Closed emorysaba closed 3 years ago
Hi,
thank you for raising this issue! We pad the signal to fill batches at the end of the signal - this can lead to cases where there are detections that extend beyond the original signal. I will add a check that prevents that from happening and will let you know once that is fixed.
I pushed a potential fix. Can you update with pip install das --upgrade
and let me know whether this fixes the issue? Thanks!
I downloaded the new version from the GitHub page and unfortunately, there seems to be a new error - now whenever I load any audio file, this pops up
These are the errors when opening the file and running predict, respectively in full
I've pushed a new version that should fix the predict issue.
Please upgrade the GUI with pip install xarray-behave --upgrade
and let me know whether this helps.
The other error is independent - I've opened a new issue #11. Looks like some issue caused by a new version of pyqtgraph or qt. Could you post the output of conda list
there? Sorry about the hassle...
Hi! The original problem seems to be fixed, but now when I run predict, the model finds 1 or no segments where it used to accurately predict many.
I messed this up, terribly sorry about that. I've pushed code that fixes the issue now. Hopefully...
Please update das via pip install das --upgrade
and let me know whether all works as expected now.
Thanks for your help and your patience!
Yes, it seems to be working as expected - thank you for all your help, I really do appreciate it!
Great - closing this.
Hello! I'm running into an issue with the predict function (screenshot attached below) - I've tried re-assembling datasets and re-training models every which way and still run into this issue about 2/3rds of the time when using it. Any idea as to what could be causing this? Thank you in advance!