-
Notifications
You must be signed in to change notification settings - Fork 1.6k
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
CLAModel name changed to HTMPredictionModel #3516
Conversation
As discussed offline, |
I kindof like |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Spittem!
Will wait for @subutai before merging. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good overall. This is a major incompatible change - recommend warning the community.
@@ -34,7 +34,7 @@ from nupic.frameworks.opf.expdescriptionhelpers import ( | |||
applyValueGettersToContainer, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If this is marked as UNDER_DEVELOPMENT can we just delete it?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There are a bunch of "UNDER_DEVELOPMENT" examples. I've removed them in another PR.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks - might want to merge in master - will likely be merge conflicts.
docs/source/quick-start/opf.rst
Outdated
@@ -34,7 +34,7 @@ using the :class:`.ModelFactory`. | |||
|
|||
.. literalinclude:: ../../examples/opf/create-model-example.py | |||
|
|||
The resulting ``model`` will be an instance of :class:`.CLAModel`. | |||
The resulting ``model`` will be an instance of :class:`.SPTMModel`. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
SPTM is not too informative. Should we consider calling it StreamingModel or PredictionModel?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@subutai @mrcslws @scottpurdy I would like it if you guys could talk in the office and come up with a name you agree on. Once you have one, I will update the PR.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@subutai - "Prediction" and "streaming" aren't good because all OPF models are streaming and prediction. One really important thing to capture is that this is HTM-related since other models may be for reference (like TwoGramModel). While "HTMModel" might be ok, we may split out certain variants into their own classes rather than shoe-horning all HTM-variants into a single OPF model.
That is how we got to "SPTMModel" but definitely open to other ideas.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
And spittem is fun to say
@@ -35,12 +35,12 @@ def claModelControlEnableSPLearningCb(claModel): | |||
|
|||
See also claModelControlDisableSPLearningCb. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The text prefix claModel was not replaced. Shows up in many places.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ok, waiting for a final name before I commit anything.
@subutai Warned the community here: https://discourse.numenta.org/t/warning-0-7-0-breaking-changes/2200/4 This PR is ready to merge if we are ok with |
Decision: |
I just realized this also means we need to update the # Type of model that the rest of these parameters apply to.
'model': "CLA", More changes coming to lots of example files, and the |
# get to cleaning up the clamodel.py file. | ||
# being called at the same time as the compute() method. Only compute() | ||
# should be called via network.run(). This flag will be removed once we | ||
# get to cleaning up the htmpredictionmodel.py file. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Shouldn't the file be htm_prediction_model.py?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I was just following previous convention. Should I change both the module name and file name?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I thought having the _ was our filename convention (not module names), but you are right that the OPF directory doesn't seem to follow that. Perhaps best handled in another PR.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ok: #3546
@@ -34,7 +34,7 @@ from nupic.frameworks.opf.expdescriptionhelpers import ( | |||
applyValueGettersToContainer, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks - might want to merge in master - will likely be merge conflicts.
sb. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should I have left this pickle alone?
@scottpurdy @subutai Ready for final review. This might break regression tests. If so I will fix tomorrow. |
fixes #3515