fairseq/models/lstm.py
Killed 166 out of 356 mutantsSurvived
Survived mutation testing. These mutants show holes in your test suite.Mutant 2137
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -20,7 +20,7 @@
from typing import Dict, List, Optional, Tuple
-DEFAULT_MAX_SOURCE_POSITIONS = 1e5
+DEFAULT_MAX_SOURCE_POSITIONS = 100001.0
DEFAULT_MAX_TARGET_POSITIONS = 1e5
Mutant 2138
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -20,7 +20,7 @@
from typing import Dict, List, Optional, Tuple
-DEFAULT_MAX_SOURCE_POSITIONS = 1e5
+DEFAULT_MAX_SOURCE_POSITIONS = None
DEFAULT_MAX_TARGET_POSITIONS = 1e5
Mutant 2139
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -21,7 +21,7 @@
DEFAULT_MAX_SOURCE_POSITIONS = 1e5
-DEFAULT_MAX_TARGET_POSITIONS = 1e5
+DEFAULT_MAX_TARGET_POSITIONS = 100001.0
@register_model('lstm')
Mutant 2140
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -21,7 +21,7 @@
DEFAULT_MAX_SOURCE_POSITIONS = 1e5
-DEFAULT_MAX_TARGET_POSITIONS = 1e5
+DEFAULT_MAX_TARGET_POSITIONS = None
@register_model('lstm')
Mutant 2143
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -29,7 +29,6 @@
def __init__(self, encoder, decoder):
super().__init__(encoder, decoder)
- @staticmethod
def add_args(parser):
"""Add model-specific arguments to the parser."""
# fmt: off
Mutant 2145
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -33,7 +33,7 @@
def add_args(parser):
"""Add model-specific arguments to the parser."""
# fmt: off
- parser.add_argument('--dropout', type=float, metavar='D',
+ parser.add_argument('--dropout', type=float, metavar='XXDXX',
help='dropout probability')
parser.add_argument('--encoder-embed-dim', type=int, metavar='N',
help='encoder embedding dimension')
Mutant 2146
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -34,7 +34,7 @@
"""Add model-specific arguments to the parser."""
# fmt: off
parser.add_argument('--dropout', type=float, metavar='D',
- help='dropout probability')
+ help='XXdropout probabilityXX')
parser.add_argument('--encoder-embed-dim', type=int, metavar='N',
help='encoder embedding dimension')
parser.add_argument('--encoder-embed-path', type=str, metavar='STR',
Mutant 2148
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -35,7 +35,7 @@
# fmt: off
parser.add_argument('--dropout', type=float, metavar='D',
help='dropout probability')
- parser.add_argument('--encoder-embed-dim', type=int, metavar='N',
+ parser.add_argument('--encoder-embed-dim', type=int, metavar='XXNXX',
help='encoder embedding dimension')
parser.add_argument('--encoder-embed-path', type=str, metavar='STR',
help='path to pre-trained encoder embedding')
Mutant 2149
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -36,7 +36,7 @@
parser.add_argument('--dropout', type=float, metavar='D',
help='dropout probability')
parser.add_argument('--encoder-embed-dim', type=int, metavar='N',
- help='encoder embedding dimension')
+ help='XXencoder embedding dimensionXX')
parser.add_argument('--encoder-embed-path', type=str, metavar='STR',
help='path to pre-trained encoder embedding')
parser.add_argument('--encoder-freeze-embed', action='store_true',
Mutant 2151
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -37,7 +37,7 @@
help='dropout probability')
parser.add_argument('--encoder-embed-dim', type=int, metavar='N',
help='encoder embedding dimension')
- parser.add_argument('--encoder-embed-path', type=str, metavar='STR',
+ parser.add_argument('--encoder-embed-path', type=str, metavar='XXSTRXX',
help='path to pre-trained encoder embedding')
parser.add_argument('--encoder-freeze-embed', action='store_true',
help='freeze encoder embeddings')
Mutant 2152
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -38,7 +38,7 @@
parser.add_argument('--encoder-embed-dim', type=int, metavar='N',
help='encoder embedding dimension')
parser.add_argument('--encoder-embed-path', type=str, metavar='STR',
- help='path to pre-trained encoder embedding')
+ help='XXpath to pre-trained encoder embeddingXX')
parser.add_argument('--encoder-freeze-embed', action='store_true',
help='freeze encoder embeddings')
parser.add_argument('--encoder-hidden-size', type=int, metavar='N',
Mutant 2153
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -39,7 +39,7 @@
help='encoder embedding dimension')
parser.add_argument('--encoder-embed-path', type=str, metavar='STR',
help='path to pre-trained encoder embedding')
- parser.add_argument('--encoder-freeze-embed', action='store_true',
+ parser.add_argument('XX--encoder-freeze-embedXX', action='store_true',
help='freeze encoder embeddings')
parser.add_argument('--encoder-hidden-size', type=int, metavar='N',
help='encoder hidden size')
Mutant 2155
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -40,7 +40,7 @@
parser.add_argument('--encoder-embed-path', type=str, metavar='STR',
help='path to pre-trained encoder embedding')
parser.add_argument('--encoder-freeze-embed', action='store_true',
- help='freeze encoder embeddings')
+ help='XXfreeze encoder embeddingsXX')
parser.add_argument('--encoder-hidden-size', type=int, metavar='N',
help='encoder hidden size')
parser.add_argument('--encoder-layers', type=int, metavar='N',
Mutant 2157
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -41,7 +41,7 @@
help='path to pre-trained encoder embedding')
parser.add_argument('--encoder-freeze-embed', action='store_true',
help='freeze encoder embeddings')
- parser.add_argument('--encoder-hidden-size', type=int, metavar='N',
+ parser.add_argument('--encoder-hidden-size', type=int, metavar='XXNXX',
help='encoder hidden size')
parser.add_argument('--encoder-layers', type=int, metavar='N',
help='number of encoder layers')
Mutant 2158
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -42,7 +42,7 @@
parser.add_argument('--encoder-freeze-embed', action='store_true',
help='freeze encoder embeddings')
parser.add_argument('--encoder-hidden-size', type=int, metavar='N',
- help='encoder hidden size')
+ help='XXencoder hidden sizeXX')
parser.add_argument('--encoder-layers', type=int, metavar='N',
help='number of encoder layers')
parser.add_argument('--encoder-bidirectional', action='store_true',
Mutant 2160
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -43,7 +43,7 @@
help='freeze encoder embeddings')
parser.add_argument('--encoder-hidden-size', type=int, metavar='N',
help='encoder hidden size')
- parser.add_argument('--encoder-layers', type=int, metavar='N',
+ parser.add_argument('--encoder-layers', type=int, metavar='XXNXX',
help='number of encoder layers')
parser.add_argument('--encoder-bidirectional', action='store_true',
help='make all layers of encoder bidirectional')
Mutant 2161
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -44,7 +44,7 @@
parser.add_argument('--encoder-hidden-size', type=int, metavar='N',
help='encoder hidden size')
parser.add_argument('--encoder-layers', type=int, metavar='N',
- help='number of encoder layers')
+ help='XXnumber of encoder layersXX')
parser.add_argument('--encoder-bidirectional', action='store_true',
help='make all layers of encoder bidirectional')
parser.add_argument('--decoder-embed-dim', type=int, metavar='N',
Mutant 2162
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -45,7 +45,7 @@
help='encoder hidden size')
parser.add_argument('--encoder-layers', type=int, metavar='N',
help='number of encoder layers')
- parser.add_argument('--encoder-bidirectional', action='store_true',
+ parser.add_argument('XX--encoder-bidirectionalXX', action='store_true',
help='make all layers of encoder bidirectional')
parser.add_argument('--decoder-embed-dim', type=int, metavar='N',
help='decoder embedding dimension')
Mutant 2164
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -46,7 +46,7 @@
parser.add_argument('--encoder-layers', type=int, metavar='N',
help='number of encoder layers')
parser.add_argument('--encoder-bidirectional', action='store_true',
- help='make all layers of encoder bidirectional')
+ help='XXmake all layers of encoder bidirectionalXX')
parser.add_argument('--decoder-embed-dim', type=int, metavar='N',
help='decoder embedding dimension')
parser.add_argument('--decoder-embed-path', type=str, metavar='STR',
Mutant 2166
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -47,7 +47,7 @@
help='number of encoder layers')
parser.add_argument('--encoder-bidirectional', action='store_true',
help='make all layers of encoder bidirectional')
- parser.add_argument('--decoder-embed-dim', type=int, metavar='N',
+ parser.add_argument('--decoder-embed-dim', type=int, metavar='XXNXX',
help='decoder embedding dimension')
parser.add_argument('--decoder-embed-path', type=str, metavar='STR',
help='path to pre-trained decoder embedding')
Mutant 2167
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -48,7 +48,7 @@
parser.add_argument('--encoder-bidirectional', action='store_true',
help='make all layers of encoder bidirectional')
parser.add_argument('--decoder-embed-dim', type=int, metavar='N',
- help='decoder embedding dimension')
+ help='XXdecoder embedding dimensionXX')
parser.add_argument('--decoder-embed-path', type=str, metavar='STR',
help='path to pre-trained decoder embedding')
parser.add_argument('--decoder-freeze-embed', action='store_true',
Mutant 2169
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -49,7 +49,7 @@
help='make all layers of encoder bidirectional')
parser.add_argument('--decoder-embed-dim', type=int, metavar='N',
help='decoder embedding dimension')
- parser.add_argument('--decoder-embed-path', type=str, metavar='STR',
+ parser.add_argument('--decoder-embed-path', type=str, metavar='XXSTRXX',
help='path to pre-trained decoder embedding')
parser.add_argument('--decoder-freeze-embed', action='store_true',
help='freeze decoder embeddings')
Mutant 2170
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -50,7 +50,7 @@
parser.add_argument('--decoder-embed-dim', type=int, metavar='N',
help='decoder embedding dimension')
parser.add_argument('--decoder-embed-path', type=str, metavar='STR',
- help='path to pre-trained decoder embedding')
+ help='XXpath to pre-trained decoder embeddingXX')
parser.add_argument('--decoder-freeze-embed', action='store_true',
help='freeze decoder embeddings')
parser.add_argument('--decoder-hidden-size', type=int, metavar='N',
Mutant 2171
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -51,7 +51,7 @@
help='decoder embedding dimension')
parser.add_argument('--decoder-embed-path', type=str, metavar='STR',
help='path to pre-trained decoder embedding')
- parser.add_argument('--decoder-freeze-embed', action='store_true',
+ parser.add_argument('XX--decoder-freeze-embedXX', action='store_true',
help='freeze decoder embeddings')
parser.add_argument('--decoder-hidden-size', type=int, metavar='N',
help='decoder hidden size')
Mutant 2173
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -52,7 +52,7 @@
parser.add_argument('--decoder-embed-path', type=str, metavar='STR',
help='path to pre-trained decoder embedding')
parser.add_argument('--decoder-freeze-embed', action='store_true',
- help='freeze decoder embeddings')
+ help='XXfreeze decoder embeddingsXX')
parser.add_argument('--decoder-hidden-size', type=int, metavar='N',
help='decoder hidden size')
parser.add_argument('--decoder-layers', type=int, metavar='N',
Mutant 2175
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -53,7 +53,7 @@
help='path to pre-trained decoder embedding')
parser.add_argument('--decoder-freeze-embed', action='store_true',
help='freeze decoder embeddings')
- parser.add_argument('--decoder-hidden-size', type=int, metavar='N',
+ parser.add_argument('--decoder-hidden-size', type=int, metavar='XXNXX',
help='decoder hidden size')
parser.add_argument('--decoder-layers', type=int, metavar='N',
help='number of decoder layers')
Mutant 2176
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -54,7 +54,7 @@
parser.add_argument('--decoder-freeze-embed', action='store_true',
help='freeze decoder embeddings')
parser.add_argument('--decoder-hidden-size', type=int, metavar='N',
- help='decoder hidden size')
+ help='XXdecoder hidden sizeXX')
parser.add_argument('--decoder-layers', type=int, metavar='N',
help='number of decoder layers')
parser.add_argument('--decoder-out-embed-dim', type=int, metavar='N',
Mutant 2178
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -55,7 +55,7 @@
help='freeze decoder embeddings')
parser.add_argument('--decoder-hidden-size', type=int, metavar='N',
help='decoder hidden size')
- parser.add_argument('--decoder-layers', type=int, metavar='N',
+ parser.add_argument('--decoder-layers', type=int, metavar='XXNXX',
help='number of decoder layers')
parser.add_argument('--decoder-out-embed-dim', type=int, metavar='N',
help='decoder output embedding dimension')
Mutant 2179
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -56,7 +56,7 @@
parser.add_argument('--decoder-hidden-size', type=int, metavar='N',
help='decoder hidden size')
parser.add_argument('--decoder-layers', type=int, metavar='N',
- help='number of decoder layers')
+ help='XXnumber of decoder layersXX')
parser.add_argument('--decoder-out-embed-dim', type=int, metavar='N',
help='decoder output embedding dimension')
parser.add_argument('--decoder-attention', type=str, metavar='BOOL',
Mutant 2181
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -57,7 +57,7 @@
help='decoder hidden size')
parser.add_argument('--decoder-layers', type=int, metavar='N',
help='number of decoder layers')
- parser.add_argument('--decoder-out-embed-dim', type=int, metavar='N',
+ parser.add_argument('--decoder-out-embed-dim', type=int, metavar='XXNXX',
help='decoder output embedding dimension')
parser.add_argument('--decoder-attention', type=str, metavar='BOOL',
help='decoder attention')
Mutant 2182
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -58,7 +58,7 @@
parser.add_argument('--decoder-layers', type=int, metavar='N',
help='number of decoder layers')
parser.add_argument('--decoder-out-embed-dim', type=int, metavar='N',
- help='decoder output embedding dimension')
+ help='XXdecoder output embedding dimensionXX')
parser.add_argument('--decoder-attention', type=str, metavar='BOOL',
help='decoder attention')
parser.add_argument('--adaptive-softmax-cutoff', metavar='EXPR',
Mutant 2184
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -59,7 +59,7 @@
help='number of decoder layers')
parser.add_argument('--decoder-out-embed-dim', type=int, metavar='N',
help='decoder output embedding dimension')
- parser.add_argument('--decoder-attention', type=str, metavar='BOOL',
+ parser.add_argument('--decoder-attention', type=str, metavar='XXBOOLXX',
help='decoder attention')
parser.add_argument('--adaptive-softmax-cutoff', metavar='EXPR',
help='comma separated list of adaptive softmax cutoff points. '
Mutant 2185
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -60,7 +60,7 @@
parser.add_argument('--decoder-out-embed-dim', type=int, metavar='N',
help='decoder output embedding dimension')
parser.add_argument('--decoder-attention', type=str, metavar='BOOL',
- help='decoder attention')
+ help='XXdecoder attentionXX')
parser.add_argument('--adaptive-softmax-cutoff', metavar='EXPR',
help='comma separated list of adaptive softmax cutoff points. '
'Must be used with adaptive_loss criterion')
Mutant 2187
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -61,7 +61,7 @@
help='decoder output embedding dimension')
parser.add_argument('--decoder-attention', type=str, metavar='BOOL',
help='decoder attention')
- parser.add_argument('--adaptive-softmax-cutoff', metavar='EXPR',
+ parser.add_argument('--adaptive-softmax-cutoff', metavar='XXEXPRXX',
help='comma separated list of adaptive softmax cutoff points. '
'Must be used with adaptive_loss criterion')
parser.add_argument('--share-decoder-input-output-embed', default=False,
Mutant 2188
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -62,7 +62,7 @@
parser.add_argument('--decoder-attention', type=str, metavar='BOOL',
help='decoder attention')
parser.add_argument('--adaptive-softmax-cutoff', metavar='EXPR',
- help='comma separated list of adaptive softmax cutoff points. '
+ help='XXcomma separated list of adaptive softmax cutoff points. XX'
'Must be used with adaptive_loss criterion')
parser.add_argument('--share-decoder-input-output-embed', default=False,
action='store_true',
Mutant 2189
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -64,7 +64,7 @@
parser.add_argument('--adaptive-softmax-cutoff', metavar='EXPR',
help='comma separated list of adaptive softmax cutoff points. '
'Must be used with adaptive_loss criterion')
- parser.add_argument('--share-decoder-input-output-embed', default=False,
+ parser.add_argument('XX--share-decoder-input-output-embedXX', default=False,
action='store_true',
help='share decoder input and output embeddings')
parser.add_argument('--share-all-embeddings', default=False, action='store_true',
Mutant 2192
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -66,7 +66,7 @@
'Must be used with adaptive_loss criterion')
parser.add_argument('--share-decoder-input-output-embed', default=False,
action='store_true',
- help='share decoder input and output embeddings')
+ help='XXshare decoder input and output embeddingsXX')
parser.add_argument('--share-all-embeddings', default=False, action='store_true',
help='share encoder, decoder and output embeddings'
' (requires shared dictionary and embed dim)')
Mutant 2193
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -67,7 +67,7 @@
parser.add_argument('--share-decoder-input-output-embed', default=False,
action='store_true',
help='share decoder input and output embeddings')
- parser.add_argument('--share-all-embeddings', default=False, action='store_true',
+ parser.add_argument('XX--share-all-embeddingsXX', default=False, action='store_true',
help='share encoder, decoder and output embeddings'
' (requires shared dictionary and embed dim)')
Mutant 2196
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -68,7 +68,7 @@
action='store_true',
help='share decoder input and output embeddings')
parser.add_argument('--share-all-embeddings', default=False, action='store_true',
- help='share encoder, decoder and output embeddings'
+ help='XXshare encoder, decoder and output embeddingsXX'
' (requires shared dictionary and embed dim)')
# Granular dropout settings (if not specified these default to --dropout)
Mutant 2198
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -72,7 +72,7 @@
' (requires shared dictionary and embed dim)')
# Granular dropout settings (if not specified these default to --dropout)
- parser.add_argument('--encoder-dropout-in', type=float, metavar='D',
+ parser.add_argument('--encoder-dropout-in', type=float, metavar='XXDXX',
help='dropout probability for encoder input embedding')
parser.add_argument('--encoder-dropout-out', type=float, metavar='D',
help='dropout probability for encoder output')
Mutant 2199
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -73,7 +73,7 @@
# Granular dropout settings (if not specified these default to --dropout)
parser.add_argument('--encoder-dropout-in', type=float, metavar='D',
- help='dropout probability for encoder input embedding')
+ help='XXdropout probability for encoder input embeddingXX')
parser.add_argument('--encoder-dropout-out', type=float, metavar='D',
help='dropout probability for encoder output')
parser.add_argument('--decoder-dropout-in', type=float, metavar='D',
Mutant 2201
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -74,7 +74,7 @@
# Granular dropout settings (if not specified these default to --dropout)
parser.add_argument('--encoder-dropout-in', type=float, metavar='D',
help='dropout probability for encoder input embedding')
- parser.add_argument('--encoder-dropout-out', type=float, metavar='D',
+ parser.add_argument('--encoder-dropout-out', type=float, metavar='XXDXX',
help='dropout probability for encoder output')
parser.add_argument('--decoder-dropout-in', type=float, metavar='D',
help='dropout probability for decoder input embedding')
Mutant 2202
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -75,7 +75,7 @@
parser.add_argument('--encoder-dropout-in', type=float, metavar='D',
help='dropout probability for encoder input embedding')
parser.add_argument('--encoder-dropout-out', type=float, metavar='D',
- help='dropout probability for encoder output')
+ help='XXdropout probability for encoder outputXX')
parser.add_argument('--decoder-dropout-in', type=float, metavar='D',
help='dropout probability for decoder input embedding')
parser.add_argument('--decoder-dropout-out', type=float, metavar='D',
Mutant 2204
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -76,7 +76,7 @@
help='dropout probability for encoder input embedding')
parser.add_argument('--encoder-dropout-out', type=float, metavar='D',
help='dropout probability for encoder output')
- parser.add_argument('--decoder-dropout-in', type=float, metavar='D',
+ parser.add_argument('--decoder-dropout-in', type=float, metavar='XXDXX',
help='dropout probability for decoder input embedding')
parser.add_argument('--decoder-dropout-out', type=float, metavar='D',
help='dropout probability for decoder output')
Mutant 2205
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -77,7 +77,7 @@
parser.add_argument('--encoder-dropout-out', type=float, metavar='D',
help='dropout probability for encoder output')
parser.add_argument('--decoder-dropout-in', type=float, metavar='D',
- help='dropout probability for decoder input embedding')
+ help='XXdropout probability for decoder input embeddingXX')
parser.add_argument('--decoder-dropout-out', type=float, metavar='D',
help='dropout probability for decoder output')
# fmt: on
Mutant 2207
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -78,7 +78,7 @@
help='dropout probability for encoder output')
parser.add_argument('--decoder-dropout-in', type=float, metavar='D',
help='dropout probability for decoder input embedding')
- parser.add_argument('--decoder-dropout-out', type=float, metavar='D',
+ parser.add_argument('--decoder-dropout-out', type=float, metavar='XXDXX',
help='dropout probability for decoder output')
# fmt: on
Mutant 2208
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -79,7 +79,7 @@
parser.add_argument('--decoder-dropout-in', type=float, metavar='D',
help='dropout probability for decoder input embedding')
parser.add_argument('--decoder-dropout-out', type=float, metavar='D',
- help='dropout probability for decoder output')
+ help='XXdropout probability for decoder outputXX')
# fmt: on
@classmethod
Mutant 2212
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -91,7 +91,7 @@
if args.encoder_layers != args.decoder_layers:
raise ValueError('--encoder-layers must match --decoder-layers')
- max_source_positions = getattr(args, 'max_source_positions', DEFAULT_MAX_SOURCE_POSITIONS)
+ max_source_positions = getattr(args, 'XXmax_source_positionsXX', DEFAULT_MAX_SOURCE_POSITIONS)
max_target_positions = getattr(args, 'max_target_positions', DEFAULT_MAX_TARGET_POSITIONS)
def load_pretrained_embedding_from_file(embed_path, dictionary, embed_dim):
Mutant 2213
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -91,7 +91,7 @@
if args.encoder_layers != args.decoder_layers:
raise ValueError('--encoder-layers must match --decoder-layers')
- max_source_positions = getattr(args, 'max_source_positions', DEFAULT_MAX_SOURCE_POSITIONS)
+ max_source_positions = None
max_target_positions = getattr(args, 'max_target_positions', DEFAULT_MAX_TARGET_POSITIONS)
def load_pretrained_embedding_from_file(embed_path, dictionary, embed_dim):
Mutant 2214
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -92,7 +92,7 @@
raise ValueError('--encoder-layers must match --decoder-layers')
max_source_positions = getattr(args, 'max_source_positions', DEFAULT_MAX_SOURCE_POSITIONS)
- max_target_positions = getattr(args, 'max_target_positions', DEFAULT_MAX_TARGET_POSITIONS)
+ max_target_positions = getattr(args, 'XXmax_target_positionsXX', DEFAULT_MAX_TARGET_POSITIONS)
def load_pretrained_embedding_from_file(embed_path, dictionary, embed_dim):
num_embeddings = len(dictionary)
Mutant 2215
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -92,7 +92,7 @@
raise ValueError('--encoder-layers must match --decoder-layers')
max_source_positions = getattr(args, 'max_source_positions', DEFAULT_MAX_SOURCE_POSITIONS)
- max_target_positions = getattr(args, 'max_target_positions', DEFAULT_MAX_TARGET_POSITIONS)
+ max_target_positions = None
def load_pretrained_embedding_from_file(embed_path, dictionary, embed_dim):
num_embeddings = len(dictionary)
Mutant 2219
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -174,7 +174,7 @@
share_input_output_embed=args.share_decoder_input_output_embed,
adaptive_softmax_cutoff=(
options.eval_str_list(args.adaptive_softmax_cutoff, type=int)
- if args.criterion == 'adaptive_loss' else None
+ if args.criterion == 'XXadaptive_lossXX' else None
),
max_target_positions=max_target_positions,
residuals=False,
Mutant 2220
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -177,7 +177,7 @@
if args.criterion == 'adaptive_loss' else None
),
max_target_positions=max_target_positions,
- residuals=False,
+ residuals=True,
)
return cls(encoder, decoder)
Mutant 2222
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -198,7 +198,7 @@
class LSTMEncoder(FairseqEncoder):
"""LSTM encoder."""
def __init__(
- self, dictionary, embed_dim=512, hidden_size=512, num_layers=1,
+ self, dictionary, embed_dim=513, hidden_size=512, num_layers=1,
dropout_in=0.1, dropout_out=0.1, bidirectional=False,
left_pad=True, pretrained_embed=None, padding_idx=None,
max_source_positions=DEFAULT_MAX_SOURCE_POSITIONS,
Mutant 2223
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -198,7 +198,7 @@
class LSTMEncoder(FairseqEncoder):
"""LSTM encoder."""
def __init__(
- self, dictionary, embed_dim=512, hidden_size=512, num_layers=1,
+ self, dictionary, embed_dim=512, hidden_size=513, num_layers=1,
dropout_in=0.1, dropout_out=0.1, bidirectional=False,
left_pad=True, pretrained_embed=None, padding_idx=None,
max_source_positions=DEFAULT_MAX_SOURCE_POSITIONS,
Mutant 2224
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -198,7 +198,7 @@
class LSTMEncoder(FairseqEncoder):
"""LSTM encoder."""
def __init__(
- self, dictionary, embed_dim=512, hidden_size=512, num_layers=1,
+ self, dictionary, embed_dim=512, hidden_size=512, num_layers=2,
dropout_in=0.1, dropout_out=0.1, bidirectional=False,
left_pad=True, pretrained_embed=None, padding_idx=None,
max_source_positions=DEFAULT_MAX_SOURCE_POSITIONS,
Mutant 2225
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -199,7 +199,7 @@
"""LSTM encoder."""
def __init__(
self, dictionary, embed_dim=512, hidden_size=512, num_layers=1,
- dropout_in=0.1, dropout_out=0.1, bidirectional=False,
+ dropout_in=1.1, dropout_out=0.1, bidirectional=False,
left_pad=True, pretrained_embed=None, padding_idx=None,
max_source_positions=DEFAULT_MAX_SOURCE_POSITIONS,
):
Mutant 2226
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -199,7 +199,7 @@
"""LSTM encoder."""
def __init__(
self, dictionary, embed_dim=512, hidden_size=512, num_layers=1,
- dropout_in=0.1, dropout_out=0.1, bidirectional=False,
+ dropout_in=0.1, dropout_out=1.1, bidirectional=False,
left_pad=True, pretrained_embed=None, padding_idx=None,
max_source_positions=DEFAULT_MAX_SOURCE_POSITIONS,
):
Mutant 2227
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -199,7 +199,7 @@
"""LSTM encoder."""
def __init__(
self, dictionary, embed_dim=512, hidden_size=512, num_layers=1,
- dropout_in=0.1, dropout_out=0.1, bidirectional=False,
+ dropout_in=0.1, dropout_out=0.1, bidirectional=True,
left_pad=True, pretrained_embed=None, padding_idx=None,
max_source_positions=DEFAULT_MAX_SOURCE_POSITIONS,
):
Mutant 2228
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -200,7 +200,7 @@
def __init__(
self, dictionary, embed_dim=512, hidden_size=512, num_layers=1,
dropout_in=0.1, dropout_out=0.1, bidirectional=False,
- left_pad=True, pretrained_embed=None, padding_idx=None,
+ left_pad=False, pretrained_embed=None, padding_idx=None,
max_source_positions=DEFAULT_MAX_SOURCE_POSITIONS,
):
super().__init__(dictionary)
Mutant 2234
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -209,7 +209,7 @@
self.dropout_out_module = FairseqDropout(dropout_out, module_name=self.__class__.__name__)
self.bidirectional = bidirectional
self.hidden_size = hidden_size
- self.max_source_positions = max_source_positions
+ self.max_source_positions = None
num_embeddings = len(dictionary)
self.padding_idx = padding_idx if padding_idx is not None else dictionary.pad()
Mutant 2235
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -211,7 +211,7 @@
self.hidden_size = hidden_size
self.max_source_positions = max_source_positions
- num_embeddings = len(dictionary)
+ num_embeddings = None
self.padding_idx = padding_idx if padding_idx is not None else dictionary.pad()
if pretrained_embed is None:
self.embed_tokens = Embedding(num_embeddings, embed_dim, self.padding_idx)
Mutant 2238
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -213,7 +213,7 @@
num_embeddings = len(dictionary)
self.padding_idx = padding_idx if padding_idx is not None else dictionary.pad()
- if pretrained_embed is None:
+ if pretrained_embed is not None:
self.embed_tokens = Embedding(num_embeddings, embed_dim, self.padding_idx)
else:
self.embed_tokens = pretrained_embed
Mutant 2240
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -222,7 +222,7 @@
input_size=embed_dim,
hidden_size=hidden_size,
num_layers=num_layers,
- dropout=self.dropout_out_module.p if num_layers > 1 else 0.,
+ dropout=self.dropout_out_module.p if num_layers >= 1 else 0.,
bidirectional=bidirectional,
)
self.left_pad = left_pad
Mutant 2241
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -222,7 +222,7 @@
input_size=embed_dim,
hidden_size=hidden_size,
num_layers=num_layers,
- dropout=self.dropout_out_module.p if num_layers > 1 else 0.,
+ dropout=self.dropout_out_module.p if num_layers > 2 else 0.,
bidirectional=bidirectional,
)
self.left_pad = left_pad
Mutant 2242
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -222,7 +222,7 @@
input_size=embed_dim,
hidden_size=hidden_size,
num_layers=num_layers,
- dropout=self.dropout_out_module.p if num_layers > 1 else 0.,
+ dropout=self.dropout_out_module.p if num_layers > 1 else 1.0,
bidirectional=bidirectional,
)
self.left_pad = left_pad
Mutant 2245
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -235,7 +235,7 @@
self,
src_tokens: Tensor,
src_lengths: Tensor,
- enforce_sorted: bool = True,
+ enforce_sorted: bool = False,
):
"""
Args:
Mutant 2257
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -281,7 +281,7 @@
packed_outs, (final_hiddens, final_cells) = self.lstm(packed_x, (h0, c0))
# unpack outputs and apply dropout
- x, _ = nn.utils.rnn.pad_packed_sequence(packed_outs, padding_value=self.padding_idx*1.0)
+ x, _ = nn.utils.rnn.pad_packed_sequence(packed_outs, padding_value=self.padding_idx/1.0)
x = self.dropout_out_module(x)
assert list(x.size()) == [seqlen, bsz, self.output_units]
Mutant 2258
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -281,7 +281,7 @@
packed_outs, (final_hiddens, final_cells) = self.lstm(packed_x, (h0, c0))
# unpack outputs and apply dropout
- x, _ = nn.utils.rnn.pad_packed_sequence(packed_outs, padding_value=self.padding_idx*1.0)
+ x, _ = nn.utils.rnn.pad_packed_sequence(packed_outs, padding_value=self.padding_idx*2.0)
x = self.dropout_out_module(x)
assert list(x.size()) == [seqlen, bsz, self.output_units]
Mutant 2263
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -316,7 +316,7 @@
class AttentionLayer(nn.Module):
- def __init__(self, input_embed_dim, source_embed_dim, output_embed_dim, bias=False):
+ def __init__(self, input_embed_dim, source_embed_dim, output_embed_dim, bias=True):
super().__init__()
self.input_proj = Linear(input_embed_dim, source_embed_dim, bias=bias)
Mutant 2268
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -330,7 +330,7 @@
x = self.input_proj(input)
# compute attention
- attn_scores = (source_hids * x.unsqueeze(0)).sum(dim=2)
+ attn_scores = (source_hids / x.unsqueeze(0)).sum(dim=2)
# don't attend over padding
if encoder_padding_mask is not None:
Mutant 2272
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -333,7 +333,7 @@
attn_scores = (source_hids * x.unsqueeze(0)).sum(dim=2)
# don't attend over padding
- if encoder_padding_mask is not None:
+ if encoder_padding_mask is None:
attn_scores = attn_scores.float().masked_fill_(
encoder_padding_mask,
float('-inf')
Mutant 2275
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -339,7 +339,7 @@
float('-inf')
).type_as(attn_scores) # FP16 support: cast to float and back
- attn_scores = F.softmax(attn_scores, dim=0) # srclen x bsz
+ attn_scores = F.softmax(attn_scores, dim=1) # srclen x bsz
# sum weighted sources
x = (attn_scores.unsqueeze(2) * source_hids).sum(dim=0)
Mutant 2278
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -342,7 +342,7 @@
attn_scores = F.softmax(attn_scores, dim=0) # srclen x bsz
# sum weighted sources
- x = (attn_scores.unsqueeze(2) * source_hids).sum(dim=0)
+ x = (attn_scores.unsqueeze(2) / source_hids).sum(dim=0)
x = torch.tanh(self.output_proj(torch.cat((x, input), dim=1)))
return x, attn_scores
Mutant 2283
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -351,7 +351,7 @@
class LSTMDecoder(FairseqIncrementalDecoder):
"""LSTM decoder."""
def __init__(
- self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=512,
+ self, dictionary, embed_dim=513, hidden_size=512, out_embed_dim=512,
num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=True,
encoder_output_units=512, pretrained_embed=None,
share_input_output_embed=False, adaptive_softmax_cutoff=None,
Mutant 2284
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -351,7 +351,7 @@
class LSTMDecoder(FairseqIncrementalDecoder):
"""LSTM decoder."""
def __init__(
- self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=512,
+ self, dictionary, embed_dim=512, hidden_size=513, out_embed_dim=512,
num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=True,
encoder_output_units=512, pretrained_embed=None,
share_input_output_embed=False, adaptive_softmax_cutoff=None,
Mutant 2285
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -351,7 +351,7 @@
class LSTMDecoder(FairseqIncrementalDecoder):
"""LSTM decoder."""
def __init__(
- self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=512,
+ self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=513,
num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=True,
encoder_output_units=512, pretrained_embed=None,
share_input_output_embed=False, adaptive_softmax_cutoff=None,
Mutant 2286
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -352,7 +352,7 @@
"""LSTM decoder."""
def __init__(
self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=512,
- num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=True,
+ num_layers=2, dropout_in=0.1, dropout_out=0.1, attention=True,
encoder_output_units=512, pretrained_embed=None,
share_input_output_embed=False, adaptive_softmax_cutoff=None,
max_target_positions=DEFAULT_MAX_TARGET_POSITIONS,
Mutant 2287
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -352,7 +352,7 @@
"""LSTM decoder."""
def __init__(
self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=512,
- num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=True,
+ num_layers=1, dropout_in=1.1, dropout_out=0.1, attention=True,
encoder_output_units=512, pretrained_embed=None,
share_input_output_embed=False, adaptive_softmax_cutoff=None,
max_target_positions=DEFAULT_MAX_TARGET_POSITIONS,
Mutant 2288
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -352,7 +352,7 @@
"""LSTM decoder."""
def __init__(
self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=512,
- num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=True,
+ num_layers=1, dropout_in=0.1, dropout_out=1.1, attention=True,
encoder_output_units=512, pretrained_embed=None,
share_input_output_embed=False, adaptive_softmax_cutoff=None,
max_target_positions=DEFAULT_MAX_TARGET_POSITIONS,
Mutant 2289
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -352,7 +352,7 @@
"""LSTM decoder."""
def __init__(
self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=512,
- num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=True,
+ num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=False,
encoder_output_units=512, pretrained_embed=None,
share_input_output_embed=False, adaptive_softmax_cutoff=None,
max_target_positions=DEFAULT_MAX_TARGET_POSITIONS,
Mutant 2290
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -353,7 +353,7 @@
def __init__(
self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=512,
num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=True,
- encoder_output_units=512, pretrained_embed=None,
+ encoder_output_units=513, pretrained_embed=None,
share_input_output_embed=False, adaptive_softmax_cutoff=None,
max_target_positions=DEFAULT_MAX_TARGET_POSITIONS,
residuals=False,
Mutant 2291
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -354,7 +354,7 @@
self, dictionary, embed_dim=512, hidden_size=512, out_embed_dim=512,
num_layers=1, dropout_in=0.1, dropout_out=0.1, attention=True,
encoder_output_units=512, pretrained_embed=None,
- share_input_output_embed=False, adaptive_softmax_cutoff=None,
+ share_input_output_embed=True, adaptive_softmax_cutoff=None,
max_target_positions=DEFAULT_MAX_TARGET_POSITIONS,
residuals=False,
):
Mutant 2292
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -356,7 +356,7 @@
encoder_output_units=512, pretrained_embed=None,
share_input_output_embed=False, adaptive_softmax_cutoff=None,
max_target_positions=DEFAULT_MAX_TARGET_POSITIONS,
- residuals=False,
+ residuals=True,
):
super().__init__(dictionary)
self.dropout_in_module = FairseqDropout(dropout_in, module_name=self.__class__.__name__)
Mutant 2299
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -364,7 +364,7 @@
self.hidden_size = hidden_size
self.share_input_output_embed = share_input_output_embed
self.need_attn = True
- self.max_target_positions = max_target_positions
+ self.max_target_positions = None
self.residuals = residuals
self.num_layers = num_layers
Mutant 2302
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -368,7 +368,7 @@
self.residuals = residuals
self.num_layers = num_layers
- self.adaptive_softmax = None
+ self.adaptive_softmax = ""
num_embeddings = len(dictionary)
padding_idx = dictionary.pad()
if pretrained_embed is None:
Mutant 2304
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -370,7 +370,7 @@
self.adaptive_softmax = None
num_embeddings = len(dictionary)
- padding_idx = dictionary.pad()
+ padding_idx = None
if pretrained_embed is None:
self.embed_tokens = Embedding(num_embeddings, embed_dim, padding_idx)
else:
Mutant 2307
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -376,7 +376,7 @@
else:
self.embed_tokens = pretrained_embed
- self.encoder_output_units = encoder_output_units
+ self.encoder_output_units = None
if encoder_output_units != hidden_size and encoder_output_units != 0:
self.encoder_hidden_proj = Linear(encoder_output_units, hidden_size)
self.encoder_cell_proj = Linear(encoder_output_units, hidden_size)
Mutant 2308
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -377,7 +377,7 @@
self.embed_tokens = pretrained_embed
self.encoder_output_units = encoder_output_units
- if encoder_output_units != hidden_size and encoder_output_units != 0:
+ if encoder_output_units == hidden_size and encoder_output_units != 0:
self.encoder_hidden_proj = Linear(encoder_output_units, hidden_size)
self.encoder_cell_proj = Linear(encoder_output_units, hidden_size)
else:
Mutant 2309
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -377,7 +377,7 @@
self.embed_tokens = pretrained_embed
self.encoder_output_units = encoder_output_units
- if encoder_output_units != hidden_size and encoder_output_units != 0:
+ if encoder_output_units != hidden_size and encoder_output_units == 0:
self.encoder_hidden_proj = Linear(encoder_output_units, hidden_size)
self.encoder_cell_proj = Linear(encoder_output_units, hidden_size)
else:
Mutant 2310
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -377,7 +377,7 @@
self.embed_tokens = pretrained_embed
self.encoder_output_units = encoder_output_units
- if encoder_output_units != hidden_size and encoder_output_units != 0:
+ if encoder_output_units != hidden_size and encoder_output_units != 1:
self.encoder_hidden_proj = Linear(encoder_output_units, hidden_size)
self.encoder_cell_proj = Linear(encoder_output_units, hidden_size)
else:
Mutant 2311
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -377,7 +377,7 @@
self.embed_tokens = pretrained_embed
self.encoder_output_units = encoder_output_units
- if encoder_output_units != hidden_size and encoder_output_units != 0:
+ if encoder_output_units != hidden_size or encoder_output_units != 0:
self.encoder_hidden_proj = Linear(encoder_output_units, hidden_size)
self.encoder_cell_proj = Linear(encoder_output_units, hidden_size)
else:
Mutant 2313
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -385,7 +385,7 @@
# disable input feeding if there is no encoder
# input feeding is described in arxiv.org/abs/1508.04025
- input_feed_size = 0 if encoder_output_units == 0 else hidden_size
+ input_feed_size = 1 if encoder_output_units == 0 else hidden_size
self.layers = nn.ModuleList([
LSTMCell(
input_size=input_feed_size + embed_dim if layer == 0 else hidden_size,
Mutant 2315
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -385,7 +385,7 @@
# disable input feeding if there is no encoder
# input feeding is described in arxiv.org/abs/1508.04025
- input_feed_size = 0 if encoder_output_units == 0 else hidden_size
+ input_feed_size = 0 if encoder_output_units == 1 else hidden_size
self.layers = nn.ModuleList([
LSTMCell(
input_size=input_feed_size + embed_dim if layer == 0 else hidden_size,
Mutant 2317
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -396,7 +396,7 @@
if attention:
# TODO make bias configurable
- self.attention = AttentionLayer(hidden_size, encoder_output_units, hidden_size, bias=False)
+ self.attention = AttentionLayer(hidden_size, encoder_output_units, hidden_size, bias=True)
else:
self.attention = None
Mutant 2319
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -400,7 +400,7 @@
else:
self.attention = None
- if hidden_size != out_embed_dim:
+ if hidden_size == out_embed_dim:
self.additional_fc = Linear(hidden_size, out_embed_dim)
if adaptive_softmax_cutoff is not None:
Mutant 2326
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -435,7 +435,7 @@
# get outputs from encoder
if encoder_out is not None:
encoder_outs = encoder_out[0]
- encoder_hiddens = encoder_out[1]
+ encoder_hiddens = encoder_out[2]
encoder_cells = encoder_out[2]
encoder_padding_mask = encoder_out[3]
else:
Mutant 2335
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -445,7 +445,7 @@
encoder_padding_mask = torch.empty(0)
srclen = encoder_outs.size(0)
- if incremental_state is not None and len(incremental_state) > 0:
+ if incremental_state is not None and len(incremental_state) >= 0:
prev_output_tokens = prev_output_tokens[:, -1:]
bsz, seqlen = prev_output_tokens.size()
Mutant 2336
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -445,7 +445,7 @@
encoder_padding_mask = torch.empty(0)
srclen = encoder_outs.size(0)
- if incremental_state is not None and len(incremental_state) > 0:
+ if incremental_state is not None and len(incremental_state) > 1:
prev_output_tokens = prev_output_tokens[:, -1:]
bsz, seqlen = prev_output_tokens.size()
Mutant 2345
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -458,7 +458,7 @@
x = x.transpose(0, 1)
# initialize previous states (or get from cache during incremental generation)
- if incremental_state is not None and len(incremental_state) > 0:
+ if incremental_state is not None and len(incremental_state) >= 0:
prev_hiddens, prev_cells, input_feed = self.get_cached_state(incremental_state)
elif encoder_out is not None:
# setup recurrent cells
Mutant 2346
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -458,7 +458,7 @@
x = x.transpose(0, 1)
# initialize previous states (or get from cache during incremental generation)
- if incremental_state is not None and len(incremental_state) > 0:
+ if incremental_state is not None and len(incremental_state) > 1:
prev_hiddens, prev_cells, input_feed = self.get_cached_state(incremental_state)
elif encoder_out is not None:
# setup recurrent cells
Mutant 2353
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -475,7 +475,7 @@
prev_cells = [zero_state for i in range(self.num_layers)]
input_feed = None
- assert srclen > 0 or self.attention is None, \
+ assert srclen >= 0 or self.attention is None, \
"attention is not supported if there are no encoder outputs"
attn_scores = x.new_zeros(srclen, seqlen, bsz) if self.attention is not None else None
outs = []
Mutant 2355
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -475,7 +475,7 @@
prev_cells = [zero_state for i in range(self.num_layers)]
input_feed = None
- assert srclen > 0 or self.attention is None, \
+ assert srclen > 0 or self.attention is not None, \
"attention is not supported if there are no encoder outputs"
attn_scores = x.new_zeros(srclen, seqlen, bsz) if self.attention is not None else None
outs = []
Mutant 2367
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -500,7 +500,7 @@
prev_cells[i] = cell
# apply attention using the last layer's hidden state
- if self.attention is not None:
+ if self.attention is None:
assert attn_scores is not None
out, attn_scores[:, j, :] = self.attention(hidden, encoder_outs, encoder_padding_mask)
else:
Mutant 2371
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -508,7 +508,7 @@
out = self.dropout_out_module(out)
# input feeding
- if input_feed is not None:
+ if input_feed is None:
input_feed = out
# save final output
Mutant 2372
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -509,7 +509,7 @@
# input feeding
if input_feed is not None:
- input_feed = out
+ input_feed = None
# save final output
outs.append(out)
Mutant 2373
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -515,7 +515,7 @@
outs.append(out)
# Stack all the necessary tensors together and store
- prev_hiddens_tensor = torch.stack(prev_hiddens)
+ prev_hiddens_tensor = None
prev_cells_tensor = torch.stack(prev_cells)
cache_state = torch.jit.annotate(
Dict[str, Optional[Tensor]],
Mutant 2374
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -516,7 +516,7 @@
# Stack all the necessary tensors together and store
prev_hiddens_tensor = torch.stack(prev_hiddens)
- prev_cells_tensor = torch.stack(prev_cells)
+ prev_cells_tensor = None
cache_state = torch.jit.annotate(
Dict[str, Optional[Tensor]],
{
Mutant 2375
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -520,7 +520,7 @@
cache_state = torch.jit.annotate(
Dict[str, Optional[Tensor]],
{
- "prev_hiddens": prev_hiddens_tensor,
+ "XXprev_hiddensXX": prev_hiddens_tensor,
"prev_cells": prev_cells_tensor,
"input_feed": input_feed,
}
Mutant 2376
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -521,7 +521,7 @@
Dict[str, Optional[Tensor]],
{
"prev_hiddens": prev_hiddens_tensor,
- "prev_cells": prev_cells_tensor,
+ "XXprev_cellsXX": prev_cells_tensor,
"input_feed": input_feed,
}
)
Mutant 2377
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -522,7 +522,7 @@
{
"prev_hiddens": prev_hiddens_tensor,
"prev_cells": prev_cells_tensor,
- "input_feed": input_feed,
+ "XXinput_feedXX": input_feed,
}
)
self.set_incremental_state(incremental_state, 'cached_state', cache_state)
Mutant 2378
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -525,7 +525,7 @@
"input_feed": input_feed,
}
)
- self.set_incremental_state(incremental_state, 'cached_state', cache_state)
+ self.set_incremental_state(incremental_state, 'XXcached_stateXX', cache_state)
# collect outputs across time steps
x = torch.cat(outs, dim=0).view(seqlen, bsz, self.hidden_size)
Mutant 2379
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -528,7 +528,7 @@
self.set_incremental_state(incremental_state, 'cached_state', cache_state)
# collect outputs across time steps
- x = torch.cat(outs, dim=0).view(seqlen, bsz, self.hidden_size)
+ x = torch.cat(outs, dim=1).view(seqlen, bsz, self.hidden_size)
# T x B x C -> B x T x C
x = x.transpose(1, 0)
Mutant 2382
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -531,7 +531,7 @@
x = torch.cat(outs, dim=0).view(seqlen, bsz, self.hidden_size)
# T x B x C -> B x T x C
- x = x.transpose(1, 0)
+ x = x.transpose(1, 1)
if hasattr(self, 'additional_fc') and self.adaptive_softmax is None:
x = self.additional_fc(x)
Mutant 2384
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -533,7 +533,7 @@
# T x B x C -> B x T x C
x = x.transpose(1, 0)
- if hasattr(self, 'additional_fc') and self.adaptive_softmax is None:
+ if hasattr(self, 'XXadditional_fcXX') and self.adaptive_softmax is None:
x = self.additional_fc(x)
x = self.dropout_out_module(x)
# srclen x tgtlen x bsz -> bsz x tgtlen x srclen
Mutant 2385
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -533,7 +533,7 @@
# T x B x C -> B x T x C
x = x.transpose(1, 0)
- if hasattr(self, 'additional_fc') and self.adaptive_softmax is None:
+ if hasattr(self, 'additional_fc') and self.adaptive_softmax is not None:
x = self.additional_fc(x)
x = self.dropout_out_module(x)
# srclen x tgtlen x bsz -> bsz x tgtlen x srclen
Mutant 2389
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -537,7 +537,7 @@
x = self.additional_fc(x)
x = self.dropout_out_module(x)
# srclen x tgtlen x bsz -> bsz x tgtlen x srclen
- if not self.training and self.need_attn and self.attention is not None:
+ if not self.training or self.need_attn and self.attention is not None:
assert attn_scores is not None
attn_scores = attn_scores.transpose(0, 2)
else:
Mutant 2391
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -539,7 +539,7 @@
# srclen x tgtlen x bsz -> bsz x tgtlen x srclen
if not self.training and self.need_attn and self.attention is not None:
assert attn_scores is not None
- attn_scores = attn_scores.transpose(0, 2)
+ attn_scores = attn_scores.transpose(1, 2)
else:
attn_scores = None
return x, attn_scores
Mutant 2394
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -546,7 +546,7 @@
def output_layer(self, x):
"""Project features to the vocabulary size."""
- if self.adaptive_softmax is None:
+ if self.adaptive_softmax is not None:
if self.share_input_output_embed:
x = F.linear(x, self.embed_tokens.weight)
else:
Mutant 2396
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -595,7 +595,7 @@
"""Maximum output length supported by the decoder."""
return self.max_target_positions
- def make_generation_fast_(self, need_attn=False, **kwargs):
+ def make_generation_fast_(self, need_attn=True, **kwargs):
self.need_attn = need_attn
Mutant 2398
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -601,7 +601,7 @@
def Embedding(num_embeddings, embedding_dim, padding_idx):
m = nn.Embedding(num_embeddings, embedding_dim, padding_idx=padding_idx)
- nn.init.uniform_(m.weight, -0.1, 0.1)
+ nn.init.uniform_(m.weight, +0.1, 0.1)
nn.init.constant_(m.weight[padding_idx], 0)
return m
Mutant 2399
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -601,7 +601,7 @@
def Embedding(num_embeddings, embedding_dim, padding_idx):
m = nn.Embedding(num_embeddings, embedding_dim, padding_idx=padding_idx)
- nn.init.uniform_(m.weight, -0.1, 0.1)
+ nn.init.uniform_(m.weight, -1.1, 0.1)
nn.init.constant_(m.weight[padding_idx], 0)
return m
Mutant 2400
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -601,7 +601,7 @@
def Embedding(num_embeddings, embedding_dim, padding_idx):
m = nn.Embedding(num_embeddings, embedding_dim, padding_idx=padding_idx)
- nn.init.uniform_(m.weight, -0.1, 0.1)
+ nn.init.uniform_(m.weight, -0.1, 1.1)
nn.init.constant_(m.weight[padding_idx], 0)
return m
Mutant 2401
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -602,7 +602,7 @@
def Embedding(num_embeddings, embedding_dim, padding_idx):
m = nn.Embedding(num_embeddings, embedding_dim, padding_idx=padding_idx)
nn.init.uniform_(m.weight, -0.1, 0.1)
- nn.init.constant_(m.weight[padding_idx], 0)
+ nn.init.constant_(m.weight[padding_idx], 1)
return m
Mutant 2403
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -609,7 +609,7 @@
def LSTM(input_size, hidden_size, **kwargs):
m = nn.LSTM(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'XXweightXX' in name or 'bias' in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2404
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -609,7 +609,7 @@
def LSTM(input_size, hidden_size, **kwargs):
m = nn.LSTM(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'weight' not in name or 'bias' in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2405
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -609,7 +609,7 @@
def LSTM(input_size, hidden_size, **kwargs):
m = nn.LSTM(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'weight' in name or 'XXbiasXX' in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2406
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -609,7 +609,7 @@
def LSTM(input_size, hidden_size, **kwargs):
m = nn.LSTM(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'weight' in name or 'bias' not in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2407
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -609,7 +609,7 @@
def LSTM(input_size, hidden_size, **kwargs):
m = nn.LSTM(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'weight' in name and 'bias' in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2408
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -610,7 +610,7 @@
m = nn.LSTM(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
if 'weight' in name or 'bias' in name:
- param.data.uniform_(-0.1, 0.1)
+ param.data.uniform_(+0.1, 0.1)
return m
Mutant 2409
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -610,7 +610,7 @@
m = nn.LSTM(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
if 'weight' in name or 'bias' in name:
- param.data.uniform_(-0.1, 0.1)
+ param.data.uniform_(-1.1, 0.1)
return m
Mutant 2410
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -610,7 +610,7 @@
m = nn.LSTM(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
if 'weight' in name or 'bias' in name:
- param.data.uniform_(-0.1, 0.1)
+ param.data.uniform_(-0.1, 1.1)
return m
Mutant 2412
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -617,7 +617,7 @@
def LSTMCell(input_size, hidden_size, **kwargs):
m = nn.LSTMCell(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'XXweightXX' in name or 'bias' in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2413
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -617,7 +617,7 @@
def LSTMCell(input_size, hidden_size, **kwargs):
m = nn.LSTMCell(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'weight' not in name or 'bias' in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2414
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -617,7 +617,7 @@
def LSTMCell(input_size, hidden_size, **kwargs):
m = nn.LSTMCell(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'weight' in name or 'XXbiasXX' in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2415
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -617,7 +617,7 @@
def LSTMCell(input_size, hidden_size, **kwargs):
m = nn.LSTMCell(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'weight' in name or 'bias' not in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2416
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -617,7 +617,7 @@
def LSTMCell(input_size, hidden_size, **kwargs):
m = nn.LSTMCell(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
- if 'weight' in name or 'bias' in name:
+ if 'weight' in name and 'bias' in name:
param.data.uniform_(-0.1, 0.1)
return m
Mutant 2417
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -618,7 +618,7 @@
m = nn.LSTMCell(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
if 'weight' in name or 'bias' in name:
- param.data.uniform_(-0.1, 0.1)
+ param.data.uniform_(+0.1, 0.1)
return m
Mutant 2418
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -618,7 +618,7 @@
m = nn.LSTMCell(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
if 'weight' in name or 'bias' in name:
- param.data.uniform_(-0.1, 0.1)
+ param.data.uniform_(-1.1, 0.1)
return m
Mutant 2419
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -618,7 +618,7 @@
m = nn.LSTMCell(input_size, hidden_size, **kwargs)
for name, param in m.named_parameters():
if 'weight' in name or 'bias' in name:
- param.data.uniform_(-0.1, 0.1)
+ param.data.uniform_(-0.1, 1.1)
return m
Mutant 2420
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -622,7 +622,7 @@
return m
-def Linear(in_features, out_features, bias=True, dropout=0.):
+def Linear(in_features, out_features, bias=False, dropout=0.):
"""Linear layer (input: N x T x C)"""
m = nn.Linear(in_features, out_features, bias=bias)
m.weight.data.uniform_(-0.1, 0.1)
Mutant 2421
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -622,7 +622,7 @@
return m
-def Linear(in_features, out_features, bias=True, dropout=0.):
+def Linear(in_features, out_features, bias=True, dropout=1.0):
"""Linear layer (input: N x T x C)"""
m = nn.Linear(in_features, out_features, bias=bias)
m.weight.data.uniform_(-0.1, 0.1)
Mutant 2423
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -625,7 +625,7 @@
def Linear(in_features, out_features, bias=True, dropout=0.):
"""Linear layer (input: N x T x C)"""
m = nn.Linear(in_features, out_features, bias=bias)
- m.weight.data.uniform_(-0.1, 0.1)
+ m.weight.data.uniform_(+0.1, 0.1)
if bias:
m.bias.data.uniform_(-0.1, 0.1)
return m
Mutant 2424
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -625,7 +625,7 @@
def Linear(in_features, out_features, bias=True, dropout=0.):
"""Linear layer (input: N x T x C)"""
m = nn.Linear(in_features, out_features, bias=bias)
- m.weight.data.uniform_(-0.1, 0.1)
+ m.weight.data.uniform_(-1.1, 0.1)
if bias:
m.bias.data.uniform_(-0.1, 0.1)
return m
Mutant 2425
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -625,7 +625,7 @@
def Linear(in_features, out_features, bias=True, dropout=0.):
"""Linear layer (input: N x T x C)"""
m = nn.Linear(in_features, out_features, bias=bias)
- m.weight.data.uniform_(-0.1, 0.1)
+ m.weight.data.uniform_(-0.1, 1.1)
if bias:
m.bias.data.uniform_(-0.1, 0.1)
return m
Mutant 2426
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -627,7 +627,7 @@
m = nn.Linear(in_features, out_features, bias=bias)
m.weight.data.uniform_(-0.1, 0.1)
if bias:
- m.bias.data.uniform_(-0.1, 0.1)
+ m.bias.data.uniform_(+0.1, 0.1)
return m
Mutant 2427
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -627,7 +627,7 @@
m = nn.Linear(in_features, out_features, bias=bias)
m.weight.data.uniform_(-0.1, 0.1)
if bias:
- m.bias.data.uniform_(-0.1, 0.1)
+ m.bias.data.uniform_(-1.1, 0.1)
return m
Mutant 2428
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -627,7 +627,7 @@
m = nn.Linear(in_features, out_features, bias=bias)
m.weight.data.uniform_(-0.1, 0.1)
if bias:
- m.bias.data.uniform_(-0.1, 0.1)
+ m.bias.data.uniform_(-0.1, 1.1)
return m
Mutant 2430
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -631,7 +631,7 @@
return m
-@register_model_architecture('lstm', 'lstm')
+@register_model_architecture('lstm', 'XXlstmXX')
def base_architecture(args):
args.dropout = getattr(args, 'dropout', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
Mutant 2431
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -630,8 +630,6 @@
m.bias.data.uniform_(-0.1, 0.1)
return m
-
-@register_model_architecture('lstm', 'lstm')
def base_architecture(args):
args.dropout = getattr(args, 'dropout', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
Mutant 2432
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -633,7 +633,7 @@
@register_model_architecture('lstm', 'lstm')
def base_architecture(args):
- args.dropout = getattr(args, 'dropout', 0.1)
+ args.dropout = getattr(args, 'XXdropoutXX', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
Mutant 2433
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -633,7 +633,7 @@
@register_model_architecture('lstm', 'lstm')
def base_architecture(args):
- args.dropout = getattr(args, 'dropout', 0.1)
+ args.dropout = getattr(args, 'dropout', 1.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
Mutant 2435
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -634,7 +634,7 @@
@register_model_architecture('lstm', 'lstm')
def base_architecture(args):
args.dropout = getattr(args, 'dropout', 0.1)
- args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
+ args.encoder_embed_dim = getattr(args, 'XXencoder_embed_dimXX', 512)
args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
Mutant 2436
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -634,7 +634,7 @@
@register_model_architecture('lstm', 'lstm')
def base_architecture(args):
args.dropout = getattr(args, 'dropout', 0.1)
- args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
+ args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 513)
args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
Mutant 2438
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -635,7 +635,7 @@
def base_architecture(args):
args.dropout = getattr(args, 'dropout', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
- args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
+ args.encoder_embed_path = getattr(args, 'XXencoder_embed_pathXX', None)
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
args.encoder_layers = getattr(args, 'encoder_layers', 1)
Mutant 2439
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -635,7 +635,7 @@
def base_architecture(args):
args.dropout = getattr(args, 'dropout', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
- args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
+ args.encoder_embed_path = None
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
args.encoder_layers = getattr(args, 'encoder_layers', 1)
Mutant 2440
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -636,7 +636,7 @@
args.dropout = getattr(args, 'dropout', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
- args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
+ args.encoder_freeze_embed = getattr(args, 'XXencoder_freeze_embedXX', False)
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
args.encoder_layers = getattr(args, 'encoder_layers', 1)
args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
Mutant 2441
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -636,7 +636,7 @@
args.dropout = getattr(args, 'dropout', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
- args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
+ args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', True)
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
args.encoder_layers = getattr(args, 'encoder_layers', 1)
args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
Mutant 2442
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -636,7 +636,7 @@
args.dropout = getattr(args, 'dropout', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
- args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
+ args.encoder_freeze_embed = None
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
args.encoder_layers = getattr(args, 'encoder_layers', 1)
args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
Mutant 2443
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -637,7 +637,7 @@
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
- args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
+ args.encoder_hidden_size = getattr(args, 'XXencoder_hidden_sizeXX', args.encoder_embed_dim)
args.encoder_layers = getattr(args, 'encoder_layers', 1)
args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
Mutant 2445
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -638,7 +638,7 @@
args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
- args.encoder_layers = getattr(args, 'encoder_layers', 1)
+ args.encoder_layers = getattr(args, 'XXencoder_layersXX', 1)
args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
Mutant 2448
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -639,7 +639,7 @@
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
args.encoder_layers = getattr(args, 'encoder_layers', 1)
- args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
+ args.encoder_bidirectional = getattr(args, 'XXencoder_bidirectionalXX', False)
args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
Mutant 2449
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -639,7 +639,7 @@
args.encoder_freeze_embed = getattr(args, 'encoder_freeze_embed', False)
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
args.encoder_layers = getattr(args, 'encoder_layers', 1)
- args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
+ args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', True)
args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
Mutant 2451
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -640,7 +640,7 @@
args.encoder_hidden_size = getattr(args, 'encoder_hidden_size', args.encoder_embed_dim)
args.encoder_layers = getattr(args, 'encoder_layers', 1)
args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
- args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
+ args.encoder_dropout_in = getattr(args, 'XXencoder_dropout_inXX', args.dropout)
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
Mutant 2453
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -641,7 +641,7 @@
args.encoder_layers = getattr(args, 'encoder_layers', 1)
args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
- args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
+ args.encoder_dropout_out = getattr(args, 'XXencoder_dropout_outXX', args.dropout)
args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
Mutant 2455
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -642,7 +642,7 @@
args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
- args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
+ args.decoder_embed_dim = getattr(args, 'XXdecoder_embed_dimXX', 512)
args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
Mutant 2456
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -642,7 +642,7 @@
args.encoder_bidirectional = getattr(args, 'encoder_bidirectional', False)
args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
- args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
+ args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 513)
args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
Mutant 2458
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -643,7 +643,7 @@
args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
- args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
+ args.decoder_embed_path = getattr(args, 'XXdecoder_embed_pathXX', None)
args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
args.decoder_layers = getattr(args, 'decoder_layers', 1)
Mutant 2459
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -643,7 +643,7 @@
args.encoder_dropout_in = getattr(args, 'encoder_dropout_in', args.dropout)
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
- args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
+ args.decoder_embed_path = None
args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
args.decoder_layers = getattr(args, 'decoder_layers', 1)
Mutant 2460
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -644,7 +644,7 @@
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
- args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
+ args.decoder_freeze_embed = getattr(args, 'XXdecoder_freeze_embedXX', False)
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
args.decoder_layers = getattr(args, 'decoder_layers', 1)
args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 512)
Mutant 2462
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -644,7 +644,7 @@
args.encoder_dropout_out = getattr(args, 'encoder_dropout_out', args.dropout)
args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
- args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
+ args.decoder_freeze_embed = None
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
args.decoder_layers = getattr(args, 'decoder_layers', 1)
args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 512)
Mutant 2463
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -645,7 +645,7 @@
args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', 512)
args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
- args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
+ args.decoder_hidden_size = getattr(args, 'XXdecoder_hidden_sizeXX', args.decoder_embed_dim)
args.decoder_layers = getattr(args, 'decoder_layers', 1)
args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 512)
args.decoder_attention = getattr(args, 'decoder_attention', '1')
Mutant 2465
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -646,7 +646,7 @@
args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
- args.decoder_layers = getattr(args, 'decoder_layers', 1)
+ args.decoder_layers = getattr(args, 'XXdecoder_layersXX', 1)
args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 512)
args.decoder_attention = getattr(args, 'decoder_attention', '1')
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
Mutant 2468
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -647,7 +647,7 @@
args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
args.decoder_layers = getattr(args, 'decoder_layers', 1)
- args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 512)
+ args.decoder_out_embed_dim = getattr(args, 'XXdecoder_out_embed_dimXX', 512)
args.decoder_attention = getattr(args, 'decoder_attention', '1')
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
Mutant 2469
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -647,7 +647,7 @@
args.decoder_freeze_embed = getattr(args, 'decoder_freeze_embed', False)
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
args.decoder_layers = getattr(args, 'decoder_layers', 1)
- args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 512)
+ args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 513)
args.decoder_attention = getattr(args, 'decoder_attention', '1')
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
Mutant 2471
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -648,7 +648,7 @@
args.decoder_hidden_size = getattr(args, 'decoder_hidden_size', args.decoder_embed_dim)
args.decoder_layers = getattr(args, 'decoder_layers', 1)
args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 512)
- args.decoder_attention = getattr(args, 'decoder_attention', '1')
+ args.decoder_attention = getattr(args, 'XXdecoder_attentionXX', '1')
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
Mutant 2474
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -649,7 +649,7 @@
args.decoder_layers = getattr(args, 'decoder_layers', 1)
args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 512)
args.decoder_attention = getattr(args, 'decoder_attention', '1')
- args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
+ args.decoder_dropout_in = getattr(args, 'XXdecoder_dropout_inXX', args.dropout)
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
Mutant 2476
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -650,7 +650,7 @@
args.decoder_out_embed_dim = getattr(args, 'decoder_out_embed_dim', 512)
args.decoder_attention = getattr(args, 'decoder_attention', '1')
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
- args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
+ args.decoder_dropout_out = getattr(args, 'XXdecoder_dropout_outXX', args.dropout)
args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
Mutant 2478
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -651,7 +651,7 @@
args.decoder_attention = getattr(args, 'decoder_attention', '1')
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
- args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
+ args.share_decoder_input_output_embed = getattr(args, 'XXshare_decoder_input_output_embedXX', False)
args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
Mutant 2479
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -651,7 +651,7 @@
args.decoder_attention = getattr(args, 'decoder_attention', '1')
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
- args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
+ args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', True)
args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
Mutant 2481
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -652,7 +652,7 @@
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
- args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
+ args.share_all_embeddings = getattr(args, 'XXshare_all_embeddingsXX', False)
args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
Mutant 2482
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -652,7 +652,7 @@
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
- args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
+ args.share_all_embeddings = getattr(args, 'share_all_embeddings', True)
args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
Mutant 2483
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -652,7 +652,7 @@
args.decoder_dropout_in = getattr(args, 'decoder_dropout_in', args.dropout)
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
- args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
+ args.share_all_embeddings = None
args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
Mutant 2484
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -653,7 +653,7 @@
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
- args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
+ args.adaptive_softmax_cutoff = getattr(args, 'XXadaptive_softmax_cutoffXX', '10000,50000,200000')
@register_model_architecture('lstm', 'lstm_wiseman_iwslt_de_en')
Mutant 2485
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -653,7 +653,7 @@
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
- args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
+ args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', 'XX10000,50000,200000XX')
@register_model_architecture('lstm', 'lstm_wiseman_iwslt_de_en')
Mutant 2486
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -653,7 +653,7 @@
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
- args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
+ args.adaptive_softmax_cutoff = None
@register_model_architecture('lstm', 'lstm_wiseman_iwslt_de_en')
Mutant 2488
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -656,7 +656,7 @@
args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
-@register_model_architecture('lstm', 'lstm_wiseman_iwslt_de_en')
+@register_model_architecture('lstm', 'XXlstm_wiseman_iwslt_de_enXX')
def lstm_wiseman_iwslt_de_en(args):
args.dropout = getattr(args, 'dropout', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 256)
Mutant 2489
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -655,8 +655,6 @@
args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', '10000,50000,200000')
-
-@register_model_architecture('lstm', 'lstm_wiseman_iwslt_de_en')
def lstm_wiseman_iwslt_de_en(args):
args.dropout = getattr(args, 'dropout', 0.1)
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 256)
Mutant 2491
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -669,7 +669,7 @@
base_architecture(args)
-@register_model_architecture('lstm', 'lstm_luong_wmt_en_de')
+@register_model_architecture('lstm', 'XXlstm_luong_wmt_en_deXX')
def lstm_luong_wmt_en_de(args):
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 1000)
args.encoder_layers = getattr(args, 'encoder_layers', 4)
Mutant 2492
--- fairseq/models/lstm.py
+++ fairseq/models/lstm.py
@@ -668,8 +668,6 @@
args.decoder_dropout_out = getattr(args, 'decoder_dropout_out', args.dropout)
base_architecture(args)
-
-@register_model_architecture('lstm', 'lstm_luong_wmt_en_de')
def lstm_luong_wmt_en_de(args):
args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 1000)
args.encoder_layers = getattr(args, 'encoder_layers', 4)