Update README.md
Browse files
README.md
CHANGED
@@ -6,4 +6,79 @@ language:
|
|
6 |
pretty_name: T
|
7 |
size_categories:
|
8 |
- 1M<n<10M
|
9 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
pretty_name: T
|
7 |
size_categories:
|
8 |
- 1M<n<10M
|
9 |
+
---
|
10 |
+
|
11 |
+
# TempKB
|
12 |
+
|
13 |
+
## Overview
|
14 |
+
|
15 |
+
TempKB is a comprehensive collection of knowledge graph data designed to train ML models on knowledge graph completion and reasoning tasks.
|
16 |
+
|
17 |
+
## Dataset Structure
|
18 |
+
|
19 |
+
The dataset is organized into the following main components:
|
20 |
+
|
21 |
+
Train Set: 7764337 instances for training models.
|
22 |
+
|
23 |
+
Validation Set: 117510 instances for validating model performance.
|
24 |
+
|
25 |
+
Test Set: 117327 instances for final evaluation.
|
26 |
+
|
27 |
+
## Data Fields
|
28 |
+
|
29 |
+
Each instance in the dataset contains the following fields:
|
30 |
+
|
31 |
+
- `head`: the head entity
|
32 |
+
- `relation`: the relationships between the head and tail entities
|
33 |
+
- `tail`: the tail entity
|
34 |
+
- `start`: start time of the event
|
35 |
+
- `end`: end time of the event
|
36 |
+
- `type`: temporal information type, which is one of the following:
|
37 |
+
- `NONE`: no start or end time
|
38 |
+
- `TIMESTAMP`: start time only
|
39 |
+
- `INTERVAL`: start and end time
|
40 |
+
|
41 |
+
## Example Instance
|
42 |
+
|
43 |
+
```json
|
44 |
+
{
|
45 |
+
"head": "Martin Bača"
|
46 |
+
"relation": "member of sports team"
|
47 |
+
"tail": "Czech Republic national under-18 football team"
|
48 |
+
"start": "2002-01-01"
|
49 |
+
"end": "2002-01-01"
|
50 |
+
"type": "INTERVAL"
|
51 |
+
}
|
52 |
+
```
|
53 |
+
|
54 |
+
## Usage
|
55 |
+
|
56 |
+
To use this dataset, you can load it using the Hugging Face datasets library:
|
57 |
+
|
58 |
+
```python
|
59 |
+
from datasets import load_dataset
|
60 |
+
|
61 |
+
dataset = load_dataset('ESITime/TempKB')
|
62 |
+
```
|
63 |
+
|
64 |
+
## License
|
65 |
+
|
66 |
+
This dataset is unlicensed. Feel free to use it as you want.
|
67 |
+
|
68 |
+
## Citation
|
69 |
+
|
70 |
+
If you use this dataset in your research, please cite it as follows:
|
71 |
+
|
72 |
+
```bib
|
73 |
+
@dataset{username_dataset_name_year,
|
74 |
+
author = {Phuong Ngo},
|
75 |
+
title = {TempKB},
|
76 |
+
year = {2024},
|
77 |
+
publisher = {Hugging Face},
|
78 |
+
url = {https://huggingface.co/datasets/ESITime/TempKB}
|
79 |
+
}
|
80 |
+
```
|
81 |
+
|
82 |
+
## Acknowledgement
|
83 |
+
|
84 |
+
This dataset is derived from the collection of knowledge graph datasets including ICEWS05-15, ICEWS14, Wikidata, and YAGO. The datasets are downloaded from the processed version by Facebook Research: https://github.com/facebookresearch/tkbc/
|