Add files using upload-large-folder tool

README.md

@@ -0,0 +1,52 @@
+---
+base_model:
+- Ultralytics/YOLOv8
+pipeline_tag: object-detection
+---
+
+# YOLOv8 Patent Text Region Detection Model
+
+### Model Description
+
+**patent_text_regions** is a [YOLOv8](https://docs.ultralytics.com/models/yolov8) model fine-tuned on a custom dataset of page-level images drawn from historical patent specifications published by the British Patent Office. It has been trained to recognize all text regions located within pages of patent specifications as a single class. We take the initialized weights from the official release of the small YOLOv8s model ([yolov8s.pt](https://huggingface.co/Ultralytics/YOLOv8/blob/main/yolov8s.pt)) and fine tune on our custom dataset. 
+
+### Usage
+
+This model can be used in the same way as any pre-trained [YOLOv8 model](https://huggingface.co/Ultralytics/YOLOv8) by setting the model path to [best.pt](https://huggingface.co/gbpatentdata/yolov8_patent_layouts/blob/main/weights/best.pt).
+
+### Training Data
+
+The dataset was created by randomly sampling 420 page images from British patent specifications published between 1850-1899. The data was randomly split 80-10-10 (train-val-test) and then standard preprocessing (images were stretched and auto-oriented to 640 x 640 pixels) and the following data augmentations were applied using [Roboflow](https://roboflow.com/):
+
+- Crop: 0% Minimum Zoom, 20% Maximum Zoom
+- Grayscale: Apply to 15% of images
+- Saturation: Between -25% and +25%
+- Blur: Up to 2.5px
+- Noise: Up to 0.1% of pixels
+
+The custom dataset consists of 1,092 labelled images in total, which are made available in this repository.
+
+### Hyperparameters 
+
+We train the model using [default hyperparameters](https://docs.ultralytics.com/modes/train/#introduction), except from the batch size (128) and the number of epochs (300). 
+
+### Evaluation 
+
+Evals on the test set are reported below:
+
+- mAP50: **0.987**
+- mAP50-95: **0.892**
+
+## Citation 
+
+If you use our model or custom training/evaluation data in your research, please cite our accompanying paper as follows:
+
+```bibtex
+@article{bct2025,
+  title = {300 Years of British Patents},
+  author = {Enrico Berkes and Matthew Lee Chen and Matteo Tranchero},
+  journal = {arXiv preprint arXiv:2401.12345},
+  year = {2025},
+  url = {https://arxiv.org/abs/2401.12345}
+}
+```

args.yaml

@@ -0,0 +1,93 @@
+task: detect
+mode: train
+model: yolov8s.yaml
+data: patents-3/data.yaml
+epochs: 300
+patience: 0
+batch: 128
+imgsz: 640
+save: true
+cache: false
+device:
+- 0
+- 1
+workers: 8
+project: null
+name: null
+exist_ok: false
+pretrained: false
+optimizer: SGD
+verbose: true
+seed: 42
+deterministic: true
+single_cls: false
+image_weights: false
+rect: false
+cos_lr: false
+close_mosaic: 10
+resume: false
+overlap_mask: true
+mask_ratio: 4
+dropout: false
+val: true
+save_json: false
+save_hybrid: false
+conf: 0.001
+iou: 0.7
+max_det: 300
+half: false
+dnn: false
+plots: false
+source: ultralytics/assets/
+show: false
+save_txt: false
+save_conf: false
+save_crop: false
+hide_labels: false
+hide_conf: false
+vid_stride: 1
+line_thickness: 3
+visualize: false
+augment: false
+agnostic_nms: false
+classes: null
+retina_masks: false
+boxes: true
+format: torchscript
+keras: false
+optimize: false
+int8: false
+dynamic: false
+simplify: false
+opset: 17
+workspace: 4
+nms: false
+lr0: 0.01
+lrf: 0.01
+momentum: 0.937
+weight_decay: 0.001
+warmup_epochs: 3.0
+warmup_momentum: 0.8
+warmup_bias_lr: 0.1
+box: 7.5
+cls: 0.5
+dfl: 1.5
+fl_gamma: 0.0
+label_smoothing: 0.0
+nbs: 64
+hsv_h: 0.015
+hsv_s: 0.7
+hsv_v: 0.4
+degrees: 0.0
+translate: 0.1
+scale: 0.5
+shear: 0.0
+perspective: 0.0
+flipud: 0.0
+fliplr: 0.5
+mosaic: 1.0
+mixup: 0.0
+copy_paste: 0.0
+cfg: null
+v5loader: false
+save_dir: runs/detect/train3

data/data.yaml

@@ -0,0 +1,6 @@
+train: ../train/images
+val: ../valid/images
+test: ../test/images
+
+nc: 1
+names: ['Text']

data/data_splits/test/labels/09ec2438-GB189114138A_1_jpg.rf.c9aa835d2836e2694b654a4278f02a06.txt

@@ -0,0 +1,3 @@
+0 0.46640625 0.16484375 0.62421875 0.04140625
+0 0.4703125 0.50546875 0.7 0.63671875
+0 0.65703125 0.8453125 0.2890625 0.05

data/data_splits/test/labels/12f5a89d-GB189207576A_4_jpg.rf.8f958fc23c25236e83ccaf17ad499133.txt

@@ -0,0 +1 @@
+0 0.5234375 0.5171875 0.721875 0.76328125

data/data_splits/test/labels/405cdb71-GB189823771A_1_jpg.rf.9d004f7bbb3d1b58df3a9df9081e15d3.txt

@@ -0,0 +1,2 @@
+0 0.4765625 0.23125 0.56484375 0.021875
+0 0.4921875 0.57109375 0.73125 0.64296875

data/data_splits/test/labels/c6ad622a-GB189425129A_2_jpg.rf.24ab52833dd5e22aa33a8a921b9b0e6e.txt

@@ -0,0 +1,2 @@
+0 0.525 0.38671875 0.7265625 0.55
+0 0.55546875 0.68125 0.62734375 0.03828125

data/data_splits/test/labels/f5b05c10-GB189010347A_1_jpg.rf.8a27f697b8d976945ddfd1cbe39271c9.txt

@@ -0,0 +1,5 @@
+0 0.48984375 0.184375 0.665625 0.03984375
+0 0.48046875 0.315625 0.73671875 0.221875
+0 0.69140625 0.4296875 0.25859375 0.0203125
+0 0.4890625 0.6515625 0.728125 0.4
+0 0.690625 0.85703125 0.2609375 0.0203125

data/data_splits/test/labels/f6a42619-GB189316375A_2_jpg.rf.fb49b8b49366af22c4fbdb6368473788.txt

@@ -0,0 +1,2 @@
+0 0.5234375 0.44921875 0.72890625 0.671875
+0 0.7828125 0.79453125 0.184375 0.025

data/data_splits/valid/labels/0291bc0f-GB189201439A_2_jpg.rf.2b0ab439f0fb6b2059f1e3f212df1d8c.txt

@@ -0,0 +1,2 @@
+0 0.5140625 0.190625 0.71796875 0.13125
+0 0.70234375 0.27421875 0.3078125 0.03671875

data/data_splits/valid/labels/06d6b139-GB188808551A_4_jpg.rf.eaf638be4da882cd2d7c27cd18a2ff7d.txt

@@ -0,0 +1 @@
+0 0.515625 0.48984375 0.74453125 0.796875

data/data_splits/valid/labels/0eb5f092-GB188616969A_2_jpg.rf.b744b9eba4a18e023eb73af9ab0f128d.txt

@@ -0,0 +1 @@
+0 0.503125 0.53125 0.7171875 0.78984375

data/data_splits/valid/labels/10e51567-GB186701790A_1_jpg.rf.3f5b12cbe34c153476c9c837d896adf6.txt

@@ -0,0 +1,3 @@
+0 0.521875 0.45625 0.74453125 0.09296875
+0 0.4609375 0.52265625 0.6140625 0.03359375
+0 0.52109375 0.74609375 0.73828125 0.3515625

data/data_splits/valid/labels/213c8656-GB189825691A_2_jpg.rf.85d0932599f1866c0e92450791a29f47.txt

@@ -0,0 +1,2 @@
+0 0.50859375 0.46875 0.75078125 0.6765625
+0 0.57265625 0.83515625 0.6 0.05

data/data_splits/valid/labels/37c40480-GB189000430A_3_jpg.rf.328f77b4ef08aadcad31d50ad99fc171.txt

@@ -0,0 +1,2 @@
+0 0.490625 0.22109375 0.75625 0.25703125
+0 0.64296875 0.36953125 0.4640625 0.0484375

data/data_splits/valid/labels/3cc8e014-GB188507196A_2_jpg.rf.79aed4cf3810fd2468bd3fae6959c564.txt

@@ -0,0 +1,3 @@
+0 0.49375 0.17265625 0.73515625 0.11328125
+0 0.6984375 0.23671875 0.284375 0.0203125
+0 0.4984375 0.6453125 0.74375 0.53828125

data/data_splits/valid/labels/3e3ad239-GB188003297A_1_jpg.rf.5222e025aaca0271c701f846710753b7.txt

@@ -0,0 +1,2 @@
+0 0.4953125 0.74296875 0.740625 0.4265625
+0 0.50078125 0.434375 0.75546875 0.08984375

data/data_splits/valid/labels/474036c6-GB185901872A_1_jpg.rf.c2bcd40dc0f6079ee7c0407c00cef7b7.txt

@@ -0,0 +1,3 @@
+0 0.5078125 0.475 0.75 0.08828125
+0 0.4453125 0.53828125 0.6171875 0.03203125
+0 0.5109375 0.74375 0.7390625 0.30703125

data/data_splits/valid/labels/4dcf118f-GB187803508A_3_jpg.rf.3ef883e72139ddd11fe23db3e758c2ec.txt

@@ -0,0 +1 @@
+0 0.4984375 0.534375 0.72890625 0.79296875

data/data_splits/valid/labels/6284ea8f-GB189316143A_1_jpg.rf.2b8a66a3307de1b56b72d4a5a367c236.txt

@@ -0,0 +1,2 @@
+0 0.46875 0.240625 0.6796875 0.08984375
+0 0.45703125 0.58359375 0.73828125 0.596875

data/data_splits/valid/labels/6b10d269-GB185600686A_1_jpg.rf.fa755ebd62e8cffbee4d39e5a1c09871.txt

@@ -0,0 +1,3 @@
+0 0.4875 0.6703125 0.7703125 0.46796875
+0 0.4109375 0.39921875 0.62578125 0.03125
+0 0.4875 0.3546875 0.7703125 0.05

data/data_splits/valid/labels/6dac99ef-GB185901872A_2_jpg.rf.3c1cf9d7ff25eabefe7ae795545579d5.txt

@@ -0,0 +1,2 @@
+0 0.4921875 0.21875 0.74453125 0.19375
+0 0.48984375 0.61484375 0.740625 0.5703125

data/data_splits/valid/labels/7137fb0e-GB189207094A_2_jpg.rf.59f444777a7b21d6b33847436750299e.txt

@@ -0,0 +1,2 @@
+0 0.50390625 0.40859375 0.72421875 0.56171875
+0 0.73828125 0.70234375 0.19609375 0.025