Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -25,10 +25,19 @@ configs:
|
|
| 25 |
<!-- DRAFT. Kurt Keutzer (UC Berkeley) contributed transcriptions. Numbers measured 2026-08-18. -->
|
| 26 |
|
| 27 |
**8,866 page images of Tibetan text with page-level Unicode transcriptions**, mostly
|
| 28 |
-
dbu-med (u-med) manuscripts with a small woodblock (uchen) portion.
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
and
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
|
| 33 |
Part of the **BDRC Tibetan OCR** release; see the model
|
| 34 |
[`BDRC/tibetan-ocr`](https://huggingface.co/BDRC/tibetan-ocr) and the
|
|
@@ -86,7 +95,7 @@ Single `train` split (training resource). For held-out evaluation use the separa
|
|
| 86 |
```bibtex
|
| 87 |
@misc{bdrc_berkeley_2026,
|
| 88 |
title = {Tibetan OCR --- Berkeley (dbu-med manuscripts)},
|
| 89 |
-
author = {
|
| 90 |
year = {2026},
|
| 91 |
howpublished = {Hugging Face},
|
| 92 |
note = {https://huggingface.co/datasets/BDRC/berkeley}
|
|
|
|
| 25 |
<!-- DRAFT. Kurt Keutzer (UC Berkeley) contributed transcriptions. Numbers measured 2026-08-18. -->
|
| 26 |
|
| 27 |
**8,866 page images of Tibetan text with page-level Unicode transcriptions**, mostly
|
| 28 |
+
dbu-med (u-med) manuscripts with a small woodblock (uchen) portion.
|
| 29 |
+
|
| 30 |
+
These works were transcribed by Geshe Dangsong Namgyal, from 2016 to the present, in his role as a Data System Analyst at University of California, Berkeley.
|
| 31 |
+
This work was initiated by Prof. Kurt Keutzer with the longstanding hope of producing training data for Handwritten Text Recognition for dbu med and OCR of blockprint of Tibetan.
|
| 32 |
+
|
| 33 |
+
The transcriptions focus on readability of the produced text rather than fidelity to the image. The orthographic shorthands are thus expanded to their full form, and some transcriptions might still contain some markup to represent emendations and inserts.
|
| 34 |
+
|
| 35 |
+
The data provided here is a modified version of the original files with the following changes:
|
| 36 |
+
- alignment image by image reviewed manually
|
| 37 |
+
- line breaks added manually
|
| 38 |
+
- manual filtering of transcriptions judged too distant from the enscribed text
|
| 39 |
+
|
| 40 |
+
These changes were implemented in a collaboration between BDRC and [Dharmaduta](https://dharmaduta.in) in the context of the project "The BDRC Etext Corpus", a grant by the Khyentse Foundation.
|
| 41 |
|
| 42 |
Part of the **BDRC Tibetan OCR** release; see the model
|
| 43 |
[`BDRC/tibetan-ocr`](https://huggingface.co/BDRC/tibetan-ocr) and the
|
|
|
|
| 95 |
```bibtex
|
| 96 |
@misc{bdrc_berkeley_2026,
|
| 97 |
title = {Tibetan OCR --- Berkeley (dbu-med manuscripts)},
|
| 98 |
+
author = {Kurt Keutzer, Geshe Dangsong Namgyal and Buddhist Digital Resource Center},
|
| 99 |
year = {2026},
|
| 100 |
howpublished = {Hugging Face},
|
| 101 |
note = {https://huggingface.co/datasets/BDRC/berkeley}
|