fairseq/data/data_utils.py

Survived

Mutant 858

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -17,7 +17,7 @@
 import numpy as np
 
 
-logger = logging.getLogger(__name__)
+logger = None
 
 
 def infer_language_pair(path):

Mutant 859

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -30,7 +30,7 @@
     return src, dst
 
 
-def collate_tokens(values, pad_idx, eos_idx=None, left_pad=False, move_eos_to_beginning=False):
+def collate_tokens(values, pad_idx, eos_idx=None, left_pad=True, move_eos_to_beginning=False):
     """Convert a list of 1d tensors into a padded 2d tensor."""
     size = max(v.size(0) for v in values)
     res = values[0].new(len(values), size).fill_(pad_idx)

Mutant 860

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -30,7 +30,7 @@
     return src, dst
 
 
-def collate_tokens(values, pad_idx, eos_idx=None, left_pad=False, move_eos_to_beginning=False):
+def collate_tokens(values, pad_idx, eos_idx=None, left_pad=False, move_eos_to_beginning=True):
     """Convert a list of 1d tensors into a padded 2d tensor."""
     size = max(v.size(0) for v in values)
     res = values[0].new(len(values), size).fill_(pad_idx)

Mutant 863

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -33,7 +33,7 @@
 def collate_tokens(values, pad_idx, eos_idx=None, left_pad=False, move_eos_to_beginning=False):
     """Convert a list of 1d tensors into a padded 2d tensor."""
     size = max(v.size(0) for v in values)
-    res = values[0].new(len(values), size).fill_(pad_idx)
+    res = values[1].new(len(values), size).fill_(pad_idx)
 
     def copy_tensor(src, dst):
         assert dst.numel() == src.numel()

Mutant 867

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -52,7 +52,7 @@
     return res
 
 
-def load_indexed_dataset(path, dictionary, dataset_impl=None, combine=False, default='cached'):
+def load_indexed_dataset(path, dictionary, dataset_impl=None, combine=True, default='cached'):
     """A helper function for loading indexed datasets.
 
     Args:

Mutant 868

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -52,7 +52,7 @@
     return res
 
 
-def load_indexed_dataset(path, dictionary, dataset_impl=None, combine=False, default='cached'):
+def load_indexed_dataset(path, dictionary, dataset_impl=None, combine=False, default='XXcachedXX'):
     """A helper function for loading indexed datasets.
 
     Args:

Mutant 869

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -96,8 +96,6 @@
     else:
         return ConcatDataset(datasets)
 
-
-@contextlib.contextmanager
 def numpy_seed(seed, *addl_seeds):
     """Context manager which seeds the NumPy PRNG with the specified seed and
     restores the state afterward"""

Mutant 870

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -131,7 +131,7 @@
             filtered.append(el)
 
 
-def _filter_by_size_dynamic(indices, size_fn, max_positions, raise_exception=False):
+def _filter_by_size_dynamic(indices, size_fn, max_positions, raise_exception=True):
     def check_size(idx):
         if isinstance(max_positions, float) or isinstance(max_positions, int):
             return size_fn(idx) <= max_positions

Mutant 871

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -164,7 +164,7 @@
     return indices, ignored
 
 
-def filter_by_size(indices, dataset, max_positions, raise_exception=False):
+def filter_by_size(indices, dataset, max_positions, raise_exception=True):
     """
     Filter indices based on their size.
 

Mutant 872

--- fairseq/data/data_utils.py
+++ fairseq/data/data_utils.py
@@ -203,7 +203,7 @@
 
 def batch_by_size(
     indices, num_tokens_fn, max_tokens=None, max_sentences=None,
-    required_batch_size_multiple=1, fixed_shapes=None,
+    required_batch_size_multiple=2, fixed_shapes=None,
 ):
     """
     Yield mini-batches of indices bucketed by size. Batches may contain