duplicated block id: 1
size: 213 cleaned lines of code
in 2 files:
 - fairseq_cli/train.py (25:306)
 - train.py (25:306)

duplicated block id: 2
size: 191 cleaned lines of code
in 2 files:
 - fairseq_cli/preprocess.py (26:254)
 - preprocess.py (26:254)

duplicated block id: 3
size: 157 cleaned lines of code
in 2 files:
 - eval_lm.py (22:225)
 - fairseq_cli/eval_lm.py (22:225)

duplicated block id: 4
size: 140 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (19:199)
 - generate.py (19:199)

duplicated block id: 5
size: 134 cleaned lines of code
in 2 files:
 - fairseq_cli/interactive.py (21:195)
 - interactive.py (21:195)

duplicated block id: 6
size: 90 cleaned lines of code
in 2 files:
 - fairseq/modules/masked_multihead_attention.py (58:171)
 - fairseq/modules/multihead_attention.py (131:242)

duplicated block id: 7
size: 59 cleaned lines of code
in 2 files:
 - scripts/oracle.py (9:86)
 - scripts/oracle_s2s.py (9:86)

duplicated block id: 8
size: 52 cleaned lines of code
in 2 files:
 - fairseq_cli/setup.py (13:68)
 - setup.py (13:68)

duplicated block id: 9
size: 31 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (695:728)
 - fairseq/models/disco_transformer.py (342:375)

duplicated block id: 10
size: 28 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (123:150)
 - fairseq/models/transformer.py (76:103)

duplicated block id: 11
size: 25 cleaned lines of code
in 2 files:
 - fairseq/modules/sparse_transformer_sentence_encoder.py (19:43)
 - fairseq/modules/transformer_sentence_encoder.py (69:93)

duplicated block id: 12
size: 23 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (198:223)
 - fairseq/models/disco_transformer.py (47:72)

duplicated block id: 13
size: 23 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (470:493)
 - fairseq/models/disco_transformer.py (313:338)

duplicated block id: 14
size: 20 cleaned lines of code
in 2 files:
 - fairseq/modules/masked_multihead_attention.py (179:205)
 - fairseq/modules/multihead_attention.py (240:266)

duplicated block id: 15
size: 17 cleaned lines of code
in 2 files:
 - fairseq/data/language_pair_dataset.py (105:121)
 - fairseq/data/language_pair_self_dataset_mask.py (130:146)

duplicated block id: 16
size: 17 cleaned lines of code
in 2 files:
 - fairseq/optim/fp16_optimizer.py (69:87)
 - fairseq/optim/fp16_optimizer.py (240:258)

duplicated block id: 17
size: 13 cleaned lines of code
in 2 files:
 - fairseq/modules/dynamic_convolution.py (200:216)
 - fairseq/modules/lightweight_convolution.py (216:232)

duplicated block id: 18
size: 13 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (366:383)
 - fairseq/models/disco_transformer.py (251:268)

duplicated block id: 19
size: 12 cleaned lines of code
in 2 files:
 - fairseq/modules/sparse_transformer_sentence_encoder_layer.py (17:28)
 - fairseq/modules/transformer_sentence_encoder_layer.py (25:36)

duplicated block id: 20
size: 12 cleaned lines of code
in 2 files:
 - fairseq/optim/fp16_optimizer.py (180:194)
 - fairseq/optim/fp16_optimizer.py (352:367)

duplicated block id: 21
size: 12 cleaned lines of code
in 2 files:
 - fairseq/tasks/translation.py (116:127)
 - fairseq/tasks/translation_self.py (47:58)

duplicated block id: 22
size: 10 cleaned lines of code
in 2 files:
 - fairseq/optim/nag.py (25:42)
 - fairseq/optim/sgd.py (24:41)

duplicated block id: 23
size: 10 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (66:77)
 - generate_disco.py (65:76)

duplicated block id: 24
size: 10 cleaned lines of code
in 2 files:
 - fairseq/modules/sparse_transformer_sentence_encoder.py (61:70)
 - fairseq/modules/transformer_sentence_encoder.py (132:141)

duplicated block id: 25
size: 10 cleaned lines of code
in 2 files:
 - fairseq/modules/dynamic_convolution.py (108:118)
 - fairseq/modules/dynamic_convolution.py (162:171)

duplicated block id: 26
size: 10 cleaned lines of code
in 2 files:
 - generate.py (66:77)
 - generate_disco.py (65:76)

duplicated block id: 27
size: 10 cleaned lines of code
in 2 files:
 - fairseq/optim/adafactor.py (130:146)
 - fairseq/optim/adam.py (96:112)

duplicated block id: 28
size: 9 cleaned lines of code
in 2 files:
 - fairseq_cli/interactive.py (92:100)
 - generate.py (54:62)

duplicated block id: 29
size: 9 cleaned lines of code
in 2 files:
 - generate.py (54:62)
 - interactive.py (92:100)

duplicated block id: 30
size: 9 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (54:62)
 - interactive.py (92:100)

duplicated block id: 31
size: 9 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (54:62)
 - fairseq_cli/interactive.py (92:100)

duplicated block id: 32
size: 9 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (493:503)
 - fairseq/models/transformer.py (736:746)

duplicated block id: 33
size: 9 cleaned lines of code
in 2 files:
 - fairseq/modules/dynamic_convolution.py (39:48)
 - fairseq/modules/lightweight_convolution.py (98:107)

duplicated block id: 34
size: 9 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (347:359)
 - fairseq/models/transformer.py (412:426)

duplicated block id: 35
size: 9 cleaned lines of code
in 2 files:
 - fairseq/modules/dynamic_convolution.py (123:131)
 - fairseq/modules/lightweight_convolution.py (162:170)

duplicated block id: 36
size: 9 cleaned lines of code
in 2 files:
 - fairseq/optim/lr_scheduler/fixed_schedule.py (47:59)
 - fairseq/optim/lr_scheduler/polynomial_decay_schedule.py (50:62)

duplicated block id: 37
size: 9 cleaned lines of code
in 2 files:
 - fairseq/modules/masked_multihead_attention.py (47:55)
 - fairseq/modules/multihead_attention.py (120:128)

duplicated block id: 38
size: 9 cleaned lines of code
in 2 files:
 - fairseq/strategies/easy_first.py (27:37)
 - fairseq/strategies/mask_predict.py (24:34)

duplicated block id: 39
size: 8 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (478:486)
 - fairseq/models/transformer.py (717:725)

duplicated block id: 40
size: 8 cleaned lines of code
in 2 files:
 - fairseq/optim/adam.py (96:110)
 - fairseq/optim/lamb.py (87:100)

duplicated block id: 41
size: 8 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (596:612)
 - fairseq/models/transformer.py (256:274)

duplicated block id: 42
size: 8 cleaned lines of code
in 2 files:
 - fairseq/optim/adam.py (92:107)
 - fairseq/optim/nag.py (50:65)

duplicated block id: 43
size: 8 cleaned lines of code
in 2 files:
 - fairseq_cli/interactive.py (164:171)
 - generate.py (140:147)

duplicated block id: 44
size: 8 cleaned lines of code
in 2 files:
 - fairseq/criterions/label_smoothed_cross_entropy.py (65:73)
 - fairseq/criterions/label_smoothed_length_cross_entropy.py (47:55)

duplicated block id: 45
size: 8 cleaned lines of code
in 2 files:
 - fairseq/models/transformer.py (574:581)
 - fairseq/models/transformer.py (724:731)

duplicated block id: 46
size: 8 cleaned lines of code
in 2 files:
 - generate.py (140:147)
 - interactive.py (164:171)

duplicated block id: 47
size: 8 cleaned lines of code
in 2 files:
 - fairseq/models/disco_transformer.py (323:331)
 - fairseq/models/transformer.py (717:725)

duplicated block id: 48
size: 8 cleaned lines of code
in 2 files:
 - fairseq/tasks/translation.py (130:143)
 - fairseq/tasks/translation_self.py (72:84)

duplicated block id: 49
size: 8 cleaned lines of code
in 2 files:
 - fairseq/tasks/translation.py (196:208)
 - fairseq/tasks/translation_self.py (193:205)

duplicated block id: 50
size: 8 cleaned lines of code
in 2 files:
 - fairseq/modules/dynamic_convolution.py (158:165)
 - fairseq/modules/lightweight_convolution.py (190:197)

duplicated block id: 51
size: 8 cleaned lines of code
in 2 files:
 - generate_disco.py (113:120)
 - generate_disco.py (142:149)

duplicated block id: 52
size: 8 cleaned lines of code
in 2 files:
 - fairseq/optim/adafactor.py (130:144)
 - fairseq/optim/lamb.py (87:100)

duplicated block id: 53
size: 8 cleaned lines of code
in 2 files:
 - fairseq/criterions/cross_entropy.py (35:43)
 - fairseq/criterions/label_smoothed_cross_entropy.py (64:72)

duplicated block id: 54
size: 8 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (140:147)
 - interactive.py (164:171)

duplicated block id: 55
size: 8 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (140:147)
 - fairseq_cli/interactive.py (164:171)

duplicated block id: 56
size: 8 cleaned lines of code
in 2 files:
 - fairseq/optim/lr_scheduler/fixed_schedule.py (15:25)
 - fairseq/optim/lr_scheduler/polynomial_decay_schedule.py (15:25)

duplicated block id: 57
size: 7 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (304:310)
 - fairseq/models/transformer.py (355:361)

duplicated block id: 58
size: 7 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (184:196)
 - fairseq/models/disco_transformer.py (28:40)

duplicated block id: 59
size: 7 cleaned lines of code
in 2 files:
 - fairseq/tasks/translation.py (59:67)
 - fairseq/tasks/translation_self.py (151:159)

duplicated block id: 60
size: 7 cleaned lines of code
in 2 files:
 - fairseq/models/transformer.py (525:532)
 - fairseq/models/transformer.py (620:627)

duplicated block id: 61
size: 7 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (632:652)
 - fairseq/models/transformer.py (499:521)

duplicated block id: 62
size: 7 cleaned lines of code
in 2 files:
 - fairseq/models/transformer.py (283:290)
 - fairseq/models/transformer.py (470:478)

duplicated block id: 63
size: 7 cleaned lines of code
in 2 files:
 - fairseq/utils.py (516:522)
 - fairseq/utils.py (558:564)

duplicated block id: 64
size: 7 cleaned lines of code
in 2 files:
 - generate_disco.py (55:61)
 - interactive.py (92:98)

duplicated block id: 65
size: 7 cleaned lines of code
in 2 files:
 - fairseq/modules/dynamic_convolution.py (190:198)
 - fairseq/modules/lightweight_convolution.py (206:213)

duplicated block id: 66
size: 7 cleaned lines of code
in 2 files:
 - fairseq_cli/interactive.py (92:98)
 - generate_disco.py (55:61)

duplicated block id: 67
size: 7 cleaned lines of code
in 2 files:
 - generate.py (54:60)
 - generate_disco.py (55:61)

duplicated block id: 68
size: 7 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (54:60)
 - generate_disco.py (55:61)

duplicated block id: 69
size: 7 cleaned lines of code
in 2 files:
 - fairseq/utils.py (436:442)
 - fairseq/utils.py (461:467)

duplicated block id: 70
size: 7 cleaned lines of code
in 2 files:
 - fairseq/modules/simple_sinusoidal_positional_embedding.py (22:29)
 - fairseq/modules/sinusoidal_positional_embedding.py (45:52)

duplicated block id: 71
size: 7 cleaned lines of code
in 2 files:
 - fairseq/criterions/cross_entropy.py (36:43)
 - fairseq/criterions/label_smoothed_length_cross_entropy.py (47:54)

duplicated block id: 72
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/lamb.py (34:47)
 - fairseq/optim/nag.py (26:39)

duplicated block id: 73
size: 6 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (495:500)
 - fairseq/models/transformer.py (584:589)

duplicated block id: 74
size: 6 cleaned lines of code
in 2 files:
 - fairseq/models/disco_transformer.py (348:353)
 - fairseq/models/transformer.py (771:776)

duplicated block id: 75
size: 6 cleaned lines of code
in 2 files:
 - fairseq/models/transformer.py (756:761)
 - fairseq/modules/dynamic_convolution.py (16:21)

duplicated block id: 76
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/adagrad.py (23:36)
 - fairseq/optim/adam.py (39:52)

duplicated block id: 77
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/lamb.py (34:47)
 - fairseq/optim/sgd.py (25:38)

duplicated block id: 78
size: 6 cleaned lines of code
in 2 files:
 - fairseq/modules/lightweight_convolution.py (46:52)
 - fairseq/modules/lightweight_convolution.py (115:121)

duplicated block id: 79
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/adam.py (116:126)
 - fairseq/optim/adam.py (275:285)

duplicated block id: 80
size: 6 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (19:24)
 - generate_disco.py (22:27)

duplicated block id: 81
size: 6 cleaned lines of code
in 2 files:
 - fairseq/criterions/label_smoothed_cross_entropy.py (79:85)
 - fairseq/criterions/label_smoothed_length_cross_entropy.py (72:78)

duplicated block id: 82
size: 6 cleaned lines of code
in 2 files:
 - fairseq/modules/sparse_transformer_sentence_encoder.py (43:49)
 - fairseq/modules/sparse_transformer_sentence_encoder_layer.py (28:34)

duplicated block id: 83
size: 6 cleaned lines of code
in 2 files:
 - fairseq/modules/adaptive_input.py (26:32)
 - fairseq/modules/adaptive_softmax.py (60:66)

duplicated block id: 84
size: 6 cleaned lines of code
in 2 files:
 - fairseq/data/indexed_dataset.py (165:172)
 - fairseq/data/indexed_dataset.py (272:279)

duplicated block id: 85
size: 6 cleaned lines of code
in 2 files:
 - fairseq/modules/dynamic_convolution.py (106:111)
 - fairseq/modules/lightweight_convolution.py (154:159)

duplicated block id: 86
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/adam.py (39:52)
 - fairseq/optim/lamb.py (34:47)

duplicated block id: 87
size: 6 cleaned lines of code
in 2 files:
 - fairseq/data/language_pair_dataset.py (151:184)
 - fairseq/data/language_pair_self_dataset_mask.py (221:249)

duplicated block id: 88
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/adagrad.py (23:36)
 - fairseq/optim/lamb.py (34:47)

duplicated block id: 89
size: 6 cleaned lines of code
in 2 files:
 - fairseq/data/language_pair_dataset.py (14:19)
 - fairseq/data/language_pair_self_dataset_mask.py (17:22)

duplicated block id: 90
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/adagrad.py (23:36)
 - fairseq/optim/sgd.py (25:38)

duplicated block id: 91
size: 6 cleaned lines of code
in 2 files:
 - fairseq/models/bert_seq2seq.py (701:706)
 - fairseq/models/transformer.py (771:776)

duplicated block id: 92
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/adam.py (39:52)
 - fairseq/optim/nag.py (26:39)

duplicated block id: 93
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/adam.py (39:52)
 - fairseq/optim/sgd.py (25:38)

duplicated block id: 94
size: 6 cleaned lines of code
in 2 files:
 - fairseq/models/transformer.py (584:589)
 - fairseq/models/transformer.py (738:743)

duplicated block id: 95
size: 6 cleaned lines of code
in 2 files:
 - fairseq/optim/adagrad.py (23:36)
 - fairseq/optim/nag.py (26:39)

duplicated block id: 96
size: 6 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (46:51)
 - interactive.py (80:85)

duplicated block id: 97
size: 6 cleaned lines of code
in 2 files:
 - generate.py (46:51)
 - interactive.py (80:85)

duplicated block id: 98
size: 6 cleaned lines of code
in 2 files:
 - fairseq/models/transformer.py (294:300)
 - fairseq/models/transformer.py (492:499)

duplicated block id: 99
size: 6 cleaned lines of code
in 2 files:
 - generate.py (19:24)
 - generate_disco.py (22:27)

duplicated block id: 100
size: 6 cleaned lines of code
in 2 files:
 - fairseq/modules/masked_multihead_attention.py (38:43)
 - fairseq/modules/multihead_attention.py (111:116)

duplicated block id: 101
size: 6 cleaned lines of code
in 2 files:
 - fairseq_cli/interactive.py (80:85)
 - generate.py (46:51)

duplicated block id: 102
size: 6 cleaned lines of code
in 2 files:
 - fairseq_cli/generate.py (46:51)
 - fairseq_cli/interactive.py (80:85)