make training input handling in keras fit case the same as CTL case

official/recommendation/ncf_keras_main.py

@@ -93,8 +93,6 @@ def _get_train_and_eval_data(producer, params):
   train_input_fn = producer.make_input_fn(is_training=True)
   train_input_dataset = train_input_fn(params).map(
       preprocess_train_input)
-  if not params["keras_use_ctl"]:
-    train_input_dataset = train_input_dataset.repeat(FLAGS.train_epochs)
 
   def preprocess_eval_input(features):
     """Pre-process the eval data.
@@ -286,8 +284,7 @@ def run_ncf(_):
   eval_input_dataset = eval_input_dataset.batch(batches_per_step)
 
   time_callback = keras_utils.TimeHistory(batch_size, FLAGS.log_steps)
-  per_epoch_callback = IncrementEpochCallback(producer)
-  callbacks = [per_epoch_callback, time_callback]
+  callbacks = [time_callback]
 
   if FLAGS.early_stopping:
     early_stopping_callback = CustomEarlyStopping(
@@ -388,7 +385,6 @@ def run_ncf(_):
       keras_model.compile(optimizer=optimizer)
 
       history = keras_model.fit(train_input_dataset,
-                                steps_per_epoch=num_train_steps,
                                 epochs=FLAGS.train_epochs,
                                 callbacks=callbacks,
                                 validation_data=eval_input_dataset,