Train Deep Learning failure with MaskRCNN

[ColinBurchill1]

I'm using ArcPro 2.9.13. I keep running into this same error whenever I try to train a model based on samples I exported using the MaskRCNN metadata format. I ran this same workflow a couple years ago and was successful but i can't identify the issue now.

Any help parsing the error message would be greatly appreciated!

[Robert_LeClair · Accepted solution]

The only thing I'm finding related to this error message is the following:

Resolution: BUG-000160687 : "IndexError: index 0 is out of bounds for axis 0 with size 0" is returned when a space character exists in the training sample feature class' Classname attributes.
Workaround: Remove the space character from the training sample feature class' Classname attributes and re-run the Export Training Data For Deep Learning tool, followed by the Train Deep Learning Model tool.

BUG-000160400 : Error 999999 is returned when running the Export Training Data For Deep Learning tool with an input mask polygon.
Workaround: Leave the Input Mask Polygons parameter empty before running the tool or specify an extent under Environments tab > Processing Extent.

[ColinBurchill1]

I appreciate your input! After a bit more testing we found that it was the space character in our Classname that was causing it to fail.