Brand Logo Classification & Detection Models (Flickr27)
This repository contains trained transfer learning checkpoints and YOLOv8 weights for brand logo classification and object detection on the Flickr27 dataset.
Checkpoint Files
EfficientNet.ckpt: EfficientNet-B0 transfer learning checkpoint (~20 MB, 4.35M parameters)InceptionV3.ckpt: InceptionV3 transfer learning checkpoint (~90 MB, 22.32M parameters)MobileNetV2.ckpt: MobileNetV2 transfer learning checkpoint (~13 MB, 2.56M parameters)Xception.ckpt: XceptionNet transfer learning checkpoint (~86 MB, 21.34M parameters)LogoYolobest.pt: YOLOv8 logo detection weights (~6 MB)
27 Canonical Brand Classes
Adidas, Apple, BMW, Citroen, Cocacola, DHL, Fedex, Ferrari, Ford, Google, HP, Heineken, Intel, McDonalds, Mini, Nbc, Nike, Pepsi, Porsche, Puma, RedBull, Sprite, Starbucks, Texaco, Unicef, Vodafone, Yahoo.