talmolab/sleap

Exporting CSV file but doesn't include predictions and instance score

msjindal opened this issue · 3 comments

Bug description

When exporting the project as a csv file the file does not include the predicted frames, only the user labeled frames

Expected behaviour

When exporting the csv after predictions, the file should include the predicted frames and their instance score, instead, it doesn't have instance scores, and it only has the user labeled frames.

Actual behaviour

Your personal set up

  • OS:
  • Version(s):
Environment packages
# paste output of `pip freeze` or `conda list` here
Logs
# paste relevant logs here, if any

Screenshots

How to reproduce

  1. Go to '...'
  2. Click on '....'
  3. Scroll down to '....'
  4. See error

Hi,
I believe I have the same issue as msjindal and a result receive error messages when importing tracking data into SimBA: the predictions seem to be needed, as well as a link of the video file within the SLP data? Could it perhaps also be an error because the predictions are saved separately from the user-labelled frames?

I tried to export the my SLEAP predictions in H5, SLP or CSV files but none of them seem to work with SimBA (they give me different error messages), so I was wondering whether perhaps some of the information is lost when saving the data from SLEAP. But of course it could also be an error on the side of SimBA...

I am very new to all of these things, so I hope what I am writing makes sense 😅
Any help/tips would be very much appreciated :) Thank you!

Hi @msjindal,

I tried to reproduce this error but I was able to generate a csv that include both user instances and predictions with instance scores:

Image

Do you mind providing a bit more info on your setup, like the SLEAP version you're using? If you don't mind sharing the .SLP file that produces this error, we can do more troubleshooting.


@glinnes12: Do you mind creating a new discussion or issue for your problem if you haven't gotten help from the simba side yet?

Thanks!

Elizabeth

This seems to have been resolved over in sgoldenlab/simba#294, but feel free to open a new issue or discussion if needed :)