26 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (123:153, 9%)
  - fairseq/modules/multihead_attention.py (331:361, 5%)

17 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (331:349, 6%)
  - fairseq/modules/multihead_attention.py (585:603, 3%)

16 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (283:303, 5%)
  - fairseq/modules/multihead_attention.py (522:542, 3%)

10 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (67:77, 3%)
  - fairseq/modules/multihead_attention.py (48:59, 1%)

9 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (250:258, 3%)
  - fairseq/modules/multihead_attention.py (483:491, 1%)

9 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (192:200, 3%)
  - fairseq/modules/multihead_attention.py (414:422, 1%)

8 duplicated lines in:
  - fairseq/model_parallel/models/pipeline_parallel_transformer/layers.py (534:541, 1%)
  - fairseq/modules/transformer_layer.py (483:491, 1%)

8 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (37:44, 2%)
  - fairseq/modules/multihead_attention.py (26:33, 1%)

8 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (37:44, 2%)
  - fairseq/modules/sparse_multihead_attention.py (24:31, 8%)

8 duplicated lines in:
  - fairseq/model_parallel/models/pipeline_parallel_transformer/layers.py (334:346, 1%)
  - fairseq/modules/transformer_layer.py (149:161, 1%)

7 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (96:102, 2%)
  - fairseq/modules/multihead_attention.py (243:249, 1%)

6 duplicated lines in:
  - fairseq/model_parallel/models/pipeline_parallel_transformer/layers.py (595:600, 1%)
  - fairseq/modules/dynamic_convolution.py (61:66, 2%)

6 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (203:208, 2%)
  - fairseq/modules/multihead_attention.py (424:429, 1%)

6 duplicated lines in:
  - fairseq/model_parallel/modules/multihead_attention.py (170:177, 2%)
  - fairseq/modules/multihead_attention.py (395:402, 1%)