Datasets:
| task_categories: | |
| - object-detection | |
| language: | |
| - hy | |
| pretty_name: hye_yolo_v0 | |
| size_categories: | |
| - n<1K | |
| tags: | |
| - handwritten text | |
| - dictation | |
| - YOLOv8 | |
| license: mit | |
| # Handwritten text detection dataset | |
| ## Data domain | |
| The blanks were provided by youth organization "Armenian Club" ([telegram](https://t.me/armenian_club), [instagram](https://www.instagram.com/armenian.club?igsh=MTJjYTN0dTdjamtxMQ==) ), Russia Moscow. | |
| The text on blanks was written during dictation "Teladrutyun" in 2018 | |
| The blanks were labeled by [Amir](https://huggingface.co/Agmiyas) and [Renal](https://huggingface.co/Renaxit) during research project in HSE MIEM | |
| ## Dataset info | |
| Contains labeled dictations blanks in YOLO format | |
| 91 image in total, 73 (80%) for train and 18 (20%) for test | |
| No image alignment or any preprocess | |
| Resolution 1320x1020, 96 dpi | |
| ## How to use | |
| 1) clone repo | |
| ``` | |
| git clone https://huggingface.co/datasets/armvectores/handwritten_text_detection | |
| cd handwritten_text_detection | |
| ``` | |
| 2) use data.yaml for training | |
| ``` | |
| from ultralytics import YOLO | |
| model = YOLO('yolov8n.pt') | |
| model.train(data='data.yaml', epochs=20) | |
| ``` | |
| ## Data sample | |
| <img src="blank_sample.png" width="700" /> |