Merge pull request #9070 from NivekNey:patch-1

PiperOrigin-RevId: 327072144

official/nlp/bert/run_classifier.py

@@ -263,6 +263,7 @@ def run_keras_compile_fit(model_dir,
 def get_predictions_and_labels(strategy,
                                trained_model,
                                eval_input_fn,
+                               is_regression=False,
                                return_probs=False):
   """Obtains predictions of trained model on evaluation data.
 
@@ -273,6 +274,7 @@ def get_predictions_and_labels(strategy,
     strategy: Distribution strategy.
     trained_model: Trained model with preloaded weights.
     eval_input_fn: Input function for evaluation data.
+    is_regression: Whether it is a regression task.
     return_probs: Whether to return probabilities of classes.
 
   Returns:
@@ -288,8 +290,11 @@ def get_predictions_and_labels(strategy,
       """Replicated predictions."""
       inputs, labels = inputs
       logits = trained_model(inputs, training=False)
-      probabilities = tf.nn.softmax(logits)
-      return probabilities, labels
+      if not is_regression:
+        probabilities = tf.nn.softmax(logits)
+        return probabilities, labels
+      else:
+        return logits, labels
 
     outputs, labels = strategy.run(_test_step_fn, args=(next(iterator),))
     # outputs: current batch logits as a tuple of shard logits
@@ -447,9 +452,10 @@ def custom_main(custom_callbacks=None, custom_metrics=None):
       include_sample_weights=include_sample_weights)
 
   if FLAGS.mode == 'predict':
+    num_labels = input_meta_data.get('num_labels', 1)
     with strategy.scope():
       classifier_model = bert_models.classifier_model(
-          bert_config, input_meta_data['num_labels'])[0]
+          bert_config, num_labels)[0]
       checkpoint = tf.train.Checkpoint(model=classifier_model)
       latest_checkpoint_file = (
           FLAGS.predict_checkpoint_path or
@@ -460,7 +466,11 @@ def custom_main(custom_callbacks=None, custom_metrics=None):
       checkpoint.restore(
           latest_checkpoint_file).assert_existing_objects_matched()
       preds, _ = get_predictions_and_labels(
-          strategy, classifier_model, eval_input_fn, return_probs=True)
+          strategy,
+          classifier_model,
+          eval_input_fn,
+          is_regression=(num_labels == 1),
+          return_probs=True)
     output_predict_file = os.path.join(FLAGS.model_dir, 'test_results.tsv')
     with tf.io.gfile.GFile(output_predict_file, 'w') as writer:
       logging.info('***** Predict results *****')