Liezl Maree
Liezl Maree
Hi @ericas12, Thanks for the follow-up. Yes, the slp file should never have multiple skeletons in it. We have closed entry points for this bug in the past, but we...
Hi @ericas12, Thanks again for the follow-up. ### The labels you sent Strange, but when opening the labels I did not see multiple `Skeleton`s - in fact, I saw 0...
Hi @ericas12, Any updates from your end? I also wonder if this might be a shared/remote GPU that you are using? Thanks, Liezl
### Update The issue described below ended up being caused by a change in the data structure for `SingleImageVideo`s. After finding the issue, the data structure change was reversed in...
THis may have been the issue we worked on in: - #2064
Hi @roaldarbol, I have an inkling that this shifting of predictions likely has to do with the `resize_and_pad_to_target` and related parameters in the training_config.json. This parameter should be set to...
Ah yes, thanks for clarifying! I was definitely adding in my own assumptions (and overlooked the number of labeled frames in the project). This is a case that we haven't...
Long awaited, but finally integrated. This issue has been fixed in the SLEAP 1.4.1 release!
Suceeded by: - #2050
We still need to allow setting the max instances for bottom-up models