Commit
·
7c414c9
1
Parent(s):
b82ddf2
Update README.md
Browse files
README.md
CHANGED
@@ -280,4 +280,180 @@ dataset_info:
|
|
280 |
num_examples: 285384149
|
281 |
download_size: 7483375536
|
282 |
dataset_size: 7999426267
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
283 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
280 |
num_examples: 285384149
|
281 |
download_size: 7483375536
|
282 |
dataset_size: 7999426267
|
283 |
+
license: cc-by-sa-4.0
|
284 |
+
task_categories:
|
285 |
+
- text-generation
|
286 |
+
language:
|
287 |
+
- sv
|
288 |
+
tags:
|
289 |
+
- newspapers
|
290 |
+
- historical
|
291 |
+
size_categories:
|
292 |
+
- 1B<n<10B
|
293 |
---
|
294 |
+
|
295 |
+
# kubhist2
|
296 |
+
|
297 |
+
## Dataset Description
|
298 |
+
|
299 |
+
- **Homepage: https://changeiskey.org**
|
300 |
+
- **Repository: https://github.com/ChangeIsKey/kubhist2**
|
301 |
+
- **Point of Contact: Simon Hengchen / iguanodon.ai**
|
302 |
+
|
303 |
+
|
304 |
+
### Dataset Summary
|
305 |
+
|
306 |
+
This is a version of the Kubhist 2 dataset originally created, curated and made available by Språkbanken Text (SBX) at the University of Gothenburg (Sweden) under the CC BY 4.0 license.
|
307 |
+
This is a a corpus of OCRed newspapers from Sweden spanning the 1640s to the 1900s.
|
308 |
+
The original data is available with many types of annotation in XML at https://spraakbanken.gu.se/en/resources/kubhist2.
|
309 |
+
A good description of the original data is available in this blog entry by Dana Dannélls: https://spraakbanken.gu.se/blogg/index.php/2019/09/15/the-kubhist-corpus-of-swedish-newspapers/.
|
310 |
+
|
311 |
+
If you use this dataset for academic research, cite it using the provided citation information at the bottom of this page.
|
312 |
+
|
313 |
+
In a nutshell, this hugginface dataset version offers:
|
314 |
+
- only the OCRed text
|
315 |
+
- available in decadal subsets
|
316 |
+
- one line per sentence, sentences shorter than 4 words were discarded
|
317 |
+
|
318 |
+
In total this dataset contains 2,819,065,590 tokens. A distribution of tokens per decade is available below.
|
319 |
+
|
320 |
+
License is CC BY 4.0 ShareAlike.
|
321 |
+
|
322 |
+
```bash
|
323 |
+
(env) simon@terminus:/mnt/user/cik/kubhist2 wc -w text/*/*.txt
|
324 |
+
39348 text/1640/1640.txt
|
325 |
+
4700 text/1650/1650.txt
|
326 |
+
8524 text/1660/1660.txt
|
327 |
+
2396 text/1670/1670.txt
|
328 |
+
199670 text/1680/1680.txt
|
329 |
+
487943 text/1690/1690.txt
|
330 |
+
619884 text/1700/1700.txt
|
331 |
+
265930 text/1710/1710.txt
|
332 |
+
355759 text/1720/1720.txt
|
333 |
+
856218 text/1730/1730.txt
|
334 |
+
1589508 text/1740/1740.txt
|
335 |
+
2211316 text/1750/1750.txt
|
336 |
+
5496545 text/1760/1760.txt
|
337 |
+
14434932 text/1770/1770.txt
|
338 |
+
22366170 text/1780/1780.txt
|
339 |
+
26768856 text/1790/1790.txt
|
340 |
+
36225842 text/1800/1800.txt
|
341 |
+
44510588 text/1810/1810.txt
|
342 |
+
65571094 text/1820/1820.txt
|
343 |
+
95359730 text/1830/1830.txt
|
344 |
+
143992956 text/1840/1840.txt
|
345 |
+
214538699 text/1850/1850.txt
|
346 |
+
392672066 text/1860/1860.txt
|
347 |
+
524802728 text/1870/1870.txt
|
348 |
+
695859650 text/1880/1880.txt
|
349 |
+
498244203 text/1890/1890.txt
|
350 |
+
31580335 text/1900/1900.txt
|
351 |
+
2819065590 total
|
352 |
+
|
353 |
+
```
|
354 |
+
|
355 |
+
|
356 |
+
### Languages
|
357 |
+
|
358 |
+
Swedish (nysvenska)
|
359 |
+
|
360 |
+
## Dataset Structure
|
361 |
+
|
362 |
+
One feature: `text`.
|
363 |
+
|
364 |
+
|
365 |
+
Load the whole corpus using
|
366 |
+
```python
|
367 |
+
dataset = load_dataset("ChangeIsKey/kubhist2")
|
368 |
+
```
|
369 |
+
or a decadal subset using
|
370 |
+
```python
|
371 |
+
dataset = load_dataset("ChangeIsKey/kubhist2", "decade")
|
372 |
+
```
|
373 |
+
The `decade` must be a string, valid values are within `range(1640, 1910, 10)`.
|
374 |
+
|
375 |
+
You can combine several decades using `concatenate_datasets` like this:
|
376 |
+
|
377 |
+
```python
|
378 |
+
from datasets import load_dataset, concatenate_datasets
|
379 |
+
|
380 |
+
ds_1800 = load_dataset("ChangeIsKey/kubhist2", "1800")
|
381 |
+
ds_1810 = load_dataset("ChangeIsKey/kubhist2", "1810")
|
382 |
+
ds_1820 = load_dataset("ChangeIsKey/kubhist2", "1820")
|
383 |
+
|
384 |
+
ds_1800_1820 = concatenate_datasets([
|
385 |
+
ds_1800["train"],
|
386 |
+
ds_1810["train"],
|
387 |
+
ds_1820["train"]
|
388 |
+
])
|
389 |
+
```
|
390 |
+
|
391 |
+
|
392 |
+
### Data Splits
|
393 |
+
|
394 |
+
The dataset has only one split, `train`.
|
395 |
+
|
396 |
+
## Dataset Creation
|
397 |
+
|
398 |
+
### Curation Rationale
|
399 |
+
|
400 |
+
The original data is in a highly-annotated XML format not ideally suited for basic NLP tasks such as unsupervised language modeling: information such as page numbers, fonts, etc. is less relevant and has thus been discarded.
|
401 |
+
Keeping only the running text of the newspaper and removing sentences shorter than 4 words further allows a 150x data size reduction (2.4TB --> 16GB).
|
402 |
+
|
403 |
+
### Source Data
|
404 |
+
|
405 |
+
The original data is available with many types of annotation in XML at https://spraakbanken.gu.se/en/resources/kubhist2.
|
406 |
+
|
407 |
+
#### Initial Data Collection and Normalization
|
408 |
+
|
409 |
+
See on Språkbanken Text's website.
|
410 |
+
|
411 |
+
#### Who are the source language producers?
|
412 |
+
|
413 |
+
Språkbanken Text: https://spraakbanken.gu.se/en/
|
414 |
+
|
415 |
+
### Personal and Sensitive Information
|
416 |
+
|
417 |
+
This is historical newspaper data, with the latest data published in 1909. Everyone mentioned in this dataset was probably already a public figure, and has been dead for a while.
|
418 |
+
|
419 |
+
## Considerations for Using the Data
|
420 |
+
|
421 |
+
### Discussion of Biases
|
422 |
+
|
423 |
+
This is historical data. As such, outdated views might be present in the data.
|
424 |
+
|
425 |
+
### Other Known Limitations
|
426 |
+
|
427 |
+
The data comes from an OCR process. The text is thus not perfect, especially so in the earlier decades.
|
428 |
+
|
429 |
+
## Additional Information
|
430 |
+
|
431 |
+
### Dataset Curators
|
432 |
+
|
433 |
+
This huggingface version of the data has been created by Simon Hengchen.
|
434 |
+
|
435 |
+
### Licensing Information
|
436 |
+
|
437 |
+
Creative Commons Attribution Share Alike 4.0: https://creativecommons.org/licenses/by-sa/4.0/
|
438 |
+
|
439 |
+
### Citation Information
|
440 |
+
|
441 |
+
You should always cite the original kubhist2 release, provided below as bibtex. If you want to additionally refer to this specific version, please also add a link to the huggingface page: https://huggingface.co/datasets/ChangeIsKey/kubhist2.
|
442 |
+
|
443 |
+
```bibtex
|
444 |
+
@misc{Kubhist2,
|
445 |
+
title = {The Kubhist Corpus, v2},
|
446 |
+
url = {https://spraakbanken.gu.se/korp/?mode=kubhist},
|
447 |
+
author = {Spr{\aa}kbanken},
|
448 |
+
year = {Downloaded in 2019},
|
449 |
+
organization = {Department of Swedish, University of Gothenburg}
|
450 |
+
}
|
451 |
+
|
452 |
+
```
|
453 |
+
|
454 |
+
### Acknowledgments
|
455 |
+
|
456 |
+
This dataset has been created in the context of the [ChangeIsKey!](https://www.changeiskey.org/) project funded by Riksbankens Jubileumsfond under reference number M21-0021, Change is Key! program.
|
457 |
+
The compute dedicated to the creation of the dataset has been provided by [iguanodon.ai](https://iguanodon.ai).
|
458 |
+
|
459 |
+
Many thanks got to Språkbanken Text for creating and curating this resource.
|