Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

SP not giving same output for same input with learning disabled #1380

Closed
scottpurdy opened this issue Jan 4, 2018 · 4 comments
Closed

SP not giving same output for same input with learning disabled #1380

scottpurdy opened this issue Jan 4, 2018 · 4 comments
Assignees
Labels

Comments

@scottpurdy
Copy link
Contributor

With learning disabled and boosting set to 0, the Python SP will give the exact same output when repeatedly given the same input.

When the C++ SP is swapped in, the results vary dramatically. This should not be the case. If there is a dtype conversion issue, then we need to check for it and throw an exception.

@lscheinkman lscheinkman self-assigned this Jan 4, 2018
lscheinkman added a commit to lscheinkman/nupic.core-legacy that referenced this issue Jan 10, 2018
lscheinkman added a commit to lscheinkman/nupic-legacy that referenced this issue Jan 10, 2018
@lscheinkman
Copy link
Contributor

@scottpurdy I was not able to reproduce the issue. Can you review the tests I added to make sure it covers the issue?

@breznak
Copy link
Member

breznak commented Jan 10, 2018

@lscheinkman can you please open a PR with the tests? Actually I also wanted to get a reproducible example first. And I am also observing a "strange" behavior with c++ SP, which I'd acknowledge to this issue.

@breznak
Copy link
Member

breznak commented Jan 12, 2018

When the C++ SP is swapped in, the results vary dramatically.

@scottpurdy can you please give us details on how to reproduce the issue?
I realized my test case was flawed (learning must be false, otherwise the change is from learning)... And w/ learniing=off I also cannot reproduce.

Also, have you observed the issue when working with c++ class, or a python wrapper?

lscheinkman added a commit to lscheinkman/nupic.core-legacy that referenced this issue Jan 12, 2018
lscheinkman added a commit to lscheinkman/nupic.core-legacy that referenced this issue Jan 12, 2018
@scottpurdy
Copy link
Contributor Author

Experienced with Python wrapper and @lscheinkman has a fix (related to dtype casting)

lscheinkman added a commit to lscheinkman/nupic.core-legacy that referenced this issue Jan 12, 2018
lscheinkman added a commit to lscheinkman/nupic.core-legacy that referenced this issue Jan 13, 2018
lscheinkman added a commit that referenced this issue Jan 16, 2018
 Issue #1380: Validate python parameters passed to SP compute method
lscheinkman added a commit to lscheinkman/nupic-legacy that referenced this issue Jan 16, 2018
lscheinkman added a commit to lscheinkman/nupic-legacy that referenced this issue Jan 16, 2018
lscheinkman added a commit to lscheinkman/nupic-legacy that referenced this issue Jan 16, 2018
lscheinkman added a commit to numenta/nupic-legacy that referenced this issue Jan 16, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging a pull request may close this issue.

3 participants