Wanting to try out some of the new deep learning tools in Pro 2.8. I have a clean install of 2.8, cloned Python environment and the deep learning modules installed by "conda install deep-learning-essentials".
I can successfully run the deep learning tools with CPU specified in the environment variables. But when I choose GPU I get an error box popup: "ArcGIS Pro has experienced a fatal authorization error and must close". Screenshot attached showing error and messages when running the Detect Objects tool.
I have ArcGIS Pro 2.8 Advanced and the Image Analyst extension licensed as a named user through ArcGIS Online. The same thing is occurring on two different machines, desktop with 5Gb Quadro P2000 and laptop with 8Gb GeForce RTX 2080. Both machines on Windows 10.
Am I just too quick off the mark trying to do deep learning in Pro 2.8?
Can you check the 2nd link versus your download to make sure the version for 2.8 was installed.
In the last link, it appears that the install instructions for 2.8 aren't up yet.
In case I missed something.
Yes, it's definitely installing deep-learning-essentials 2.8. Conda install (dry run) and conda list outputs attached.
Further investigation shows these tools have the error when using GPU:
These tools work without problems when using GPU:
All the tools above work without problems when using CPU.
Additionally, I can perform inferencing with a Change Detection model using the GPU when using a Jupyter Notebook.
@Tim_McGinnes you might want to look at the first issue at...
It isn't your gpu, however, it provides background.
You might be advised to file an Issue there if you haven't done one through Tech Support yet.
Issue logged at the Esri Deep Learning Frameworks GitHub site, and someone else has independently confirmed the error: https://github.com/Esri/deep-learning-frameworks/issues/19
Any updates on this? I am also having the same issue with 2.8.
I see this particular error dialog when executing "Classify Objects Using Deep Learning" from both GUI and arcpy function. It creates an empty output feature class. Please advise if we need to reinstall 2.7 from 2.8 to use deep learning tools.
Thanks for the quick response. I think I will need to revert back to 2.7 as well. It is unfortunate/headache to revert back to 2.7 with python downgrades after all the improvements and bug fixes were announced.