Logging with Accelerate

Accelerate has its own logging utility to handle logging while in a distributed system. To utilize this replace cases of logging with accelerate.logging:

- import logging
+ from accelerate.logging import get_logger
- logger = logging.getLogger(__name__)
+ logger = get_logger(__name__)

accelerate.logging.get_logger

< >

( name: str )

Parameters

  • name (str) — The name for the logger, such as __file__

Returns a logging.Logger for name that can handle multiprocessing.

If a log should be called on all processes, pass main_process_only=False

Example:

>>> from accelerate.logging import get_logger

>>> logger = get_logger(__name__)

>>> logger.info("My log", main_process_only=False)
>>> logger.debug("My log", main_process_only=True)