TagLab Crash when Exporting Training Dataset

[Miguel-Figuerola]

Hello,

I am trying to export a training dataset to eventually train my network. This is the error message I get:

After a few seconds TagLab closes.

I am using Python version: 3.10.1 and CUDA Toolkit 11.6

Any idea what could be happening?

Thanks!

[Miguel-Figuerola]

I get this error using the "small_annotated_area_project" sample project as well as my own projects.

[maxcorsini]

Hello,

this is a known issue that will be fixed in future release. It may happens that some image coordinates / image size are stored as floating points instead of integers. Qt wants INTEGER, otherwise it crashes in this way. For the moment, replace the line 213 with the following:

qimage_cropped = qimage_map.copy(int(left), int(top), int(w), int(h))

Best,

[Miguel-Figuerola]

Hello Max,

Thanks for the reply! That edit worked for exporting the training dataset. However, I am now trying to train my network and I get the following error:

Thanks for your attention,

[recyclerjh]

I am having a similar issue when trying to export training dataset. Do I need to have every portion of the workspace labeled? Also I am confused on the target scale. Should I leave that at 1.0 if my orthos are all the same or similar pixel size?