You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For textcat the dictionary other_scores contains cats_score_desc which has a description, e.g. "F1 macro", rather than a loss value. When passed on to AIM through the call aim_run.track(loss_value, name=loss_name, context={'type':f'other_scores_{score_name}'}, epoch=epoch, step=step) the normalization of the loss value results in an error.
Solution: add a test for the score name ending in '_desc' in base_logger.py in the method aim_log_step.
ifisinstance(other_scores, dict):
forscore_name, loss_valueinother_scores.items():
ifnotisinstance(loss_value, dict) andnotscore_name.endswith('_desc'): # <-- extra condition for _descaim_run.track(loss_value, name=loss_name, context={'type':f'other_scores_{score_name}'},
epoch=epoch, step=step)
The text was updated successfully, but these errors were encountered:
For
textcat
the dictionaryother_scores
containscats_score_desc
which has a description, e.g. "F1 macro", rather than a loss value. When passed on to AIM through the callaim_run.track(loss_value, name=loss_name, context={'type':f'other_scores_{score_name}'}, epoch=epoch, step=step)
the normalization of the loss value results in an error.Solution: add a test for the score name ending in '_desc' in base_logger.py in the method
aim_log_step
.The text was updated successfully, but these errors were encountered: