Internal change

PiperOrigin-RevId: 423199224

Author: saberkun

official/nlp/data/pretrain_dynamic_dataloader.py

@@ -79,17 +79,29 @@ def __init__(self, params):
   def _decode(self, record: tf.Tensor):
     """Decodes a serialized tf.Example."""
     name_to_features = {
-        'input_ids': tf.io.VarLenFeature(tf.int64),
         'input_mask': tf.io.VarLenFeature(tf.int64),
-        'segment_ids': tf.io.VarLenFeature(tf.int64),
         'masked_lm_positions': tf.io.VarLenFeature(tf.int64),
         'masked_lm_ids': tf.io.VarLenFeature(tf.int64),
         'masked_lm_weights': tf.io.VarLenFeature(tf.float32),
     }
+    if self._params.use_v2_feature_names:
+      input_ids_key = 'input_word_ids'
+      segment_key = 'input_type_ids'
+      name_to_features.update({
+          input_ids_key: tf.io.VarLenFeature(tf.int64),
+          segment_key: tf.io.VarLenFeature(tf.int64),
+      })
+    else:
+      input_ids_key = 'input_ids'
+      segment_key = 'segment_ids'
+      name_to_features.update({
+          input_ids_key: tf.io.VarLenFeature(tf.int64),
+          segment_key: tf.io.VarLenFeature(tf.int64),
+      })
     if self._use_next_sentence_label:
       name_to_features['next_sentence_labels'] = tf.io.FixedLenFeature([1],
                                                                        tf.int64)
-    dynamic_keys = ['input_ids', 'input_mask', 'segment_ids']
+    dynamic_keys = [input_ids_key, 'input_mask', segment_key]
     if self._use_position_id:
       name_to_features['position_ids'] = tf.io.VarLenFeature(tf.int64)
       dynamic_keys.append('position_ids')
@@ -102,7 +114,7 @@ def _decode(self, record: tf.Tensor):
     # sequence length dimension.
     # Pad before the first non pad from the back should not be removed.
     mask = tf.math.greater(
-        tf.math.cumsum(example['input_ids'], reverse=True), 0)
+        tf.math.cumsum(example[input_ids_key], reverse=True), 0)
     for key in dynamic_keys:
       example[key] = tf.boolean_mask(example[key], mask)