nglaura commited on
Commit
fe026d9
1 Parent(s): 7349632

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +44 -0
README.md CHANGED
@@ -1,3 +1,47 @@
1
  ---
2
  license: apache-2.0
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: apache-2.0
3
+ task_categories:
4
+ - summarization
5
+ language:
6
+ - en
7
+ pretty_name: arXiv-Lay
8
  ---
9
+
10
+ # LoRaLay: A Multilingual and Multimodal Dataset for Long Range and Layout-Aware Summarization
11
+
12
+ A collaboration between [reciTAL](https://recital.ai/en/), [MLIA](https://mlia.lip6.fr/) (ISIR, Sorbonne Université), [Meta AI](https://ai.facebook.com/), and [Università di Trento](https://www.unitn.it/)
13
+
14
+ ## Arxiv-Lay dataset for summarization
15
+
16
+ ArXiv-Lay is an enhanced version of the arXiv summarization dataset, for which layout information is provided.
17
+
18
+ ### Data Fields
19
+
20
+ - `article_id`: article id
21
+ - `article_words`: sequence of words constituting the body of the article
22
+ - `article_bboxes`: sequence of corresponding word bounding boxes
23
+ - `norm_article_bboxes`: sequence of corresponding normalized word bounding boxes
24
+ - `abstract`: a string containing the abstract of the article
25
+ - `article_pdf_url`: URL of the article's PDF
26
+
27
+ ### Data Splits
28
+
29
+ This dataset has 3 splits: _train_, _validation_, and _test_. \
30
+
31
+ | Dataset Split | Number of Instances |
32
+ | ------------- | --------------------|
33
+ | Train | 122,189 |
34
+ | Validation | 4,374 |
35
+ | Test | 4,356 |
36
+
37
+
38
+ ## Citation
39
+
40
+ ``` latex
41
+ @article{nguyen2023loralay,
42
+ title={LoRaLay: A Multilingual and Multimodal Dataset for Long Range and Layout-Aware Summarization},
43
+ author={Nguyen, Laura and Scialom, Thomas and Piwowarski, Benjamin and Staiano, Jacopo},
44
+ journal={arXiv preprint arXiv:2301.11312},
45
+ year={2023}
46
+ }
47
+ ```