Use tf.function for list column operations (#89)

Author: edknv

tests/unit/dataloader/test_tf_dataloader.py

@@ -94,25 +94,36 @@ def test_nested_list():
     schema = ds.schema
     schema["label"] = schema["label"].with_tags([Tags.TARGET])
     ds.schema = schema
-    train_dataset = tf_dataloader.Loader(
+    loader = tf_dataloader.Loader(
         ds,
         batch_size=batch_size,
         shuffle=False,
     )
 
-    batch = next(train_dataset)
+    batch = next(loader)
+
     # [[1,2,3],[3,1],[...],[]]
-    nested_data_col = tf.RaggedTensor.from_row_lengths(
-        batch[0]["data"][0][:, 0], tf.cast(batch[0]["data"][1][:, 0], tf.int32)
-    ).to_tensor()
+    @tf.function
+    def _ragged_for_nested_data_col():
+        nested_data_col = tf.RaggedTensor.from_row_lengths(
+            batch[0]["data"][0][:, 0], tf.cast(batch[0]["data"][1][:, 0], tf.int32)
+        ).to_tensor()
+        return nested_data_col
+
+    nested_data_col = _ragged_for_nested_data_col()
     true_data_col = tf.reshape(
-        tf.ragged.constant(df.iloc[:batch_size, 0].tolist()).to_tensor(),
-        [batch_size, -1],
+        tf.ragged.constant(df.iloc[:batch_size, 0].tolist()).to_tensor(), [batch_size, -1]
     )
+
     # [1,2,3]
-    multihot_data2_col = tf.RaggedTensor.from_row_lengths(
-        batch[0]["data2"][0][:, 0], tf.cast(batch[0]["data2"][1][:, 0], tf.int32)
-    ).to_tensor()
+    @tf.function
+    def _ragged_for_multihot_data_col():
+        multihot_data2_col = tf.RaggedTensor.from_row_lengths(
+            batch[0]["data2"][0][:, 0], tf.cast(batch[0]["data2"][1][:, 0], tf.int32)
+        ).to_tensor()
+        return multihot_data2_col
+
+    multihot_data2_col = _ragged_for_multihot_data_col()
     true_data2_col = tf.reshape(
         tf.ragged.constant(df.iloc[:batch_size, 1].tolist()).to_tensor(),
         [batch_size, -1],