-
Notifications
You must be signed in to change notification settings - Fork 855
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
FusionModel + SVM as FewShotSVMPredictor
#2850
Conversation
…p all annotation xml files
…when labels.txt is not present.
� Conflicts: � multimodal/src/autogluon/multimodal/predictor.py
…s` to using `get_detection_classes`. Refactoring saving results to utils.
� Conflicts: � multimodal/src/autogluon/multimodal/predictor.py
…ing visualization to a separate function call outside predict, removing prints, modifying doc strings, adding TODO to use mmdet visualization later,
FewShotSVMPredictor
Job PR-2850-f53113d is done. |
Just verified for image classification. This implementation also works. |
Job PR-2850-b3b059a is done. |
Job PR-2850-10c5885 is done. |
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.
- Do we want to expose
FewShotSVMPredictor
to users, i.e., lettingfrom autogluon.multimodal import FewShotSVMPredictor
?
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.
PR #2827 seems to have a similar functionality. What is the consideration of having 2 PRs of doing the same thing?
self._automm_predictor = MultiModalPredictor( | ||
label=label, | ||
hyperparameters=hyperparameters, | ||
problem_type=FEATURE_EXTRACTION, |
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.
FEATURE_EXTRACTION
is for text only now. Do we also need to handle image data?
Job PR-2850-76b063a is done. |
…fewshot_svm_predictor
I think we'll leave it as a util function for now. We can design our learner after 0.7 and expose that to the users. |
Job PR-2850-ae55a94 is done. |
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.
LGTM overall. We can refactor the source code later
Issue #, if available:
Description of changes:
Integrating FusionModel + SVM as a utility
FewShotSVMPredictor
class. The way to use this predictor is the same as inMultiModalPredictor
. See below:By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.