Hugging Face's logo

Datasets:
katebor
/
TableEval

Tasks:
Table Question Answering
Table to Text
Modalities:
Image
Languages:
English
Size:
1K<n<10K
ArXiv:
License:
Dataset card Data Studio Files
xet
Community
katebor commited on
Commit
8e3483b
·
verified ·
1 Parent(s): 23ab4af

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +86 -1
README.md CHANGED
@@ -22,4 +22,89 @@ configs:
22
  path: ComTQA/SciGen/scigen.json
23
  - split: numericnlg
24
  path: ComTQA/numericNLG/numericnlg.json
25
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
22
  path: ComTQA/SciGen/scigen.json
23
  - split: numericnlg
24
  path: ComTQA/numericNLG/numericnlg.json
25
+ ---
26
+
27
+ # TableEval dataset
28
+
29
+ **TableEval** is developed to benchmark and compare the performance of (M)LLMs on tables from scientific vs. non-scientific sources, represented as images vs. text.
30
+ It comprises six data subsets derived from existing benchmarks for question answering (QA) and table-to-text (T2T) tasks, containing a total of **3017 tables** and **11312 instances**.
31
+ The scienfific subset includes tables from pre-prints and peer-reviewed scholarly publications, while the non-scientific subset involves tables from Wikipedia and financial reports.
32
+ Each table is available as a **PNG** image and in four textual formats: **HTML**, **XML**, **LaTeX**, and **Dictionary (Dict)**.
33
+ All task annotations are taken from the source datasets.
34
+
35
+
36
+ ## Overview and statistics
37
+
38
+
39
+ | Dataset | Task | Source | Image | Dict | LaTeX | HTML | XML |
40
+ |-----------------------|--------------------|-------------------|---------------|---------------|---------------|---------------|---------------|
41
+ | ComTQA (PubTables-1M) <img src='https://img.shields.io/badge/arXiv-2024-red'> <a href='https://arxiv.org/abs/2406.01326'><img src='https://img.shields.io/badge/PDF-blue'></a> <a href='https://huggingface.co/datasets/ByteDance/ComTQA'><img src='https://img.shields.io/badge/Dataset-gold'> | VQA | PubMed Central | ⬇️ | ⚙️ | ⚙️ | ⚙️ | 📄 |
42
+ | numericNLG <img src='https://img.shields.io/badge/ACL-2021-red'> <a href='https://aclanthology.org/2021.acl-long.115.pdf'><img src='https://img.shields.io/badge/PDF-blue'></a> <a href='https://huggingface.co/datasets/kasnerz/numericnlg?row=0'><img src='https://img.shields.io/badge/Dataset-gold'></a> | T2T | ACL Anthology | 📄 | ⬇️ | ⚙️ | ⬇️ | ⚙️ |
43
+ | SciGen <img src='https://img.shields.io/badge/arXiv-2021-red'> <a href='https://arxiv.org/abs/2104.08296'><img src='https://img.shields.io/badge/PDF-blue'></a> <a href='https://github.com/UKPLab/SciGen/tree/main'><img src='https://img.shields.io/badge/Dataset-gold'></a> | T2T | arXiv and ACL Anthology| 📄 | ⬇️ | 📄 | ⚙️ | ⚙️ |
44
+ | ComTQA (FinTabNet) <img src='https://img.shields.io/badge/arXiv-2024-red'> <a href='https://arxiv.org/abs/2406.01326'><img src='https://img.shields.io/badge/PDF-blue'></a> <a href='https://huggingface.co/datasets/ByteDance/ComTQA'><img src='https://img.shields.io/badge/Dataset-gold'> | VQA | Earnings reports of S&P 500 companies | 📄 | ⚙️ | ⚙️ | ⚙️ | ⚙️ |
45
+ | LogicNLG <img src='https://img.shields.io/badge/ACL-2020-red'> <a href='https://aclanthology.org/2020.acl-main.708/'><img src='https://img.shields.io/badge/PDF-blue'></a> <a href='https://huggingface.co/datasets/kasnerz/logicnlg'><img src='https://img.shields.io/badge/Dataset-gold'></a> | T2T | Wikipedia | ⚙️ | ⬇️ | ⚙️ | 📄 | ⚙️ |
46
+ | Logic2Text <img src='https://img.shields.io/badge/ACL-2020-red'> <a href='https://aclanthology.org/2020.findings-emnlp.190/'><img src='https://img.shields.io/badge/PDF-blue'></a> <a href='https://huggingface.co/datasets/kasnerz/logic2text'><img src='https://img.shields.io/badge/Dataset-gold'></a> | T2T | Wikipedia | ⚙️ | ⬇️ | ⚙️ | 📄 | ⚙️ |
47
+
48
+ **Symbol ⬇️ indicates formats already available in the given corpus, while 📄 and ⚙️ denote formats extracted from the table source files (e. g., article PDF, Wikipedia page) and generated from other formats in this study, respectively.
49
+
50
+ #### Number of tables per format and data subset
51
+
52
+ | Dataset | Image | Dict | LaTeX | HTML | XML |
53
+ |------------------------- |--------------------|-------------------|---------------|---------------|---------------|
54
+ | ComTQA (PubTables-1M) | 932 | 932 | 932 | 932 | 932 |
55
+ | numericNLG | 135 | 135 | 135 | 135 | 135 |
56
+ | SciGen | 1035 | 1035 | 928 | 985 | 961 |
57
+ | ComTQA (FinTabNet) | 659 | 659 | 659 | 659 | 659 |
58
+ | LogicNLG | 184 | 184 | 184 | 184 | 184 |
59
+ | Logic2Text | 72 | 72 | 72 | 72 | 72 |
60
+ | **Total** | **3017** | **3017** | **2910** | **2967** | **2943** |
61
+
62
+
63
+ #### Total number of instances per format and data subset
64
+
65
+ | Dataset | Image | Dict | LaTeX | HTML | XML |
66
+ |------------------------- |--------------------|-------------------|---------------|---------------|---------------|
67
+ | ComTQA (PubTables-1M) | 6232 | 6232 | 6232 | 6232 | 6232 |
68
+ | numericNLG | 135 | 135 | 135 | 135 | 135 |
69
+ | SciGen | 1035 | 1035 | 928 | 985 | 961 |
70
+ | ComTQA (FinTabNet) | 2838 | 2838 | 2838 | 2838 | 2838 |
71
+ | LogicNLG | 917 | 917 | 917 | 917 | 917 |
72
+ | Logic2Text | 155 | 155 | 155 | 155 | 155 |
73
+ | **Total** | **11312** | **11312** | **11205** | **11262** | **11238** |
74
+
75
+
76
+ ## Structure
77
+
78
+ ├── ComTQA
79
+ │ ├── FinTabNet
80
+ │ │ ├── comtqa_fintabnet.json
81
+ │ │ ├── comtqa_fintabnet_imgs.zip
82
+ │ ├── PubTab1M
83
+ │ │ ├── comtqa_pubtab1m.json
84
+ │ │ ├── comtqa_pubtab1m_imgs.zip
85
+ │ ├── Logic2Text
86
+ │ │ ├── logic2text.json
87
+ │ │ ├── logic2text_imgs.zip
88
+ │ ├── LogicNLG
89
+ │ │ ├── logicnlg.json
90
+ │ │ ├── logicnlg_imgs.zip
91
+ │ ├── SciGen
92
+ │ │ ├── scigen.json
93
+ │ │ ├── scigen_imgs.zip
94
+ │ ├── numericNLG
95
+ │ │ ├── numericnlg.json
96
+ └── └── └── numericnlg_imgs.zip
97
+
98
+ For more details on each subset, please, refer to the respective README.md files: FinTabNet, PubTab1M, Logic2Text, LogicNLG, SciGen, numericNLG.
99
+
100
+ ## Citation
101
+
102
+ TBA
103
+
104
+ ## Funding
105
+
106
+ This work has received funding through the DFG project [NFDI4DS](https://www.nfdi4datascience.de) (no. 460234259).
107
+
108
+ <div style="position: relative; width: 100%;">
109
+ <img src="NFDI4DS.png" alt="drawing" width="200" style="position: absolute; bottom: 0; right: 0;"/>
110
+ </div>