To avoid the error when use_program_cache is True in Transformer inference. (#2485)

To avoid the error when use_program_cache is True in Transformer infe…rence.

To avoid the data decode error.

PaddleNLP/neural_machine_translation/transformer/infer.py

@@ -281,11 +281,11 @@ def fast_infer(args):
                     feed=feed_dict_list[0]
                     if feed_dict_list is not None else None,
                     return_numpy=False,
-                    use_program_cache=True)
+                    use_program_cache=False)
             seq_ids_list, seq_scores_list = [seq_ids], [
                 seq_scores
-            ] if isinstance(
-                seq_ids, paddle.fluid.LoDTensor) else (seq_ids, seq_scores)
+            ] if isinstance(seq_ids,
+                            paddle.fluid.LoDTensor) else (seq_ids, seq_scores)
             for seq_ids, seq_scores in zip(seq_ids_list, seq_scores_list):
                 # How to parse the results:
                 #   Suppose the lod of seq_ids is:

PaddleNLP/neural_machine_translation/transformer/reader.py

@@ -266,7 +266,7 @@ class DataReader(object):
                 with open(fpath, "rb") as f:
                     for line in f:
                         if six.PY3:
-                            line = line.decode()
+                            line = line.decode("utf8", errors="ignore")
                         fields = line.strip("\n").split(self._field_delimiter)
                         if (not self._only_src and len(fields) == 2) or (
                                 self._only_src and len(fields) == 1):

PaddleNLP/unarchived/neural_machine_translation/transformer/infer.py

@@ -280,7 +280,7 @@ def fast_infer(args):
                     feed=feed_dict_list[0]
                     if feed_dict_list is not None else None,
                     return_numpy=False,
-                    use_program_cache=True)
+                    use_program_cache=False)
             seq_ids_list, seq_scores_list = [
                 seq_ids
             ], [seq_scores] if isinstance(

PaddleNLP/unarchived/neural_machine_translation/transformer/model.py

@@ -875,7 +875,7 @@ def fast_decode(src_vocab_size,
             accu_scores = layers.elementwise_add(
                 x=layers.log(topk_scores), y=pre_scores, axis=0)
             # beam_search op uses lod to differentiate branches.
-            topk_indices = layers.lod_reset(accu_scores, pre_ids)
+            accu_scores = layers.lod_reset(accu_scores, pre_ids)
             # topK reduction across beams, also contain special handle of
             # end beams and end sentences(batch reduction)
             selected_ids, selected_scores, gather_idx = layers.beam_search(

PaddleNLP/unarchived/neural_machine_translation/transformer/reader.py

@@ -266,7 +266,7 @@ class DataReader(object):
                 with open(fpath, "rb") as f:
                     for line in f:
                         if six.PY3:
-                            line = line.decode()
+                            line = line.decode("utf8", errors="ignore")
                         fields = line.strip("\n").split(self._field_delimiter)
                         if (not self._only_src and len(fields) == 2) or (
                                 self._only_src and len(fields) == 1):