Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,104 @@
+---
+library_name: ultralytics
+tags:
+  - yolo
+  - yolo11
+  - object-detection
+  - logo-detection
+  - logodet-3k
+  - brandspotter
+license: mit
+datasets:
+  - LogoDet-3K
+metrics:
+  - mAP50
+  - mAP50-95
+  - precision
+  - recall
+pipeline_tag: object-detection
+---
+
+# BrandSpotter — Logo Detection & Brand Identification
+
+A three-stage pipeline for detecting and identifying brand logos in images: **YOLO11 detection**, **ResNet50 classification**, and **open-set rejection** for unknown brands.
+
+This repo contains the trained model weights. Source code: [github.com/daa2618/brandspotter](https://github.com/daa2618/brandspotter)
+
+## Models
+
+### YOLO11m — Logo Detection (`yolo/`)
+
+Fine-tuned YOLO11m for single-class logo detection on [LogoDet-3K](https://github.com/Wangjing1551/LogoDet-3K-Dataset).
+
+| Metric | Value |
+|--------|-------|
+| mAP@0.5 | **0.894** |
+| mAP@0.5:0.95 | **0.639** |
+| Precision | 0.829 |
+| Recall | 0.863 |
+
+**Training details:**
+- Base model: `yolo11m.pt` (pretrained)
+- Epochs: 50 (best checkpoint at epoch 47)
+- Image size: 640x640
+- Optimizer: auto (AdamW)
+- Learning rate: 0.001
+- Batch size: auto
+- Hardware: Google Colab T4 GPU (~2 hours)
+- Dataset: LogoDet-3K (single-class: "logo")
+- Augmentation: mosaic, randaugment, erasing (0.4), horizontal flip (0.5)
+
+### ResNet50 — Brand Classification (`resnet/`)
+
+_Coming soon._
+
+## Usage
+
+```python
+from ultralytics import YOLO
+
+# Download from HuggingFace
+model = YOLO("hf://vectorized-dev/brandspotter/yolo/best.pt")
+
+# Run inference
+results = model("path/to/image.jpg")
+results[0].show()
+```
+
+Or download manually and load from a local path:
+
+```python
+model = YOLO("path/to/best.pt")
+results = model("path/to/image.jpg")
+```
+
+## Files
+
+```
+yolo/
+  best.pt       — Trained weights (best checkpoint, ~39 MB)
+  args.yaml     — Full training arguments
+  results.csv   — Per-epoch training metrics
+```
+
+## Dataset
+
+[LogoDet-3K](https://github.com/Wangjing1551/LogoDet-3K-Dataset) (Wang et al., ACM TOMM 2022). 158,652 images across 3,000 logo classes. The detection model treats all logos as a single class for region proposal; brand identification is handled by the downstream classifier.
+
+## Citation
+
+```bibtex
+@article{wang2022logodet3k,
+  title={LogoDet-3K: A Large-scale Image Dataset for Logo Detection},
+  author={Wang, Jing and Min, Weiqing and Hou, Sujuan and Ma, Shengnan and Zheng, Yuanjie and Jiang, Shuqiang},
+  journal={ACM Transactions on Multimedia Computing, Communications, and Applications},
+  volume={18},
+  number={3},
+  year={2022},
+  publisher={ACM}
+}
+```
+
+## License
+
+MIT