Upload folder using huggingface_hub

README.md

@@ -1,83 +1,75 @@
 ---
 language:
-- bo
+  - bo
 license: cc0-1.0
-size_categories:
-- 1k<n<10k
 task_categories:
-- image-classification
+  - image-classification
 tags:
-- tibetan
-- manuscript
-- uchen
-- ume
-- script-classification
-- training-data
-dataset_info:
-  features:
-  - name: image
-    dtype: image
-  - name: label
-    dtype:
-      class_label:
-        names:
-          0: uchen
-          1: ume
-  - name: source
-    dtype: string
-  splits:
-  - name: train
-    num_examples: 3048
-  - name: validation
-    num_examples: 762
-  - name: test
-    num_examples: 762
+  - tibetan
+  - uchen
+  - ume
+  - manuscript
+size_categories:
+  - 10K<n<100K
+configs:
+  - config_name: default
+    data_files:
+      - split: train
+        path: train.parquet
+      - split: validation
+        path: validation.parquet
+      - split: test
+        path: test.parquet
+      - split: benchmark
+        path: benchmark.parquet
 ---
 
-# Uchen-Ume Classification Training Dataset
+# Uchen vs Umê classification (Parquet)
+
+Binary Tibetan script labels: **uchen** / **ume**.
+
+## Load with ``datasets``
 
-This repository contains the primary training and evaluation data for the **Uchen-Ume Classifier**. This dataset is used to train models to distinguish between formal "headed" block scripts (**Uchen**) and cursive "headless" script families (**Ume**).
+```python
+from datasets import load_dataset
 
-##  Repository Structure
+repo = "openpecha/uchen-ume-classification-benchmark"  # replace after upload
 
-The dataset is organized into three main directories to support different training and validation stages:
+train = load_dataset(repo, split="train")
+val = load_dataset(repo, split="validation")
+test = load_dataset(repo, split="test")
+bench = load_dataset(repo, split="benchmark")
 
-* **`source/`**: Contains the raw, original manuscript images as provided by BDRC.
-* **`preprocessed/`**: Contains three variants of the data used for different experiments:
-    * `whole_page`: Images resized (short-edge 224px) and **center-cropped** to 224x224.
-    * `patches_color`: Images split into **5 to 6 patches** per page with a **10% overlap**.
-    * `patches_clahe`: Same 10% overlap patch layout, but with **Contrast Limited Adaptive Histogram Equalization (CLAHE)** applied to improve stroke visibility in low-contrast scans.
-* **`benchmark/`**: A "Gold Standard" evaluation set containing exactly **5 images per class** across 18 script types. These images were selected after a strict taxonomical cleanup and were **never seen** by the model during the training, validation, or testing phases.
+# Or all splits at once:
+ds = load_dataset(repo)
+# ds["train"], ds["validation"], ds["test"], ds["benchmark"]
+```
 
-##  Project Context
-- **Project:** The BDRC Etext Corpus
-- **Developed by:** Dharmaduta in collaboration with BDRC
-- **Source of Images:** [Buddhist Digital Resource Center (BDRC)](https://www.bdrc.io)
-- **Funding:** Khyentse Foundation
-- **License:** CC0 (Public Domain)
+## Columns
 
-##  Dataset Stats & Splits
+| Column | Type | Description |
+|--------|------|-------------|
+| `id` | string | BDRC page id (filename stem) |
+| `image_bytes` | binary | Raw image file bytes (JPEG/PNG, …) |
+| `script` | string | `uchen` or `ume` |
 
-The following numbers represent the data used for the main Uchen-Ume binary classification task:
+Decode an image:
 
-- **Train:** 3,048 images
-- **Validation:** 762 images
-- **Test:** 762 images
-- **Total Samples:** 4,572
+```python
+from io import BytesIO
+from PIL import Image
 
-### Balanced Class Distribution
-- **Uchen (Class 0):** 2,286 images (includes `uchen_sugdring`, `uchen_sugthung`)
-- **Ume (Class 1):** 2,286 images (includes `petsuk`, `peri`, `tsegdrig`, `khyuyig`, etc.)
+row = train[0]
+img = Image.open(BytesIO(row["image_bytes"])).convert("RGB")
+```
 
-##  Selection & Tagging Methodology
+## Splits
 
-### Selection Process
-Images were selected from the BDRC digital library to provide a comprehensive representation of Tibetan script geometry. To maintain a clean binary signal, we manually excluded pages that were "Difficult," "Multi-script," or "Non-Tibetan" from this specific training set.
+| Split | Source |
+|-------|--------|
+| `train` / `validation` / `test` | Work-stratified split from training corpus (benchmark pages excluded) |
+| `benchmark` | Held-out diverse pages (`select_benchmark.py`) — **not** in train/val/test |
 
-### Annotation
-Tagging was performed using a custom-developed script classification tool developed by **Dharmaduta** and **BDRC**. 
-- **Expert Review:** A team of expert Tibetan paleographers reviewed every scan to assign ground-truth labels.
-- **Taxonomical Cleanup:** The `benchmark` folder specifically reflects a cleaned taxonomy, ensuring that the 18 sub-classes are correctly represented for the most rigorous model testing possible.
+## Training exclusion
 
-##  Acknowledgments
-We thank the **Buddhist Digital Resource Center** for their archive and the **Khyentse Foundation** for funding **The BDRC Etext Corpus** project.
+Pages in ``benchmark`` are excluded from train/val/test via ``benchmark_page_ids.json`` when building splits.

benchmark.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4bd950edb423259afd414a7815d444f8c3a025bbd82fb4d5b5db2a6fc853a6df
+size 40931078

test.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94176eac9f6ec887eb7995c5fee886b32e9c888f186a1bb73affd1f3dbf53b23
+size 556902632

train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b48881cbabdf7f50132c363a7e146b175069c84488e7d23e760730ab7495095
+size 5506315731

validation.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e82c22321ff25170779682a1bc8b46171db2c0234ef5ac0c28a399f8463712c
+size 630217530