Open cherrywoods opened 2 years ago
Sorry for the late reply, my life is a bit busy right now. Thanks for another interesting bug. I'm not quite sure where the issue is yet, but my guess is that we are incorrectly inferring the shape of some tensor which causes a future simplification to be incorrect. This may take some time to figure out.
Thank you for taking your time to answer me and for the update. I have tried to directly get verifiers running on my inputs, but without much luck. Therefore, the issue is still relevant.
Hallo, I am running into problems using DNNV that probably stem from my input network and property, but I can't find out what the issue is. I isolated the parts that looked suspicious to me, but couldn't reproduce the error. I would be thankful for any pointers what might be wrong with my network.
The network and property are in this file: inputs.zip
DNNV shows network and property like this:
This is the remaining DNNV output including the exception trace:
I tested:
I am using the latest commit on the develop branch.
I am running:
Best regards, David