tf-models-nightly 2.17.0.dev20240712__py2.py3-none-any.whl → 2.17.0.dev20240714__py2.py3-none-any.whl

official/recommendation/ranking/configs/config.py

@@ -74,6 +74,7 @@ class DataConfig(hyperparams.Config):
   cycle_length: int = 10
   sharding: bool = True
   num_shards_per_host: int = 8
+  use_cached_data: bool = False
 
 
 @dataclasses.dataclass

official/recommendation/ranking/data/data_pipeline.py

@@ -38,12 +38,14 @@ class CriteoTsvReader:
                params: config.DataConfig,
                num_dense_features: int,
                vocab_sizes: List[int],
-               use_synthetic_data: bool = False):
+               use_synthetic_data: bool = False,
+               use_cached_data: bool = False):
     self._file_pattern = file_pattern
     self._params = params
     self._num_dense_features = num_dense_features
     self._vocab_sizes = vocab_sizes
     self._use_synthetic_data = use_synthetic_data
+    self._use_cached_data = use_cached_data
 
   def __call__(self, ctx: tf.distribute.InputContext) -> tf.data.Dataset:
     params = self._params
@@ -117,6 +119,8 @@ class CriteoTsvReader:
         num_parallel_calls=tf.data.experimental.AUTOTUNE)
 
     dataset = dataset.prefetch(tf.data.experimental.AUTOTUNE)
+    if self._use_cached_data:
+      dataset = dataset.take(1).cache().repeat()
 
     return dataset
 
@@ -173,6 +177,9 @@ class CriteoTsvReader:
     if params.is_training:
       dataset = dataset.repeat()
 
+    if self._use_cached_data:
+      dataset = dataset.take(1).cache().repeat()
+
     return dataset.batch(batch_size, drop_remainder=True)
 
 

official/recommendation/ranking/data/data_pipeline_multi_hot.py

@@ -45,13 +45,15 @@ class CriteoTsvReaderMultiHot:
                num_dense_features: int,
                vocab_sizes: List[int],
                multi_hot_sizes: List[int],
-               use_synthetic_data: bool = False):
+               use_synthetic_data: bool = False,
+               use_cached_data: bool = False):
     self._file_pattern = file_pattern
     self._params = params
     self._num_dense_features = num_dense_features
     self._vocab_sizes = vocab_sizes
     self._use_synthetic_data = use_synthetic_data
     self._multi_hot_sizes = multi_hot_sizes
+    self._use_cached_data = use_cached_data
 
   def __call__(self, ctx: tf.distribute.InputContext) -> tf.data.Dataset:
     params = self._params
@@ -144,6 +146,8 @@ class CriteoTsvReaderMultiHot:
         num_parallel_calls=tf.data.experimental.AUTOTUNE)
 
     dataset = dataset.prefetch(tf.data.experimental.AUTOTUNE)
+    if self._use_cached_data:
+      dataset = dataset.take(1).cache().repeat()
 
     return dataset
 
@@ -215,12 +219,14 @@ class CriteoTFRecordReader(object):
                params: config.DataConfig,
                num_dense_features: int,
                vocab_sizes: List[int],
-               multi_hot_sizes: List[int],):
+               multi_hot_sizes: List[int],
+               use_cached_data: bool = False):
     self._file_pattern = file_pattern
     self._params = params
     self._num_dense_features = num_dense_features
     self._vocab_sizes = vocab_sizes
     self._multi_hot_sizes = multi_hot_sizes
+    self._use_cached_data = use_cached_data
 
     self.label_features = 'label'
     self.dense_features = ['dense-feature-%d' % x for x in range(1, 14)]
@@ -307,6 +313,8 @@ class CriteoTFRecordReader(object):
         num_parallel_calls=tf.data.experimental.AUTOTUNE,
     )
     dataset = dataset.prefetch(buffer_size=tf.data.experimental.AUTOTUNE)
+    if self._use_cached_data:
+      dataset = dataset.take(1).cache().repeat()
 
     return dataset
 

official/recommendation/ranking/data/data_pipeline_multi_hot_test.py

@@ -23,9 +23,13 @@ from official.recommendation.ranking.data import data_pipeline_multi_hot
 
 class DataPipelineTest(parameterized.TestCase, tf.test.TestCase):
 
-  @parameterized.named_parameters(('Train', True),
-                                  ('Eval', False))
-  def testSyntheticDataPipeline(self, is_training):
+  @parameterized.named_parameters(
+      ('TrainCached', True, True),
+      ('EvalNotCached', False, False),
+      ('TrainNotCached', True, False),
+      ('EvalCached', False, True),
+  )
+  def testSyntheticDataPipeline(self, is_training, use_cached_data):
     task = config.Task(
         model=config.ModelConfig(
             embedding_dim=4,
@@ -39,8 +43,10 @@ class DataPipelineTest(parameterized.TestCase, tf.test.TestCase):
             dcn_low_rank_dim=64,
             bottom_mlp=[64, 32, 4],
             top_mlp=[64, 32, 1]),
-        train_data=config.DataConfig(global_batch_size=16),
-        validation_data=config.DataConfig(global_batch_size=16),
+        train_data=config.DataConfig(global_batch_size=16,
+                                     use_cached_data=use_cached_data),
+        validation_data=config.DataConfig(global_batch_size=16,
+                                          use_cached_data=use_cached_data),
         use_synthetic_data=True)
 
     num_dense_features = task.model.num_dense_features

official/recommendation/ranking/data/data_pipeline_test.py

@@ -23,19 +23,29 @@ from official.recommendation.ranking.data import data_pipeline
 
 class DataPipelineTest(parameterized.TestCase, tf.test.TestCase):
 
-  @parameterized.named_parameters(('Train', True),
-                                  ('Eval', False))
-  def testSyntheticDataPipeline(self, is_training):
+  @parameterized.named_parameters(
+      ('TrainCached', True, True),
+      ('EvalNotCached', False, False),
+      ('TrainNotCached', True, False),
+      ('EvalCached', False, True),
+  )
+  def testSyntheticDataPipeline(self, is_training, use_cached_data):
     task = config.Task(
         model=config.ModelConfig(
             embedding_dim=4,
             num_dense_features=8,
             vocab_sizes=[40, 12, 11, 13, 2, 5],
             bottom_mlp=[64, 32, 4],
-            top_mlp=[64, 32, 1]),
-        train_data=config.DataConfig(global_batch_size=16),
-        validation_data=config.DataConfig(global_batch_size=16),
-        use_synthetic_data=True)
+            top_mlp=[64, 32, 1],
+        ),
+        train_data=config.DataConfig(
+            global_batch_size=16, use_cached_data=use_cached_data
+        ),
+        validation_data=config.DataConfig(
+            global_batch_size=16, use_cached_data=use_cached_data
+        ),
+        use_synthetic_data=True,
+    )
 
     num_dense_features = task.model.num_dense_features
     num_sparse_features = len(task.model.vocab_sizes)

{tf_models_nightly-2.17.0.dev20240712.dist-info → tf_models_nightly-2.17.0.dev20240714.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: tf-models-nightly
-Version: 2.17.0.dev20240712
+Version: 2.17.0.dev20240714
 Summary: TensorFlow Official Models
 Home-page: https://github.com/tensorflow/models
 Author: Google Inc.

{tf_models_nightly-2.17.0.dev20240712.dist-info → tf_models_nightly-2.17.0.dev20240714.dist-info}/RECORD

@@ -882,13 +882,13 @@ official/recommendation/ranking/task_test.py,sha256=vPN_5oq1tWF3r7GuRTuAppKSF_mP
 official/recommendation/ranking/train.py,sha256=_7zC2SVsOOOBN--If0XlWw4gtdaEeTT04PbCQRSzieo,6604
 official/recommendation/ranking/train_test.py,sha256=n96JN-gI5wj4cbZNJCwQ9XLViKt0ZAg9f9dtaTpAdV8,8756
 official/recommendation/ranking/configs/__init__.py,sha256=7oiypy0N82PDw9aSdcJBLVoGTd_oRSUOdvuJhMv4leQ,609
-official/recommendation/ranking/configs/config.py,sha256=03KupVFJ9QeslF81Ec7n4VyPn-bVvyZaEd10GQWomVA,14844
+official/recommendation/ranking/configs/config.py,sha256=_VICZEsRaLS4gbKeVv6T4RDnktqU9ughgx0s-GfgMgA,14876
 official/recommendation/ranking/configs/config_test.py,sha256=4_W0YUwJ2q0D8HdPdIvpJcAsaySjb_pDMWj26G7T1Ws,1474
 official/recommendation/ranking/data/__init__.py,sha256=7oiypy0N82PDw9aSdcJBLVoGTd_oRSUOdvuJhMv4leQ,609
-official/recommendation/ranking/data/data_pipeline.py,sha256=5GSF1naOl6GdqOVin7_Yx_hprmzaq0WAImHqDxysAoI,7329
-official/recommendation/ranking/data/data_pipeline_multi_hot.py,sha256=BOTItacIrH7f6ALYVx3vHkl_2ad4jKNpyQ4wFma7Hmc,11943
-official/recommendation/ranking/data/data_pipeline_multi_hot_test.py,sha256=arLjX742h_J6p2qqVH0vX86kl8WABWGIrU3hne5g-Dw,2702
-official/recommendation/ranking/data/data_pipeline_test.py,sha256=VRYo7WqURRkM3lbmfctvSZxyH1EzUfqwxR3sy2sZxdc,2345
+official/recommendation/ranking/data/data_pipeline.py,sha256=jtNyeAEZQFcFJaatTJhxYPcokN56OtETbF2fFqzWk8k,7578
+official/recommendation/ranking/data/data_pipeline_multi_hot.py,sha256=b8JGO0UyFOEjWn43NRXhEAs2wVBzEJuewwnwOGqTluM,12280
+official/recommendation/ranking/data/data_pipeline_multi_hot_test.py,sha256=-NKJ4-0wwSjveCY7sMP9lChDkN6lWH5VahYMHUa9z2A,2949
+official/recommendation/ranking/data/data_pipeline_test.py,sha256=34hKGxUBRjxyh9d2c7_eBhISpXI5F5PfgPfqBQoHM3Q,2573
 official/recommendation/uplift/__init__.py,sha256=_jZilTPWKu-MfMaz1IgBjEW6wqkK3FNZ1QAP4a8my3I,990
 official/recommendation/uplift/keras_test_case.py,sha256=gF5Z2FzXlKAvhuJDdj7PmFj3jsW_ZmUAv_F9Xokvs2M,6156
 official/recommendation/uplift/keys.py,sha256=7zkxkPIcXceIN5hWm4ATai4h8ymwCj85dU2r8-3XifM,1032
@@ -1212,9 +1212,9 @@ tensorflow_models/tensorflow_models_test.py,sha256=nc6A9K53OGqF25xN5St8EiWvdVbda
 tensorflow_models/nlp/__init__.py,sha256=4tA5Pf4qaFwT-fIFOpX7x7FHJpnyJT-5UgOeFYTyMlc,807
 tensorflow_models/uplift/__init__.py,sha256=mqfa55gweOdpKoaQyid4A_4u7xw__FcQeSIF0k_pYmI,999
 tensorflow_models/vision/__init__.py,sha256=zBorY_v5xva1uI-qxhZO3Qh-Dii-Suq6wEYh6hKHDfc,833
-tf_models_nightly-2.17.0.dev20240712.dist-info/AUTHORS,sha256=1dG3fXVu9jlo7bul8xuix5F5vOnczMk7_yWn4y70uw0,337
-tf_models_nightly-2.17.0.dev20240712.dist-info/LICENSE,sha256=WxeBS_DejPZQabxtfMOM_xn8qoZNJDQjrT7z2wG1I4U,11512
-tf_models_nightly-2.17.0.dev20240712.dist-info/METADATA,sha256=5IeomGvSkx2QmWWBiwbuYtPaUADvjMZk-UT6XdjCtOU,1432
-tf_models_nightly-2.17.0.dev20240712.dist-info/WHEEL,sha256=kGT74LWyRUZrL4VgLh6_g12IeVl_9u9ZVhadrgXZUEY,110
-tf_models_nightly-2.17.0.dev20240712.dist-info/top_level.txt,sha256=gum2FfO5R4cvjl2-QtP-S1aNmsvIZaFFT6VFzU0f4-g,33
-tf_models_nightly-2.17.0.dev20240712.dist-info/RECORD,,
+tf_models_nightly-2.17.0.dev20240714.dist-info/AUTHORS,sha256=1dG3fXVu9jlo7bul8xuix5F5vOnczMk7_yWn4y70uw0,337
+tf_models_nightly-2.17.0.dev20240714.dist-info/LICENSE,sha256=WxeBS_DejPZQabxtfMOM_xn8qoZNJDQjrT7z2wG1I4U,11512
+tf_models_nightly-2.17.0.dev20240714.dist-info/METADATA,sha256=YTQD7k71nGpFpJBjeSrX1NchtMazonQ6j-I5fzO3dcI,1432
+tf_models_nightly-2.17.0.dev20240714.dist-info/WHEEL,sha256=kGT74LWyRUZrL4VgLh6_g12IeVl_9u9ZVhadrgXZUEY,110
+tf_models_nightly-2.17.0.dev20240714.dist-info/top_level.txt,sha256=gum2FfO5R4cvjl2-QtP-S1aNmsvIZaFFT6VFzU0f4-g,33
+tf_models_nightly-2.17.0.dev20240714.dist-info/RECORD,,