Update README.md
Browse files
README.md
CHANGED
@@ -48,60 +48,24 @@ VisualSimpleQA is a multimodal fact-seeking benchmark with two key features. Fir
|
|
48 |
Experiments on 15 LVLMs show that even state-of-the-art models such as GPT-4o achieve merely 60%+ correctness in multimodal fact-seeking QA on VisualSimpleQA and 30%+ on VisualSimpleQA-hard.
|
49 |
Furthermore, the decoupled evaluation based on this benchmark across different models highlights substantial opportunities for improvement in both visual and linguistic modules.
|
50 |
|
51 |
-
|
52 |
-
|
53 |
-
`original_image/`
|
54 |
-
This directory contains all image files, where each filename follows the format `original_image_{ID}.png`, matching the unique ID of the corresponding sample in VisualSimpleQA.
|
55 |
-
|
56 |
-
`cropped_image/`
|
57 |
-
This directory contains all cropped rationales from the original images. Each filename follows the format `cropped_image_{ID}.painting`, matching the unique ID of the corresponding sample in VisualSimpleQA.
|
58 |
-
|
59 |
-
`data.json`
|
60 |
-
This JSON file provides detailed information about each sample.
|
61 |
-
|
62 |
-
`hard_data.json`
|
63 |
-
This JSON file provides detailed information about each hard sample in the following format:
|
64 |
-
|
65 |
-
**Example:**
|
66 |
-
```json
|
67 |
-
{
|
68 |
-
"id": 369,
|
69 |
-
"multimodal_question": "Which institution did the creator of this cartoon duck donate her natural science-related paintings to?",
|
70 |
-
"answer": "The Armitt Museum, Gallery, Library",
|
71 |
-
"rationale": "Jemima Puddle-Duck",
|
72 |
-
"text_only_question": "Which institution did the creator of Jemima Puddle-Duck donate her natural science-related paintings to?",
|
73 |
-
"image_source": "https://www.gutenberg.org/files/14814/14814-h/images/15-tb.jpg",
|
74 |
-
"evidence": "https://www.armitt.com/beatrix-potter-exhibition/\nhttps://en.wikipedia.org/wiki/Beatrix_Potter",
|
75 |
-
"resolution": "400x360",
|
76 |
-
"proportion_of_roi": 0.2232,
|
77 |
-
"category": "academic and education",
|
78 |
-
"text_in_image": "absence",
|
79 |
-
"rationale_granularity": "fine-grained"
|
80 |
-
}
|
81 |
```
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
|
90 |
-
|
91 |
-
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
96 |
-
# Randomly select one sample
|
97 |
-
random_sample = random.choice(data)
|
98 |
-
image_id = random_sample.get('id')
|
99 |
-
image_path = f'./original_image/original_image_{image_id}.png'
|
100 |
-
multimodal_question = random_sample.get('multimodal_question')
|
101 |
-
text_only_question = random_sample.get('text_only_question')
|
102 |
-
answer = random_sample.get('answer')
|
103 |
```
|
104 |
|
|
|
105 |
## Disclaimer
|
106 |
|
107 |
This dataset contains images collected from various sources. The authors do NOT claim ownership or copyright over the images. The images may be subject to third-party rights, and users are solely responsible for verifying the legal status of any content before use.
|
|
|
48 |
Experiments on 15 LVLMs show that even state-of-the-art models such as GPT-4o achieve merely 60%+ correctness in multimodal fact-seeking QA on VisualSimpleQA and 30%+ on VisualSimpleQA-hard.
|
49 |
Furthermore, the decoupled evaluation based on this benchmark across different models highlights substantial opportunities for improvement in both visual and linguistic modules.
|
50 |
|
51 |
+
**Data Example:**
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
```
|
53 |
+
{'id': 369,
|
54 |
+
'multimodal_question': 'Which institution did the creator of this cartoon duck donate her natural science-related paintings to?',
|
55 |
+
'answer': 'The Armitt Museum, Gallery, Library', 'rationale': 'Jemima Puddle-Duck',
|
56 |
+
'text_only_question': 'Which institution did the creator of Jemima Puddle-Duck donate her natural science-related paintings to?',
|
57 |
+
'image_source': 'https://www.gutenberg.org/files/14814/14814-h/images/15-tb.jpg',
|
58 |
+
'evidence': 'https://www.armitt.com/beatrix-potter-exhibition/\nhttps://en.wikipedia.org/wiki/Beatrix_Potter',
|
59 |
+
'resolution': '400x360',
|
60 |
+
'proportion_of_roi': '0.2232',
|
61 |
+
'category': 'research and education',
|
62 |
+
'text_in_image': 'absence',
|
63 |
+
'rationale_granularity': 'fine-grained',
|
64 |
+
'image': <PIL.PngImagePlugin.PngImageFile image mode=RGB size=400x360 at 0x7FE82C270D70>,
|
65 |
+
'cropped_image': <PIL.PngImagePlugin.PngImageFile image mode=RGB size=164x196 at 0x7FE82C329550>}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
66 |
```
|
67 |
|
68 |
+
|
69 |
## Disclaimer
|
70 |
|
71 |
This dataset contains images collected from various sources. The authors do NOT claim ownership or copyright over the images. The images may be subject to third-party rights, and users are solely responsible for verifying the legal status of any content before use.
|