-
- Google colaboratory GPU
-
- Chose 7 common dog breeds from Stanford Dogs Dataset. Training on 1193 images
-
-
Classification part was inspired by ResNet architecture principles and for localization only changed last resnet layers to regression layers to predict x1, y1, x2, y2 boundary boxes corners coordinates. Training notebook named train.ipynb, testing notebook - test.ipynb
-
Classification head training accuracy during 300 epochs on Google colab GPU
-
Classification head training loss during 300 epochs on Google colab GPU
-
Regression head training accuracy during 15 epochs on Google colab GPU
-
Regression head training loss during 15 epochs on Google colab GPU
-
-
Notifications
You must be signed in to change notification settings - Fork 0
Antanskas/Single-object-detection-classification
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
Classify and localize 7 dog breeds using resnet + regression heads
Resources
Stars
Watchers
Forks
Releases
No releases published