Datasets:

Languages:
Japanese
Size Categories:
1K<n<10K
ArXiv:
Tags:
License:
parquet-converter commited on
Commit
c130a59
1 Parent(s): 8e8d061

Update parquet files

Browse files
README.md DELETED
@@ -1,105 +0,0 @@
1
- ---
2
- language:
3
- - ja
4
- license: cc-by-4.0
5
- size_categories:
6
- - 1K<n<10K
7
- task_categories:
8
- - visual-question-answering
9
- dataset_info:
10
- features:
11
- - name: image_id
12
- dtype: int64
13
- - name: url
14
- dtype: string
15
- - name: width
16
- dtype: int64
17
- - name: height
18
- dtype: int64
19
- - name: coco_id
20
- dtype: float64
21
- - name: flickr_id
22
- dtype: float64
23
- - name: qas
24
- list:
25
- - name: a_objects
26
- sequence: 'null'
27
- - name: answer
28
- dtype: string
29
- - name: q_objects
30
- sequence: 'null'
31
- - name: qa_id
32
- dtype: int64
33
- - name: question
34
- dtype: string
35
- - name: image
36
- dtype: image
37
- splits:
38
- - name: test
39
- num_bytes: 73348776.0
40
- num_examples: 500
41
- download_size: 355977607
42
- dataset_size: 357916367.0
43
- configs:
44
- - config_name: default
45
- data_files:
46
- - split: test
47
- path: data/test-*
48
- ---
49
-
50
- # JA-VG-VQA-500
51
-
52
-
53
- ## Dataset Description
54
-
55
- **JA-VG-VQA-500** is a 500-sample subset of [Japanese Visual Genome VQA dataset](https://github.com/yahoojapan/ja-vg-vqa).
56
- This dataset was used in the evaluation of [EvoVLM-JP-v1-7B](https://huggingface.co/SakanaAI/EvoVLM-JP-v1-7B).
57
- Please refer to our [report](https://arxiv.org/abs/2403.13187) and [blog](https://sakana.ai/evolutionary-model-merge/) for more details.
58
- We are grateful to the developers for making the dataset available under [Creative Commons Attribution 4.0 License](https://creativecommons.org/licenses/by/4.0/legalcode).
59
-
60
- - [Visual Genome](https://homes.cs.washington.edu/~ranjay/visualgenome/index.html)
61
- - [Japanese Visual Genome VQA dataset](https://github.com/yahoojapan/ja-vg-vqa)
62
-
63
- ## Usage
64
-
65
- Use the code below to get started with the dataset.
66
-
67
- ```python
68
- from datasets import load_dataset
69
-
70
- dataset = load_dataset("SakanaAI/JA-VG-VQA-500", split="test")
71
- ```
72
-
73
- See [our GitHub repository](https://github.com/SakanaAI/evolutionary-model-merge) to evaluate Japanese VLMs.
74
-
75
- ## Acknowledgement
76
-
77
- We would like to thank the developers of the source datasets for their contributions and for making their work available.
78
-
79
- ## Citation
80
-
81
- ```bibtex
82
- @article{Krishna2016VisualGC,
83
- title = {Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations},
84
- author. = {Ranjay Krishna and Yuke Zhu and Oliver Groth and Justin Johnson and Kenji Hata and Joshua Kravitz and Stephanie Chen and Yannis Kalantidis and Li-Jia Li and David A. Shamma and Michael S. Bernstein and Li Fei-Fei},
85
- journal = {International Journal of Computer Vision},
86
- year. = {2017},
87
- volume. = {123},
88
- pages. = {32-73},
89
- URL = {https://doi.org/10.1007/s11263-016-0981-7},
90
- doi = {10.1007/s11263-016-0981-7}
91
- }
92
- ```
93
-
94
- ```bibtex
95
- @InProceedings{C18-1163,
96
- author = "Shimizu, Nobuyuki and Rong, Na and Miyazaki, Takashi",
97
- title = "Visual Question Answering Dataset for Bilingual Image Understanding: A Study of Cross-Lingual Transfer Using Attention Maps",
98
- booktitle = "Proceedings of the 27th International Conference on Computational Linguistics",
99
- year = "2018",
100
- publisher = "Association for Computational Linguistics",
101
- pages = "1918--1928",
102
- location = "Santa Fe, New Mexico, USA",
103
- url = "http://aclweb.org/anthology/C18-1163"
104
- }
105
- ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
data/test-00000-of-00001.parquet → default/test/0000.parquet RENAMED
File without changes