@<1523701087100473344:profile|SuccessfulKoala55> This is just an example. In my actual use case, the model training code developed by my team uses the logging module, so I didn’t use the task logger. Also I don’t think the task logger supports different levels of logging
Hi @<1610083512721870848:profile|CleanPelican60> , it should appear there - how are you running your code?
I see. The ClearML module, when loaded, reconfigures the Python logging module - default configuration is in the ClearML logging
section of the configuration (see here ). You can try to override parts of it.
I’m running my code line by line in VSCode’s terminal. Here’s my snippet:
from clearml import Task
import logging
task = Task.init(
project_name='abc',
task_name='abc',
output_uri=True, # Does not upload anywhere.
)
logger = task.get_logger()
logging.info('This does not show up')
logging.error('This does show up')
Wait, why are you getting the task logger and not using it?
@<1523701087100473344:profile|SuccessfulKoala55> could you help me with this please? Thanks! 😄
You can use the clearml.conf file, just add a section there as follows:
...
logging {
}
and change what you want
@<1523701087100473344:profile|SuccessfulKoala55> Thank you. How can I override this? Do you mean modifying the logging.conf file? And what settings should I change here to make my logging.info ()
calls show up? Sorry I’m quite new to this.