[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

fairseq/checkpoint_utils.py

@@ -18,16 +18,15 @@
 
 import numpy as np
 import torch
+from omegaconf import DictConfig, OmegaConf, open_dict
+
 from fairseq.data import data_utils
 from fairseq.dataclass.configs import CheckpointConfig
-from fairseq.dataclass.utils import (
-    convert_namespace_to_omegaconf,
-    overwrite_args_by_name,
-)
+from fairseq.dataclass.utils import (convert_namespace_to_omegaconf,
+                                     overwrite_args_by_name)
 from fairseq.distributed.fully_sharded_data_parallel import FSDP, has_FSDP
 from fairseq.file_io import PathManager
 from fairseq.models import FairseqDecoder, FairseqEncoder
-from omegaconf import DictConfig, OmegaConf, open_dict
 
 logger = logging.getLogger(__name__)
 
@@ -100,9 +99,9 @@ def is_better(a, b):
                 cfg.best_checkpoint_metric, val_loss, rand_sfx, suffix
             )
         ] = worst_best is None or is_better(val_loss, worst_best)
-    checkpoint_conds[
-        "checkpoint_last{}.pt".format(suffix)
-    ] = not cfg.no_last_checkpoints
+    checkpoint_conds["checkpoint_last{}.pt".format(suffix)] = (
+        not cfg.no_last_checkpoints
+    )
 
     extra_state = {
         "train_iterator": epoch_itr.state_dict(),
@@ -116,7 +115,9 @@ def is_better(a, b):
     # attributes
     if hasattr(trainer.task, "get_checkpoint_dict"):
         extra_state = {**extra_state, **trainer.task.get_checkpoint_dict()}
-        logger.info(f"State of {trainer.task.__class__.__name__} is ready to be persisted with the checkpoint")
+        logger.info(
+            f"State of {trainer.task.__class__.__name__} is ready to be persisted with the checkpoint"
+        )
 
     if hasattr(save_checkpoint, "best"):
         extra_state.update({"best": save_checkpoint.best})

fairseq/data/dictionary.py

@@ -8,6 +8,7 @@
 from multiprocessing import Pool
 
 import torch
+
 # from fairseq import utils
 # from fairseq.data import data_utils
 # from fairseq.file_chunker_utils import Chunker, find_offsets
@@ -358,7 +359,7 @@ def merge_result(counter):
             chunks = zip(offsets, offsets[1:])
             pool = Pool(processes=num_workers)
             results = []
-            for (start_offset, end_offset) in chunks:
+            for start_offset, end_offset in chunks:
                 results.append(
                     pool.apply_async(
                         Dictionary._add_file_to_dictionary_single_worker,

fairseq/data/encoders/__init__.py

@@ -9,7 +9,6 @@
 
 from fairseq import registry
 
-
 build_tokenizer, register_tokenizer, TOKENIZER_REGISTRY, _ = registry.setup_registry(
     "--tokenizer",
     default=None,

fairseq/data/fairseq_dataset.py

@@ -4,8 +4,10 @@
 # LICENSE file in the root directory of this source tree.
 
 import logging
+
 import numpy as np
 import torch.utils.data
+
 from fairseq.data import data_utils
 
 logger = logging.getLogger(__name__)

fairseq/data/iterators.py

@@ -15,8 +15,8 @@
 
 import numpy as np
 import torch
-from fairseq.data import data_utils
 
+from fairseq.data import data_utils
 
 logger = logging.getLogger(__name__)
 

fairseq/dataclass/__init__.py

@@ -6,7 +6,6 @@
 from .configs import FairseqDataclass
 from .constants import ChoiceEnum
 
-
 __all__ = [
     "FairseqDataclass",
     "ChoiceEnum",

fairseq/dataclass/configs.py

@@ -11,17 +11,15 @@
 import torch
 from omegaconf import II, MISSING
 
-from fairseq.dataclass.constants import (
-    DATASET_IMPL_CHOICES,
-    DDP_BACKEND_CHOICES,
-    DDP_COMM_HOOK_CHOICES,
-    GENERATION_CONSTRAINTS_CHOICES,
-    GENERATION_DECODING_FORMAT_CHOICES,
-    LOG_FORMAT_CHOICES,
-    PIPELINE_CHECKPOINT_CHOICES,
-    PRINT_ALIGNMENT_CHOICES,
-    ZERO_SHARDING_CHOICES,
-)
+from fairseq.dataclass.constants import (DATASET_IMPL_CHOICES,
+                                         DDP_BACKEND_CHOICES,
+                                         DDP_COMM_HOOK_CHOICES,
+                                         GENERATION_CONSTRAINTS_CHOICES,
+                                         GENERATION_DECODING_FORMAT_CHOICES,
+                                         LOG_FORMAT_CHOICES,
+                                         PIPELINE_CHECKPOINT_CHOICES,
+                                         PRINT_ALIGNMENT_CHOICES,
+                                         ZERO_SHARDING_CHOICES)
 
 
 @dataclass
@@ -108,7 +106,7 @@ class CommonConfig(FairseqDataclass):
             "help": "log progress every N batches (when progress bar is disabled)"
         },
     )
-    log_format: Optional[LOG_FORMAT_CHOICES] = field( # type: ignore
+    log_format: Optional[LOG_FORMAT_CHOICES] = field(  # type: ignore
         default=None, metadata={"help": "log format to use"}
     )
     log_file: Optional[str] = field(
@@ -298,10 +296,10 @@ class DistributedTrainingConfig(FairseqDataclass):
             "help": "do not spawn multiple processes even if multiple GPUs are visible"
         },
     )
-    ddp_backend: DDP_BACKEND_CHOICES = field( # type: ignore
+    ddp_backend: DDP_BACKEND_CHOICES = field(  # type: ignore
         default="pytorch_ddp", metadata={"help": "DistributedDataParallel backend"}
     )
-    ddp_comm_hook: DDP_COMM_HOOK_CHOICES = field( # type: ignore
+    ddp_comm_hook: DDP_COMM_HOOK_CHOICES = field(  # type: ignore
         default="none", metadata={"help": "communication hook"}
     )
     bucket_cap_mb: int = field(
@@ -428,11 +426,11 @@ class DistributedTrainingConfig(FairseqDataclass):
             "equal the length of the --pipeline-decoder-balance argument"
         },
     )
-    pipeline_checkpoint: PIPELINE_CHECKPOINT_CHOICES = field( # type: ignore
+    pipeline_checkpoint: PIPELINE_CHECKPOINT_CHOICES = field(  # type: ignore
         default="never",
         metadata={"help": "checkpointing mode for pipeline model parallelism"},
     )
-    zero_sharding: ZERO_SHARDING_CHOICES = field( # type: ignore
+    zero_sharding: ZERO_SHARDING_CHOICES = field(  # type: ignore
         default="none", metadata={"help": "ZeRO sharding"}
     )
     fp16: bool = II("common.fp16")
@@ -488,7 +486,7 @@ class DatasetConfig(FairseqDataclass):
             "help": "maximum sequence length in batch will be a multiplier of this value"
         },
     )
-    dataset_impl: Optional[DATASET_IMPL_CHOICES] = field( # type: ignore
+    dataset_impl: Optional[DATASET_IMPL_CHOICES] = field(  # type: ignore
         default=None, metadata={"help": "output dataset implementation"}
     )
     data_buffer_size: int = field(
@@ -921,7 +919,7 @@ class GenerationConfig(FairseqDataclass):
             "help": "sample from the smallest set whose cumulative probability mass exceeds p for next words"
         },
     )
-    constraints: Optional[GENERATION_CONSTRAINTS_CHOICES] = field( # type: ignore
+    constraints: Optional[GENERATION_CONSTRAINTS_CHOICES] = field(  # type: ignore
         default=None,
         metadata={
             "help": "enables lexically constrained decoding",
@@ -944,7 +942,7 @@ class GenerationConfig(FairseqDataclass):
         default=-1.0,
         metadata={"help": "strength of diversity penalty for Diverse Siblings Search"},
     )
-    print_alignment: Optional[PRINT_ALIGNMENT_CHOICES] = field( # type: ignore
+    print_alignment: Optional[PRINT_ALIGNMENT_CHOICES] = field(  # type: ignore
         default=None,
         metadata={
             "help": "if set, uses attention feedback to compute and print alignment to source tokens "
@@ -1012,7 +1010,7 @@ class GenerationConfig(FairseqDataclass):
         },
     )
     # special decoding format for advanced decoding.
-    decoding_format: Optional[GENERATION_DECODING_FORMAT_CHOICES] = field( # type: ignore
+    decoding_format: Optional[GENERATION_DECODING_FORMAT_CHOICES] = field(  # type: ignore
         default=None,
         metadata={"help": "special decoding format for advanced decoding."},
     )

fairseq/dataclass/utils.py

@@ -13,11 +13,12 @@
 from enum import Enum
 from typing import Any, Dict, List, Optional, Tuple, Type
 
-from fairseq.dataclass import FairseqDataclass
-from fairseq.dataclass.configs import FairseqConfig
 from hydra.core.global_hydra import GlobalHydra
 from hydra.experimental import compose, initialize
-from omegaconf import DictConfig, OmegaConf, open_dict, _utils
+from omegaconf import DictConfig, OmegaConf, _utils, open_dict
+
+from fairseq.dataclass import FairseqDataclass
+from fairseq.dataclass.configs import FairseqConfig
 
 logger = logging.getLogger(__name__)
 
@@ -344,7 +345,8 @@ def override_module_args(args: Namespace) -> Tuple[List[str], List[str]]:
 
         no_dc = True
         if hasattr(args, "arch"):
-            from fairseq.models import ARCH_MODEL_REGISTRY, ARCH_MODEL_NAME_REGISTRY
+            from fairseq.models import (ARCH_MODEL_NAME_REGISTRY,
+                                        ARCH_MODEL_REGISTRY)
 
             if args.arch in ARCH_MODEL_REGISTRY:
                 m_cls = ARCH_MODEL_REGISTRY[args.arch]

fairseq/distributed/__init__.py

@@ -3,14 +3,11 @@
 # This source code is licensed under the MIT license found in the
 # LICENSE file in the root directory of this source tree.
 
-from .fully_sharded_data_parallel import (
-    fsdp_enable_wrap,
-    fsdp_wrap,
-    FullyShardedDataParallel,
-)
+from .fully_sharded_data_parallel import (FullyShardedDataParallel,
+                                          fsdp_enable_wrap, fsdp_wrap)
 
 __all__ = [
     "fsdp_enable_wrap",
     "fsdp_wrap",
     "FullyShardedDataParallel",
-]
+]

fairseq/distributed/fully_sharded_data_parallel.py

@@ -7,12 +7,13 @@
 from typing import Optional
 
 import torch
+
 from fairseq.dataclass.configs import DistributedTrainingConfig
 from fairseq.distributed import utils as dist_utils
 
-
 try:
-    from fairscale.nn.data_parallel import FullyShardedDataParallel as FSDP # type: ignore
+    from fairscale.nn.data_parallel import \
+        FullyShardedDataParallel as FSDP  # type: ignore
 
     has_FSDP = True
 except ImportError:
@@ -91,7 +92,7 @@ def size(self) -> int:
 @contextlib.contextmanager
 def fsdp_enable_wrap(cfg: DistributedTrainingConfig):
     try:
-        from fairscale.nn import enable_wrap # type: ignore
+        from fairscale.nn import enable_wrap  # type: ignore
     except ImportError:
         raise ImportError(
             "Cannot find FullyShardedDataParallel. "
@@ -131,7 +132,7 @@ def fsdp_wrap(module, min_num_params: Optional[int] = None, **kwargs):
         min_num_params (int, Optional): minimum number of layer params to wrap
     """
     try:
-        from fairscale.nn import wrap # type: ignore
+        from fairscale.nn import wrap  # type: ignore
 
         if min_num_params is not None:
             num_params = sum(p.numel() for p in module.parameters())

fairseq/file_io.py

@@ -10,7 +10,6 @@
 import shutil
 from typing import List, Optional
 
-
 logger = logging.getLogger(__file__)
 
 

fairseq/logging/metrics.py

@@ -18,7 +18,6 @@
 
 from .meters import *
 
-
 # Aggregation contexts are considered "active" when inside the scope
 # created by the :func:`aggregate` context manager.
 _aggregators = OrderedDict()
@@ -329,7 +328,7 @@ def load_state_dict(state_dict):
 
 def xla_metrics_report():
     try:
-        import torch_xla.debug.metrics as met # type: ignore
+        import torch_xla.debug.metrics as met  # type: ignore
 
         print(met.metrics_report())
     except ImportError:

fairseq/models/fairseq_decoder.py

@@ -6,9 +6,10 @@
 from typing import Dict, List, Optional, Tuple
 
 import torch.nn as nn
-from fairseq import utils
 from torch import Tensor
 
+from fairseq import utils
+
 
 class FairseqDecoder(nn.Module):
     """Base class for decoders."""

fairseq/models/fairseq_encoder.py

@@ -9,7 +9,6 @@
 import torch.nn as nn
 from torch import Tensor
 
-
 EncoderOut = NamedTuple(
     "EncoderOut",
     [

fairseq/models/fairseq_model.py

@@ -13,15 +13,14 @@
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
+from omegaconf import DictConfig
+from torch import Tensor
+
 from fairseq import utils
 from fairseq.data import Dictionary
-from fairseq.dataclass.utils import (
-    convert_namespace_to_omegaconf,
-    gen_parser_from_dataclass,
-)
+from fairseq.dataclass.utils import (convert_namespace_to_omegaconf,
+                                     gen_parser_from_dataclass)
 from fairseq.models import FairseqDecoder, FairseqEncoder
-from omegaconf import DictConfig
-from torch import Tensor
 
 logger = logging.getLogger(__name__)