Reverting changes on run_pretrainng.py

TensorFlow/LanguageModeling/BERT/run_pretraining.py

@@ -72,7 +72,7 @@ flags.DEFINE_integer("num_train_steps", 100000, "Number of training steps.")
 
 flags.DEFINE_integer("num_warmup_steps", 10000, "Number of warmup steps.")
 
-flags.DEFINE_integer("save_checkpoint_steps", 1000,
+flags.DEFINE_integer("save_checkpoints_steps", 1000,
                      "How often to save the model checkpoint.")
 
 flags.DEFINE_integer("iterations_per_loop", 1000,
@@ -477,7 +477,7 @@ def main(_):
   run_config = tf.estimator.RunConfig(
       model_dir=FLAGS.output_dir,
       session_config=config,
-      save_checkpoints_steps=FLAGS.save_checkpoint_steps if not FLAGS.horovod or hvd.rank() == 0 else None,
+      save_checkpoints_steps=FLAGS.save_checkpoints_steps if not FLAGS.horovod or hvd.rank() == 0 else None,
       # This variable controls how often estimator reports examples/sec.
       # Default value is every 100 steps.
       # When --report_loss is True, we set to very large value to prevent