ignite.utils¶

Module with helper methods

`convert_tensor`	Move tensors to relevant device.
`apply_to_tensor`	Apply a function on a tensor or mapping, or sequence of tensors.
`apply_to_type`	Apply a function on a object of input_type or mapping, or sequence of objects of input_type.
`to_onehot`	Convert a tensor of indices of any shape (N, …) to a tensor of one-hot indicators of shape `(N, num_classes, …) and of type uint8.
`setup_logger`	Setups logger: name, level, format etc.
`manual_seed`	Setup random state from a seed for torch, random and optionally numpy (if can be imported).

ignite.utils.apply_to_tensor(input_: Union[torch.Tensor, collections.abc.Sequence, collections.abc.Mapping, str, bytes], func: Callable) → Union[torch.Tensor, collections.abc.Sequence, collections.abc.Mapping, str, bytes][source]¶: Apply a function on a tensor or mapping, or sequence of tensors.

ignite.utils.apply_to_type(input_: Union[Any, collections.abc.Sequence, collections.abc.Mapping, str, bytes], input_type: Union[Type, Tuple[Type[Any], Any]], func: Callable) → Union[Any, collections.abc.Sequence, collections.abc.Mapping, str, bytes][source]¶: Apply a function on a object of input_type or mapping, or sequence of objects of input_type.

ignite.utils.convert_tensor(input_: Union[torch.Tensor, collections.abc.Sequence, collections.abc.Mapping, str, bytes], device: Optional[Union[str, torch.device]] = None, non_blocking: bool = False) → Union[torch.Tensor, collections.abc.Sequence, collections.abc.Mapping, str, bytes][source]¶: Move tensors to relevant device.

ignite.utils.manual_seed(seed: int) → None[source]¶

Setup random state from a seed for torch, random and optionally numpy (if can be imported).

Parameters: seed (int) – Random state seed

Changed in version 0.4.3: Added torch.cuda.manual_seed_all(seed).

ignite.utils.setup_logger(name: Optional[str] = None, level: int = 20, stream: Optional[TextIO] = None, format: str = '%(asctime)s %(name)s %(levelname)s: %(message)s', filepath: Optional[str] = None, distributed_rank: Optional[int] = None) → logging.Logger[source]¶

Setups logger: name, level, format etc.

Parameters

name (str, optional) – new name for the logger. If None, the standard logger is used.
level (int) – logging level, e.g. CRITICAL, ERROR, WARNING, INFO, DEBUG.
stream (TextIO, optional) – logging stream. If None, the standard stream is used (sys.stderr).
format (str) – logging format. By default, %(asctime)s %(name)s %(levelname)s: %(message)s.
filepath (str, optional) – Optional logging file path. If not None, logs are written to the file.
distributed_rank (int, optional) – Optional, rank in distributed configuration to avoid logger setup for workers. If None, distributed_rank is initialized to the rank of process.

Returns

logging.Logger

For example, to improve logs readability when training with a trainer and evaluator:

from ignite.utils import setup_logger

trainer = ...
evaluator = ...

trainer.logger = setup_logger("trainer")
evaluator.logger = setup_logger("evaluator")

trainer.run(data, max_epochs=10)

# Logs will look like
# 2020-01-21 12:46:07,356 trainer INFO: Engine run starting with max_epochs=5.
# 2020-01-21 12:46:07,358 trainer INFO: Epoch[1] Complete. Time taken: 00:5:23
# 2020-01-21 12:46:07,358 evaluator INFO: Engine run starting with max_epochs=1.
# 2020-01-21 12:46:07,358 evaluator INFO: Epoch[1] Complete. Time taken: 00:01:02
# ...

Changed in version 0.4.3: Added stream parameter.

ignite.utils.to_onehot(indices: torch.Tensor, num_classes: int) → torch.Tensor[source]¶: Convert a tensor of indices of any shape (N, …) to a tensor of one-hot indicators of shape (N, num_classes, …) and of type uint8. Output’s device is equal to the input’s device.

Changed in version 0.4.3: This functions is now torchscriptable.