U 0È-e»ã@sDddlZddlZddlmZGdd„dejƒZd eedœdd„ZdS) éNé)Ú PartialStatec@s$eZdZdZedd„ƒZdd„ZdS)ÚMultiProcessAdaptera@ An adapter to assist with logging in multiprocess. `log` takes in an additional `main_process_only` kwarg, which dictates whether it should be called on all processes or only the main executed one. Default is `main_process_only=True`. Does not require an `Accelerator` object to be created first. cCstƒ}| p|o|jS)z Check if log should be performed)rZis_main_process)Úmain_process_onlyÚstate©rúS/var/www/html/Darija-Ai-Train/env/lib/python3.8/site-packages/accelerate/logging.pyÚ _should_logszMultiProcessAdapter._should_logc Os¼tjikrtdƒ‚| dd¡}| dd¡}| |¡r¸| |¡rf| ||¡\}}|jj||f|ž|ŽnR|r¸tƒ}t |j ƒD]<}||j kr®| ||¡\}}|jj||f|ž|Ž|  ¡qzdS)a\ Delegates logger call after checking if we should log. Accepts a new kwarg of `main_process_only`, which will dictate whether it will be logged across all processes or only the main executed one. Default is `True` if not passed Also accepts "in_order", which if `True` makes the processes log one by one, in order. This is much easier to read, but comes at the cost of sometimes needing to wait for the other processes. Default is `False` to not break with the previous behavior. `in_order` is ignored if `main_process_only` is passed. z€You must initialize the accelerate state by calling either `PartialState()` or `Accelerator()` before using the logging utility.rTÚin_orderFN) rZ _shared_stateÚ RuntimeErrorÚpopÚ isEnabledForr ÚprocessÚloggerÚlogÚrangeZ num_processesZ process_indexZwait_for_everyone) ÚselfÚlevelÚmsgÚargsÚkwargsrr rÚirrrr%s" ÿ     zMultiProcessAdapter.logN)Ú__name__Ú __module__Ú __qualname__Ú__doc__Ú staticmethodr rrrrrrs  r)ÚnameÚ log_levelcCsP|dkrtj dd¡}t |¡}|dk rF| | ¡¡|j | ¡¡t|iƒS)a= Returns a `logging.Logger` for `name` that can handle multiprocessing. If a log should be called on all processes, pass `main_process_only=False` If a log should be called on all processes and in order, also pass `in_order=True` Args: name (`str`): The name for the logger, such as `__file__` log_level (`str`, *optional*): The log level to use. If not passed, will default to the `LOG_LEVEL` environment variable, or `INFO` if not Example: ```python >>> 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) >>> logger = get_logger(__name__, log_level="DEBUG") >>> logger.info("My log") >>> logger.debug("My second log") >>> from accelerate import Accelerator >>> accelerator = Accelerator() >>> array = ["a", "b", "c", "d"] >>> letter_at_rank = array[accelerator.process_index] >>> logger.info(letter_at_rank, in_order=True) ``` NZACCELERATE_LOG_LEVEL) ÚosÚenvironÚgetÚloggingÚ getLoggerÚsetLevelÚupperÚrootr)rrrrrrÚ get_loggerGs# r')N)r"rrrÚ LoggerAdapterrÚstrr'rrrrÚs 2