"
-    argspec: "args=[\'x\', \'height_factor\', \'width_factor\', \'data_format\'], varargs=None, keywords=None, defaults=None"
+    argspec: "args=[\'x\', \'height_factor\', \'width_factor\', \'data_format\', \'interpolation\'], varargs=None, keywords=None, defaults=[\'nearest\'], "
   }
   member_method {
     name: "resize_volumes"
diff --git a/tensorflow/tools/api/golden/v1/tensorflow.keras.callbacks.-early-stopping.pbtxt b/tensorflow/tools/api/golden/v1/tensorflow.keras.callbacks.-early-stopping.pbtxt
index f71292856c..ed0f37647f 100644
--- a/tensorflow/tools/api/golden/v1/tensorflow.keras.callbacks.-early-stopping.pbtxt
+++ b/tensorflow/tools/api/golden/v1/tensorflow.keras.callbacks.-early-stopping.pbtxt
@@ -5,7 +5,11 @@ tf_class {
   is_instance: "<type \'object\'>"
   member_method {
     name: "__init__"
-    argspec: "args=[\'self\', \'monitor\', \'min_delta\', \'patience\', \'verbose\', \'mode\', \'baseline\'], varargs=None, keywords=None, defaults=[\'val_loss\', \'0\', \'0\', \'0\', \'auto\', \'None\'], "
+    argspec: "args=[\'self\', \'monitor\', \'min_delta\', \'patience\', \'verbose\', \'mode\', \'baseline\', \'restore_best_weights\'], varargs=None, keywords=None, defaults=[\'val_loss\', \'0\', \'0\', \'0\', \'auto\', \'None\', \'False\'], "
+  }
+  member_method {
+    name: "get_monitor_value"
+    argspec: "args=[\'self\', \'logs\'], varargs=None, keywords=None, defaults=None"
   }
   member_method {
     name: "on_batch_begin"
diff --git a/tensorflow/tools/api/golden/v1/tensorflow.keras.callbacks.-tensor-board.pbtxt b/tensorflow/tools/api/golden/v1/tensorflow.keras.callbacks.-tensor-board.pbtxt
index e58ba18c1c..e9d53b7225 100644
--- a/tensorflow/tools/api/golden/v1/tensorflow.keras.callbacks.-tensor-board.pbtxt
+++ b/tensorflow/tools/api/golden/v1/tensorflow.keras.callbacks.-tensor-board.pbtxt
@@ -5,7 +5,7 @@ tf_class {
   is_instance: "<type \'object\'>"
   member_method {
     name: "__init__"
-    argspec: "args=[\'self\', \'log_dir\', \'histogram_freq\', \'batch_size\', \'write_graph\', \'write_grads\', \'write_images\', \'embeddings_freq\', \'embeddings_layer_names\', \'embeddings_metadata\', \'embeddings_data\'], varargs=None, keywords=None, defaults=[\'./logs\', \'0\', \'32\', \'True\', \'False\', \'False\', \'0\', \'None\', \'None\', \'None\'], "
+    argspec: "args=[\'self\', \'log_dir\', \'histogram_freq\', \'batch_size\', \'write_graph\', \'write_grads\', \'write_images\', \'embeddings_freq\', \'embeddings_layer_names\', \'embeddings_metadata\', \'embeddings_data\', \'update_freq\'], varargs=None, keywords=None, defaults=[\'./logs\', \'0\', \'32\', \'True\', \'False\', \'False\', \'0\', \'None\', \'None\', \'None\', \'epoch\'], "
   }
   member_method {
     name: "on_batch_begin"
diff --git a/tensorflow/tools/api/golden/v1/tensorflow.keras.layers.-up-sampling2-d.pbtxt b/tensorflow/tools/api/golden/v1/tensorflow.keras.layers.-up-sampling2-d.pbtxt
index 40a56a0c94..b05e5ec84d 100644
--- a/tensorflow/tools/api/golden/v1/tensorflow.keras.layers.-up-sampling2-d.pbtxt
+++ b/tensorflow/tools/api/golden/v1/tensorflow.keras.layers.-up-sampling2-d.pbtxt
@@ -82,7 +82,7 @@ tf_class {
   }
   member_method {
     name: "__init__"
-    argspec: "args=[\'self\', \'size\', \'data_format\'], varargs=None, keywords=kwargs, defaults=[\'(2, 2)\', \'None\'], "
+    argspec: "args=[\'self\', \'size\', \'data_format\', \'interpolation\'], varargs=None, keywords=kwargs, defaults=[\'(2, 2)\', \'None\', \'nearest\'], "
   }
   member_method {
     name: "add_loss"
diff --git a/tensorflow/tools/api/golden/v2/tensorflow.keras.backend.pbtxt b/tensorflow/tools/api/golden/v2/tensorflow.keras.backend.pbtxt
index 9feb7c09b8..5f0dfd7ae7 100644
--- a/tensorflow/tools/api/golden/v2/tensorflow.keras.backend.pbtxt
+++ b/tensorflow/tools/api/golden/v2/tensorflow.keras.backend.pbtxt
@@ -386,7 +386,7 @@ tf_module {
   }
   member_method {
     name: "resize_images"
-    argspec: "args=[\'x\', \'height_factor\', \'width_factor\', \'data_format\'], varargs=None, keywords=None, defaults=None"
+    argspec: "args=[\'x\', \'height_factor\', \'width_factor\', \'data_format\', \'interpolation\'], varargs=None, keywords=None, defaults=[\'nearest\'], "
   }
   member_method {
     name: "resize_volumes"
diff --git a/tensorflow/tools/api/golden/v2/tensorflow.keras.callbacks.-early-stopping.pbtxt b/tensorflow/tools/api/golden/v2/tensorflow.keras.callbacks.-early-stopping.pbtxt
index f71292856c..ed0f37647f 100644
--- a/tensorflow/tools/api/golden/v2/tensorflow.keras.callbacks.-early-stopping.pbtxt
+++ b/tensorflow/tools/api/golden/v2/tensorflow.keras.callbacks.-early-stopping.pbtxt
@@ -5,7 +5,11 @@ tf_class {
   is_instance: "<type \'object\'>"
   member_method {
     name: "__init__"
-    argspec: "args=[\'self\', \'monitor\', \'min_delta\', \'patience\', \'verbose\', \'mode\', \'baseline\'], varargs=None, keywords=None, defaults=[\'val_loss\', \'0\', \'0\', \'0\', \'auto\', \'None\'], "
+    argspec: "args=[\'self\', \'monitor\', \'min_delta\', \'patience\', \'verbose\', \'mode\', \'baseline\', \'restore_best_weights\'], varargs=None, keywords=None, defaults=[\'val_loss\', \'0\', \'0\', \'0\', \'auto\', \'None\', \'False\'], "
+  }
+  member_method {
+    name: "get_monitor_value"
+    argspec: "args=[\'self\', \'logs\'], varargs=None, keywords=None, defaults=None"
   }
   member_method {
     name: "on_batch_begin"
diff --git a/tensorflow/tools/api/golden/v2/tensorflow.keras.callbacks.-tensor-board.pbtxt b/tensorflow/tools/api/golden/v2/tensorflow.keras.callbacks.-tensor-board.pbtxt
index e58ba18c1c..e9d53b7225 100644
--- a/tensorflow/tools/api/golden/v2/tensorflow.keras.callbacks.-tensor-board.pbtxt
+++ b/tensorflow/tools/api/golden/v2/tensorflow.keras.callbacks.-tensor-board.pbtxt
@@ -5,7 +5,7 @@ tf_class {
   is_instance: "<type \'object\'>"
   member_method {
     name: "__init__"
-    argspec: "args=[\'self\', \'log_dir\', \'histogram_freq\', \'batch_size\', \'write_graph\', \'write_grads\', \'write_images\', \'embeddings_freq\', \'embeddings_layer_names\', \'embeddings_metadata\', \'embeddings_data\'], varargs=None, keywords=None, defaults=[\'./logs\', \'0\', \'32\', \'True\', \'False\', \'False\', \'0\', \'None\', \'None\', \'None\'], "
+    argspec: "args=[\'self\', \'log_dir\', \'histogram_freq\', \'batch_size\', \'write_graph\', \'write_grads\', \'write_images\', \'embeddings_freq\', \'embeddings_layer_names\', \'embeddings_metadata\', \'embeddings_data\', \'update_freq\'], varargs=None, keywords=None, defaults=[\'./logs\', \'0\', \'32\', \'True\', \'False\', \'False\', \'0\', \'None\', \'None\', \'None\', \'epoch\'], "
   }
   member_method {
     name: "on_batch_begin"
diff --git a/tensorflow/tools/api/golden/v2/tensorflow.keras.layers.-up-sampling2-d.pbtxt b/tensorflow/tools/api/golden/v2/tensorflow.keras.layers.-up-sampling2-d.pbtxt
index 40a56a0c94..b05e5ec84d 100644
--- a/tensorflow/tools/api/golden/v2/tensorflow.keras.layers.-up-sampling2-d.pbtxt
+++ b/tensorflow/tools/api/golden/v2/tensorflow.keras.layers.-up-sampling2-d.pbtxt
@@ -82,7 +82,7 @@ tf_class {
   }
   member_method {
     name: "__init__"
-    argspec: "args=[\'self\', \'size\', \'data_format\'], varargs=None, keywords=kwargs, defaults=[\'(2, 2)\', \'None\'], "
+    argspec: "args=[\'self\', \'size\', \'data_format\', \'interpolation\'], varargs=None, keywords=kwargs, defaults=[\'(2, 2)\', \'None\', \'nearest\'], "
   }
   member_method {
     name: "add_loss"