Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -129,13 +129,16 @@ The dataset is built from scraping Wikipedia and doing HTML to Markdown conversi
|
|
| 129 |
|
| 130 |
## Comparison to existing sources
|
| 131 |
|
| 132 |
-
###
|
| 133 |
- [alexandrainst/nordjylland-news-image-captioning](https://huggingface.co/datasets/alexandrainst/nordjylland-news-image-captioning): 11.7k image pairs
|
| 134 |
- [alexandrainst/da-wit](https://huggingface.co/datasets/alexandrainst/da-wit): 168k image pairs
|
| 135 |
- [V4ldeLund/da-wiki-icc](https://huggingface.co/datasets/alexandrainst/alexandrainst/da-wit): 170k image pairs
|
| 136 |
|
| 137 |
### Qualitative comparison
|
|
|
|
| 138 |
By focusing on one language, we have managed to create a more accurate parsing pipeline for the Danish Wikipedia. For example, in the Wikipedia article for [Aarhus Hovedbanegård](https://da.wikipedia.org/wiki/Aarhus_Hovedbaneg%C3%A5rd) our dataset contains 29 images as opposed to only 9 in da-wit.
|
|
|
|
|
|
|
| 139 |
---
|
| 140 |
|
| 141 |
## Maintainer / Contact
|
|
|
|
| 129 |
|
| 130 |
## Comparison to existing sources
|
| 131 |
|
| 132 |
+
### Quantitative comparison
|
| 133 |
- [alexandrainst/nordjylland-news-image-captioning](https://huggingface.co/datasets/alexandrainst/nordjylland-news-image-captioning): 11.7k image pairs
|
| 134 |
- [alexandrainst/da-wit](https://huggingface.co/datasets/alexandrainst/da-wit): 168k image pairs
|
| 135 |
- [V4ldeLund/da-wiki-icc](https://huggingface.co/datasets/alexandrainst/alexandrainst/da-wit): 170k image pairs
|
| 136 |
|
| 137 |
### Qualitative comparison
|
| 138 |
+
|
| 139 |
By focusing on one language, we have managed to create a more accurate parsing pipeline for the Danish Wikipedia. For example, in the Wikipedia article for [Aarhus Hovedbanegård](https://da.wikipedia.org/wiki/Aarhus_Hovedbaneg%C3%A5rd) our dataset contains 29 images as opposed to only 9 in da-wit.
|
| 140 |
+
|
| 141 |
+
|
| 142 |
---
|
| 143 |
|
| 144 |
## Maintainer / Contact
|