Adam Van Etten
Adam Van Etten
The most recent dockerfile builds fine as of 19 July 2019. Give that a shot. The temporary fix in [#53](https://github.com/avanetten/simrdwn/issues/53) may build successfully, but leads to errors when running inference.
Without knowing your training command and output it's impossible to tell if training ran correctly. I would check the loss file in the training directory to ensure that training converged....
Add 80% of bounding box extent for YOLT
Remove labels where only a small portion of the building remains in the chip
Just updated environment_gpu.yml within Solaris. Hopefully this will fix your issue.
_clean_sub_graphs() is indeed a function within apls.py, so I'm not sure of the source of this error. Try the updated version of apls to see if this is resolved.