-
While trying Ignite (I love it!) I came across a situation where I wanted to pass the trainer engine into a call to Example:
|
Beta Was this translation helpful? Give feedback.
Replies: 2 comments
-
@aksg87 thanks for the question and the feedback about the library ! Unfortunately, in the current state of the library, we can not propagate args/kwargs to the handler but I could be easy to add as a new feature. Currently, instead of using with TensorboardLogger(log_dir="quick-start-segmentation") as tb_logger:
e = Events.EPOCH_COMPLETED
evaluator.add_event_handler(
e, predictions_gt_images_handler, tb_logger, e, trainer=trainer
) PS: by the way, in case if you haven't seen previously our message about a survey:
|
Beta Was this translation helpful? Give feedback.
-
Thanks @vfdev-5 -- I completed the survey as well :) |
Beta Was this translation helpful? Give feedback.
@aksg87 thanks for the question and the feedback about the library !
Unfortunately, in the current state of the library, we can not propagate args/kwargs to the handler but I could be easy to add as a new feature.
Currently, instead of using
tb_logger.attach
you can write it like that:PS: by the way, in case if you haven't seen previously our message about a survey: