fairseq/tasks/language_modeling.py
Killed 4 out of 49 mutantsSurvived
Survived mutation testing. These mutants show holes in your test suite.Mutant 1258
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -29,7 +29,7 @@
from fairseq.tasks import FairseqTask, register_task
-logger = logging.getLogger(__name__)
+logger = None
@register_task("language_modeling")
Mutant 1259
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -32,7 +32,7 @@
logger = logging.getLogger(__name__)
-@register_task("language_modeling")
+@register_task("XXlanguage_modelingXX")
class LanguageModelingTask(FairseqTask):
"""
Train a language model.
Mutant 1260
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -31,8 +31,6 @@
logger = logging.getLogger(__name__)
-
-@register_task("language_modeling")
class LanguageModelingTask(FairseqTask):
"""
Train a language model.
Mutant 1261
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -62,7 +62,6 @@
:prog:
"""
- @staticmethod
def add_args(parser):
"""Add task-specific arguments to the parser."""
# fmt: off
Mutant 1262
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -66,7 +66,7 @@
def add_args(parser):
"""Add task-specific arguments to the parser."""
# fmt: off
- parser.add_argument('data', help='path to data directory')
+ parser.add_argument('XXdataXX', help='path to data directory')
parser.add_argument('--sample-break-mode', default='none',
choices=['none', 'complete', 'complete_doc', 'eos'],
help='If omitted or "none", fills each sample with tokens-per-sample '
Mutant 1263
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -66,7 +66,7 @@
def add_args(parser):
"""Add task-specific arguments to the parser."""
# fmt: off
- parser.add_argument('data', help='path to data directory')
+ parser.add_argument('data', help='XXpath to data directoryXX')
parser.add_argument('--sample-break-mode', default='none',
choices=['none', 'complete', 'complete_doc', 'eos'],
help='If omitted or "none", fills each sample with tokens-per-sample '
Mutant 1264
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -67,7 +67,7 @@
"""Add task-specific arguments to the parser."""
# fmt: off
parser.add_argument('data', help='path to data directory')
- parser.add_argument('--sample-break-mode', default='none',
+ parser.add_argument('XX--sample-break-modeXX', default='none',
choices=['none', 'complete', 'complete_doc', 'eos'],
help='If omitted or "none", fills each sample with tokens-per-sample '
'tokens. If set to "complete", splits samples only at the end '
Mutant 1265
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -67,7 +67,7 @@
"""Add task-specific arguments to the parser."""
# fmt: off
parser.add_argument('data', help='path to data directory')
- parser.add_argument('--sample-break-mode', default='none',
+ parser.add_argument('--sample-break-mode', default='XXnoneXX',
choices=['none', 'complete', 'complete_doc', 'eos'],
help='If omitted or "none", fills each sample with tokens-per-sample '
'tokens. If set to "complete", splits samples only at the end '
Mutant 1266
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -68,7 +68,7 @@
# fmt: off
parser.add_argument('data', help='path to data directory')
parser.add_argument('--sample-break-mode', default='none',
- choices=['none', 'complete', 'complete_doc', 'eos'],
+ choices=['XXnoneXX', 'complete', 'complete_doc', 'eos'],
help='If omitted or "none", fills each sample with tokens-per-sample '
'tokens. If set to "complete", splits samples only at the end '
'of sentence, but may include multiple sentences per sample. '
Mutant 1267
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -68,7 +68,7 @@
# fmt: off
parser.add_argument('data', help='path to data directory')
parser.add_argument('--sample-break-mode', default='none',
- choices=['none', 'complete', 'complete_doc', 'eos'],
+ choices=['none', 'XXcompleteXX', 'complete_doc', 'eos'],
help='If omitted or "none", fills each sample with tokens-per-sample '
'tokens. If set to "complete", splits samples only at the end '
'of sentence, but may include multiple sentences per sample. '
Mutant 1268
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -68,7 +68,7 @@
# fmt: off
parser.add_argument('data', help='path to data directory')
parser.add_argument('--sample-break-mode', default='none',
- choices=['none', 'complete', 'complete_doc', 'eos'],
+ choices=['none', 'complete', 'XXcomplete_docXX', 'eos'],
help='If omitted or "none", fills each sample with tokens-per-sample '
'tokens. If set to "complete", splits samples only at the end '
'of sentence, but may include multiple sentences per sample. '
Mutant 1269
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -68,7 +68,7 @@
# fmt: off
parser.add_argument('data', help='path to data directory')
parser.add_argument('--sample-break-mode', default='none',
- choices=['none', 'complete', 'complete_doc', 'eos'],
+ choices=['none', 'complete', 'complete_doc', 'XXeosXX'],
help='If omitted or "none", fills each sample with tokens-per-sample '
'tokens. If set to "complete", splits samples only at the end '
'of sentence, but may include multiple sentences per sample. '
Mutant 1270
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -69,7 +69,7 @@
parser.add_argument('data', help='path to data directory')
parser.add_argument('--sample-break-mode', default='none',
choices=['none', 'complete', 'complete_doc', 'eos'],
- help='If omitted or "none", fills each sample with tokens-per-sample '
+ help='XXIf omitted or "none", fills each sample with tokens-per-sample XX'
'tokens. If set to "complete", splits samples only at the end '
'of sentence, but may include multiple sentences per sample. '
'"complete_doc" is similar but respects doc boundaries. '
Mutant 1271
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -74,7 +74,7 @@
'of sentence, but may include multiple sentences per sample. '
'"complete_doc" is similar but respects doc boundaries. '
'If set to "eos", includes only one sentence per sample.')
- parser.add_argument('--tokens-per-sample', default=1024, type=int,
+ parser.add_argument('XX--tokens-per-sampleXX', default=1024, type=int,
help='max number of tokens per sample for LM dataset')
parser.add_argument('--output-dictionary-size', default=-1, type=int,
help='limit the size of output dictionary')
Mutant 1272
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -74,7 +74,7 @@
'of sentence, but may include multiple sentences per sample. '
'"complete_doc" is similar but respects doc boundaries. '
'If set to "eos", includes only one sentence per sample.')
- parser.add_argument('--tokens-per-sample', default=1024, type=int,
+ parser.add_argument('--tokens-per-sample', default=1025, type=int,
help='max number of tokens per sample for LM dataset')
parser.add_argument('--output-dictionary-size', default=-1, type=int,
help='limit the size of output dictionary')
Mutant 1273
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -75,7 +75,7 @@
'"complete_doc" is similar but respects doc boundaries. '
'If set to "eos", includes only one sentence per sample.')
parser.add_argument('--tokens-per-sample', default=1024, type=int,
- help='max number of tokens per sample for LM dataset')
+ help='XXmax number of tokens per sample for LM datasetXX')
parser.add_argument('--output-dictionary-size', default=-1, type=int,
help='limit the size of output dictionary')
parser.add_argument('--self-target', action='store_true',
Mutant 1274
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -76,7 +76,7 @@
'If set to "eos", includes only one sentence per sample.')
parser.add_argument('--tokens-per-sample', default=1024, type=int,
help='max number of tokens per sample for LM dataset')
- parser.add_argument('--output-dictionary-size', default=-1, type=int,
+ parser.add_argument('XX--output-dictionary-sizeXX', default=-1, type=int,
help='limit the size of output dictionary')
parser.add_argument('--self-target', action='store_true',
help='include self target')
Mutant 1275
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -76,7 +76,7 @@
'If set to "eos", includes only one sentence per sample.')
parser.add_argument('--tokens-per-sample', default=1024, type=int,
help='max number of tokens per sample for LM dataset')
- parser.add_argument('--output-dictionary-size', default=-1, type=int,
+ parser.add_argument('--output-dictionary-size', default=+1, type=int,
help='limit the size of output dictionary')
parser.add_argument('--self-target', action='store_true',
help='include self target')
Mutant 1276
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -76,7 +76,7 @@
'If set to "eos", includes only one sentence per sample.')
parser.add_argument('--tokens-per-sample', default=1024, type=int,
help='max number of tokens per sample for LM dataset')
- parser.add_argument('--output-dictionary-size', default=-1, type=int,
+ parser.add_argument('--output-dictionary-size', default=-2, type=int,
help='limit the size of output dictionary')
parser.add_argument('--self-target', action='store_true',
help='include self target')
Mutant 1277
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -77,7 +77,7 @@
parser.add_argument('--tokens-per-sample', default=1024, type=int,
help='max number of tokens per sample for LM dataset')
parser.add_argument('--output-dictionary-size', default=-1, type=int,
- help='limit the size of output dictionary')
+ help='XXlimit the size of output dictionaryXX')
parser.add_argument('--self-target', action='store_true',
help='include self target')
parser.add_argument('--future-target', action='store_true',
Mutant 1278
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -78,7 +78,7 @@
help='max number of tokens per sample for LM dataset')
parser.add_argument('--output-dictionary-size', default=-1, type=int,
help='limit the size of output dictionary')
- parser.add_argument('--self-target', action='store_true',
+ parser.add_argument('XX--self-targetXX', action='store_true',
help='include self target')
parser.add_argument('--future-target', action='store_true',
help='include future target')
Mutant 1280
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -79,7 +79,7 @@
parser.add_argument('--output-dictionary-size', default=-1, type=int,
help='limit the size of output dictionary')
parser.add_argument('--self-target', action='store_true',
- help='include self target')
+ help='XXinclude self targetXX')
parser.add_argument('--future-target', action='store_true',
help='include future target')
parser.add_argument('--past-target', action='store_true',
Mutant 1281
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -80,7 +80,7 @@
help='limit the size of output dictionary')
parser.add_argument('--self-target', action='store_true',
help='include self target')
- parser.add_argument('--future-target', action='store_true',
+ parser.add_argument('XX--future-targetXX', action='store_true',
help='include future target')
parser.add_argument('--past-target', action='store_true',
help='include past target')
Mutant 1283
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -81,7 +81,7 @@
parser.add_argument('--self-target', action='store_true',
help='include self target')
parser.add_argument('--future-target', action='store_true',
- help='include future target')
+ help='XXinclude future targetXX')
parser.add_argument('--past-target', action='store_true',
help='include past target')
parser.add_argument('--add-bos-token', action='store_true',
Mutant 1284
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -82,7 +82,7 @@
help='include self target')
parser.add_argument('--future-target', action='store_true',
help='include future target')
- parser.add_argument('--past-target', action='store_true',
+ parser.add_argument('XX--past-targetXX', action='store_true',
help='include past target')
parser.add_argument('--add-bos-token', action='store_true',
help='prepend beginning of sentence token ()')
Mutant 1286
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -83,7 +83,7 @@
parser.add_argument('--future-target', action='store_true',
help='include future target')
parser.add_argument('--past-target', action='store_true',
- help='include past target')
+ help='XXinclude past targetXX')
parser.add_argument('--add-bos-token', action='store_true',
help='prepend beginning of sentence token ()')
parser.add_argument('--max-target-positions', type=int, metavar='N',
Mutant 1287
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -84,7 +84,7 @@
help='include future target')
parser.add_argument('--past-target', action='store_true',
help='include past target')
- parser.add_argument('--add-bos-token', action='store_true',
+ parser.add_argument('XX--add-bos-tokenXX', action='store_true',
help='prepend beginning of sentence token ()')
parser.add_argument('--max-target-positions', type=int, metavar='N',
help='max number of tokens in the target sequence')
Mutant 1289
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -85,7 +85,7 @@
parser.add_argument('--past-target', action='store_true',
help='include past target')
parser.add_argument('--add-bos-token', action='store_true',
- help='prepend beginning of sentence token ()')
+ help='XXprepend beginning of sentence token ()XX')
parser.add_argument('--max-target-positions', type=int, metavar='N',
help='max number of tokens in the target sequence')
parser.add_argument('--shorten-method', default='none',
Mutant 1290
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -86,7 +86,7 @@
help='include past target')
parser.add_argument('--add-bos-token', action='store_true',
help='prepend beginning of sentence token ()')
- parser.add_argument('--max-target-positions', type=int, metavar='N',
+ parser.add_argument('XX--max-target-positionsXX', type=int, metavar='N',
help='max number of tokens in the target sequence')
parser.add_argument('--shorten-method', default='none',
choices=['none', 'truncate', 'random_crop'],
Mutant 1291
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -86,7 +86,7 @@
help='include past target')
parser.add_argument('--add-bos-token', action='store_true',
help='prepend beginning of sentence token ()')
- parser.add_argument('--max-target-positions', type=int, metavar='N',
+ parser.add_argument('--max-target-positions', type=int, metavar='XXNXX',
help='max number of tokens in the target sequence')
parser.add_argument('--shorten-method', default='none',
choices=['none', 'truncate', 'random_crop'],
Mutant 1292
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -87,7 +87,7 @@
parser.add_argument('--add-bos-token', action='store_true',
help='prepend beginning of sentence token ()')
parser.add_argument('--max-target-positions', type=int, metavar='N',
- help='max number of tokens in the target sequence')
+ help='XXmax number of tokens in the target sequenceXX')
parser.add_argument('--shorten-method', default='none',
choices=['none', 'truncate', 'random_crop'],
help='if not none, shorten sequences that exceed --tokens-per-sample')
Mutant 1293
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -88,7 +88,7 @@
help='prepend beginning of sentence token ()')
parser.add_argument('--max-target-positions', type=int, metavar='N',
help='max number of tokens in the target sequence')
- parser.add_argument('--shorten-method', default='none',
+ parser.add_argument('XX--shorten-methodXX', default='none',
choices=['none', 'truncate', 'random_crop'],
help='if not none, shorten sequences that exceed --tokens-per-sample')
parser.add_argument('--shorten-data-split-list', default='',
Mutant 1294
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -88,7 +88,7 @@
help='prepend beginning of sentence token ()')
parser.add_argument('--max-target-positions', type=int, metavar='N',
help='max number of tokens in the target sequence')
- parser.add_argument('--shorten-method', default='none',
+ parser.add_argument('--shorten-method', default='XXnoneXX',
choices=['none', 'truncate', 'random_crop'],
help='if not none, shorten sequences that exceed --tokens-per-sample')
parser.add_argument('--shorten-data-split-list', default='',
Mutant 1295
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -89,7 +89,7 @@
parser.add_argument('--max-target-positions', type=int, metavar='N',
help='max number of tokens in the target sequence')
parser.add_argument('--shorten-method', default='none',
- choices=['none', 'truncate', 'random_crop'],
+ choices=['XXnoneXX', 'truncate', 'random_crop'],
help='if not none, shorten sequences that exceed --tokens-per-sample')
parser.add_argument('--shorten-data-split-list', default='',
help='comma-separated list of dataset splits to apply shortening to, '
Mutant 1296
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -89,7 +89,7 @@
parser.add_argument('--max-target-positions', type=int, metavar='N',
help='max number of tokens in the target sequence')
parser.add_argument('--shorten-method', default='none',
- choices=['none', 'truncate', 'random_crop'],
+ choices=['none', 'XXtruncateXX', 'random_crop'],
help='if not none, shorten sequences that exceed --tokens-per-sample')
parser.add_argument('--shorten-data-split-list', default='',
help='comma-separated list of dataset splits to apply shortening to, '
Mutant 1297
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -89,7 +89,7 @@
parser.add_argument('--max-target-positions', type=int, metavar='N',
help='max number of tokens in the target sequence')
parser.add_argument('--shorten-method', default='none',
- choices=['none', 'truncate', 'random_crop'],
+ choices=['none', 'truncate', 'XXrandom_cropXX'],
help='if not none, shorten sequences that exceed --tokens-per-sample')
parser.add_argument('--shorten-data-split-list', default='',
help='comma-separated list of dataset splits to apply shortening to, '
Mutant 1298
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -90,7 +90,7 @@
help='max number of tokens in the target sequence')
parser.add_argument('--shorten-method', default='none',
choices=['none', 'truncate', 'random_crop'],
- help='if not none, shorten sequences that exceed --tokens-per-sample')
+ help='XXif not none, shorten sequences that exceed --tokens-per-sampleXX')
parser.add_argument('--shorten-data-split-list', default='',
help='comma-separated list of dataset splits to apply shortening to, '
'e.g., "train,valid" (default: all dataset splits)')
Mutant 1299
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -91,7 +91,7 @@
parser.add_argument('--shorten-method', default='none',
choices=['none', 'truncate', 'random_crop'],
help='if not none, shorten sequences that exceed --tokens-per-sample')
- parser.add_argument('--shorten-data-split-list', default='',
+ parser.add_argument('XX--shorten-data-split-listXX', default='',
help='comma-separated list of dataset splits to apply shortening to, '
'e.g., "train,valid" (default: all dataset splits)')
# fmt: on
Mutant 1300
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -91,7 +91,7 @@
parser.add_argument('--shorten-method', default='none',
choices=['none', 'truncate', 'random_crop'],
help='if not none, shorten sequences that exceed --tokens-per-sample')
- parser.add_argument('--shorten-data-split-list', default='',
+ parser.add_argument('--shorten-data-split-list', default='XXXX',
help='comma-separated list of dataset splits to apply shortening to, '
'e.g., "train,valid" (default: all dataset splits)')
# fmt: on
Mutant 1301
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -92,7 +92,7 @@
choices=['none', 'truncate', 'random_crop'],
help='if not none, shorten sequences that exceed --tokens-per-sample')
parser.add_argument('--shorten-data-split-list', default='',
- help='comma-separated list of dataset splits to apply shortening to, '
+ help='XXcomma-separated list of dataset splits to apply shortening to, XX'
'e.g., "train,valid" (default: all dataset splits)')
# fmt: on
Mutant 1302
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -105,7 +105,6 @@
targets = ["future"]
self.targets = targets
- @classmethod
def setup_task(cls, args, **kwargs):
"""Setup the task (e.g., load dictionaries).
Mutant 1303
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -153,7 +153,7 @@
return model
- def load_dataset(self, split, epoch=1, combine=False, **kwargs):
+ def load_dataset(self, split, epoch=2, combine=False, **kwargs):
"""Load a given dataset split.
Args:
Mutant 1304
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -153,7 +153,7 @@
return model
- def load_dataset(self, split, epoch=1, combine=False, **kwargs):
+ def load_dataset(self, split, epoch=1, combine=True, **kwargs):
"""Load a given dataset split.
Args:
Mutant 1305
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -269,7 +269,6 @@
models, sample, prefix_tokens=prefix_tokens, bos_token=bos_token,
)
- @property
def source_dictionary(self):
"""Return the :class:`~fairseq.data.Dictionary` for the language
model."""
Mutant 1306
--- fairseq/tasks/language_modeling.py
+++ fairseq/tasks/language_modeling.py
@@ -275,7 +275,6 @@
model."""
return self.dictionary
- @property
def target_dictionary(self):
"""Return the :class:`~fairseq.data.Dictionary` for the language
model."""