🚀 feat(model): add UniNet #2797
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Ashwin Vaidya <ashwinnitinvaidya@gmail.com>
Signed-off-by: Ashwin Vaidya <ashwinnitinvaidya@gmail.com>
samet-akcay
changed the title
samet-akcay
reviewed
Jun 26, 2025
Signed-off-by: Ashwin Vaidya <ashwinnitinvaidya@gmail.com>
Signed-off-by: Ashwin Vaidya <ashwinnitinvaidya@gmail.com>
Signed-off-by: Ashwin Vaidya <ashwinnitinvaidya@gmail.com>
Signed-off-by: Ashwin Vaidya <ashwinnitinvaidya@gmail.com>
samet-akcay
requested changes
Jul 14, 2025
Signed-off-by: Ashwin Vaidya <ashwin.vaidya@intel.com>
Signed-off-by: Ashwin Vaidya <ashwin.vaidya@intel.com>
Signed-off-by: Ashwin Vaidya <ashwin.vaidya@intel.com>
There was a problem hiding this comment.
Pull Request Overview
This PR adds the UniNet model implementation for anomaly detection to the anomalib library. UniNet is designed for diverse domains and supports both supervised and unsupervised anomaly detection with a focus on multi-class scenarios.
- Implements complete UniNet model architecture with teachers, student, and bottleneck modules
- Adds domain-related feature selection and attention bottleneck components
- Refactors ResNet decoder to be shared between Reverse Distillation and UniNet models
Reviewed Changes
Copilot reviewed 17 out of 17 changed files in this pull request and generated 6 comments.
Show a summary per file
| File | Description |
|---|---|
| src/anomalib/models/image/uninet/ | Complete UniNet model implementation with PyTorch and Lightning modules |
| src/anomalib/models/components/backbone/ | Refactored ResNet decoder to be shared across models |
| src/anomalib/models/image/reverse_distillation/ | Updated imports to use shared decoder component |
| examples/configs/model/uninet.yaml | Configuration file for UniNet model |
| third-party-programs.txt | Added license attribution for UniNet |
Signed-off-by: Ashwin Vaidya <ashwin.vaidya@intel.com>
Signed-off-by: Ashwin Vaidya <ashwin.vaidya@intel.com>
Signed-off-by: Ashwin Vaidya <ashwin.vaidya@intel.com>
| Decoder network implementation | ||
| """ | ||
|
|
||
| # Copyright (C) 2025 Intel Corporation |
There was a problem hiding this comment.
copyright goes to the top?
Same for other files as well.
| class AttentionBottleneck(nn.Module): | ||
| """Attention Bottleneck for UniNet.""" | ||
|
|
||
| expansion: int = 4 |
There was a problem hiding this comment.
can you rename expansion to be a bit more specific or add a comment.
| loss: nn.Module, | ||
| ) -> None: | ||
| super().__init__() | ||
| self.teachers = Teachers(teacher_backbone) |
There was a problem hiding this comment.
Is there any advantage of creating the student and bottleneck in lightning model and the teachers in torch model ?
| self.bce_loss = torch.nn.BCEWithLogitsLoss() | ||
| # Used to post-process the student features from the de_resnet model to get the predictions | ||
| self.avgpool = nn.AdaptiveAvgPool2d((1, 1)) | ||
| self.fc = nn.Linear(256, 1) |
There was a problem hiding this comment.
would this 256 in nn.Linear(256, 1) be valid for the set of possible backbones?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
📝 Description
uninet.yamldoes not work as Early Stopping from config is broken for all the models.Results
Combined
Original
Anomalib
Anomaly Map Examples
✨ Changes
Select what type of change your PR is:
✅ Checklist
Before you submit your pull request, please make sure you have completed the following steps:
For more information about code review checklists, see the Code Review Checklist.