diff --git a/deepspeech/__init__.py b/deepspeech/__init__.py
index 90ab2223..493f10a6 100644
--- a/deepspeech/__init__.py
+++ b/deepspeech/__init__.py
@@ -362,19 +362,11 @@ def ctc_loss(logits,
              label_lengths,
              blank=0,
              reduction='mean',
-             norm_by_times=True,
-             norm_by_batchsize=False,
-             norm_by_total_logits_len=False):
+             norm_by_times=True):
     #logger.info("my ctc loss with norm by times")
     ## https://github.com/PaddlePaddle/Paddle/blob/f5ca2db2cc/paddle/fluid/operators/warpctc_op.h#L403
-    loss_out = paddle.fluid.layers.warpctc(
-        logits,
-        labels,
-        blank,
-        norm_by_times,
-        input_lengths,
-        label_lengths,
-        norm_by_batchsize, )
+    loss_out = paddle.fluid.layers.warpctc(logits, labels, blank, norm_by_times,
+                                           input_lengths, label_lengths)
 
     loss_out = paddle.fluid.layers.squeeze(loss_out, [-1])
     assert reduction in ['mean', 'sum', 'none']
diff --git a/deepspeech/models/ds2/deepspeech2.py b/deepspeech/models/ds2/deepspeech2.py
index a2aa31f7..63327a8c 100644
--- a/deepspeech/models/ds2/deepspeech2.py
+++ b/deepspeech/models/ds2/deepspeech2.py
@@ -219,10 +219,10 @@ class DeepSpeech2Model(nn.Layer):
             The model built from pretrained result.
         """
         model = cls(
-            #feat_size=dataloader.collate_fn.feature_size,
-            feat_size=dataloader.dataset.feature_size,
-            #dict_size=dataloader.collate_fn.vocab_size,
-            dict_size=dataloader.dataset.vocab_size,
+            feat_size=dataloader.collate_fn.feature_size,
+            #feat_size=dataloader.dataset.feature_size,
+            dict_size=dataloader.collate_fn.vocab_size,
+            #dict_size=dataloader.dataset.vocab_size,
             num_conv_layers=config.model.num_conv_layers,
             num_rnn_layers=config.model.num_rnn_layers,
             rnn_size=config.model.rnn_layer_size,