Validate output based on user-defined algorithm #937
Unanswered
MaxMuehlbauer
asked this question in
Q&A
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I have an input with size 76 and an output of size 6, where the output basically consists of:
Although the network is quite good at it from the training, there are some fixed rules, that must always apply to the output. I think when it was somehow possible to check during training that these rules apply, the training outcome would be even better.
The rules are (so you get an idea; in pseudo code):
Is it somehow possible to achieve this? Does it even make sense to you?
Beta Was this translation helpful? Give feedback.
All reactions