seanytak/ncf test tuning #1092
seanytak/ncf test tuning #1092
Conversation
…enders into seanytak/ncf_test_tuning
seanytak
requested review from
gramhagen,
miguelgfierro and
yueguoguo
as code owners
|
Check out this pull request on You'll be able to see Jupyter notebook diff and discuss changes. Powered by ReviewNB. |
|
|
|
@niwilso I'm trying to add you as reviewer but for some reason I can't. Feel free to add comments and suggestions as well |
| @@ -0,0 +1,930 @@ | |||
| { | |||
There was a problem hiding this comment.
it came as a surprise that you are comparing both algos instead of doing hyperparameter tuning on NCF. I think for simplicity it would be better to have a notebook with hyperparameter tuning for NCF.
About having a comparison between SVD and NCF, is there a benefit of having 1 notebook with the comparison vs comparing the outputs of both independent notebooks?
Reply via ReviewNB
There was a problem hiding this comment.
Originally I was thinking that this new notebook would focus more on the comparison between model architectures after they've been tuned through NNI. With only the SVD and NCF model though comparing the outputs of the two notebooks may serve the same purpose. I'll refactor the notebook to focus on tuning for NCF, thanks for the suggestion!
There was a problem hiding this comment.
could you please ping me here when the change is done? sorry I have so many messages coming from github
|
There is a broken link to the training script. |
|
Hi @anargyri addressed your comment in recent commits. Thanks! |
|
There is still a broken link (NCF Training Script). |
Description
This PR adds a training notebook and script for the NCF model to show readers how NNI can be applied to different types of models (the SVD and NCF) as a part of the model selection process.
Adds a notebook comparing NCF and SVD under the NNI framework.
Adds a training script for NCF to be used with the notebook.
Related Issues
#1091
Checklist:
stagingand notmaster.