Datasets:
Tasks:
Visual Question Answering
Modalities:
Image
Sub-tasks:
visual-question-answering
Languages:
English
ArXiv:
License:
Updated dataset card
Browse files
README.md
CHANGED
@@ -36,13 +36,7 @@ task_ids:
|
|
36 |
- [Data Fields](#data-fields)
|
37 |
- [Data Splits](#data-splits)
|
38 |
- [Dataset Creation](#dataset-creation)
|
39 |
-
- [Curation Rationale](#curation-rationale)
|
40 |
-
- [Source Data](#source-data)
|
41 |
-
- [Annotations](#annotations)
|
42 |
-
- [Personal and Sensitive Information](#personal-and-sensitive-information)
|
43 |
- [Considerations for Using the Data](#considerations-for-using-the-data)
|
44 |
-
- [Social Impact of Dataset](#social-impact-of-dataset)
|
45 |
-
- [Discussion of Biases](#discussion-of-biases)
|
46 |
- [Other Known Limitations](#other-known-limitations)
|
47 |
- [Additional Information](#additional-information)
|
48 |
- [Dataset Curators](#dataset-curators)
|
@@ -60,11 +54,48 @@ task_ids:
|
|
60 |
|
61 |
### Dataset Summary
|
62 |
|
63 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
64 |
|
65 |
### Supported Tasks and Leaderboards
|
66 |
|
67 |
-
|
68 |
|
69 |
### Languages
|
70 |
|
@@ -74,11 +105,12 @@ The dataset is currently only available in English. To extend the dataset to oth
|
|
74 |
|
75 |
### Data Instances
|
76 |
|
77 |
-
|
|
|
|
|
78 |
|
79 |
### Data Fields
|
80 |
|
81 |
-
[More Information Needed]
|
82 |
|
83 |
```
|
84 |
features = datasets.Features(
|
@@ -94,48 +126,15 @@ features = datasets.Features(
|
|
94 |
|
95 |
### Data Splits
|
96 |
|
97 |
-
|
98 |
|
99 |
## Dataset Creation
|
100 |
|
101 |
-
|
102 |
-
|
103 |
-
[More Information Needed]
|
104 |
-
|
105 |
-
### Source Data
|
106 |
|
107 |
-
#### Initial Data Collection and Normalization
|
108 |
-
|
109 |
-
[More Information Needed]
|
110 |
-
|
111 |
-
#### Who are the source language producers?
|
112 |
-
|
113 |
-
[More Information Needed]
|
114 |
-
|
115 |
-
### Annotations
|
116 |
-
|
117 |
-
#### Annotation process
|
118 |
-
|
119 |
-
[More Information Needed]
|
120 |
-
|
121 |
-
#### Who are the annotators?
|
122 |
-
|
123 |
-
[More Information Needed]
|
124 |
-
|
125 |
-
### Personal and Sensitive Information
|
126 |
-
|
127 |
-
[More Information Needed]
|
128 |
|
129 |
## Considerations for Using the Data
|
130 |
|
131 |
-
### Social Impact of Dataset
|
132 |
-
|
133 |
-
[More Information Needed]
|
134 |
-
|
135 |
-
### Discussion of Biases
|
136 |
-
|
137 |
-
[More Information Needed]
|
138 |
-
|
139 |
### Other Known Limitations
|
140 |
|
141 |
[More Information Needed]
|
@@ -144,7 +143,6 @@ features = datasets.Features(
|
|
144 |
|
145 |
### Dataset Curators
|
146 |
|
147 |
-
[More Information Needed]
|
148 |
Adam Dahlgren Lindström - [email protected]
|
149 |
|
150 |
### Licensing Information
|
|
|
36 |
- [Data Fields](#data-fields)
|
37 |
- [Data Splits](#data-splits)
|
38 |
- [Dataset Creation](#dataset-creation)
|
|
|
|
|
|
|
|
|
39 |
- [Considerations for Using the Data](#considerations-for-using-the-data)
|
|
|
|
|
40 |
- [Other Known Limitations](#other-known-limitations)
|
41 |
- [Additional Information](#additional-information)
|
42 |
- [Dataset Curators](#dataset-curators)
|
|
|
54 |
|
55 |
### Dataset Summary
|
56 |
|
57 |
+
Dataset for compositional multimodal mathematical reasoning based on CLEVR.
|
58 |
+
|
59 |
+
#### Loading the data, preprocessing text with CLIP
|
60 |
+
|
61 |
+
```
|
62 |
+
from transformers import CLIPPreprocessor
|
63 |
+
from datasets import load_dataset, DownloadConfig
|
64 |
+
|
65 |
+
|
66 |
+
dl_config = DownloadConfig(resume_download=True,
|
67 |
+
num_proc=8,
|
68 |
+
force_download=True)
|
69 |
+
|
70 |
+
# Load 'general' instance of dataset
|
71 |
+
dataset = load_dataset('dali-does/clevr-math', download_config=dl_config)
|
72 |
+
|
73 |
+
# Load version with only multihop in test data
|
74 |
+
dataset_multihop = load_dataset('dali-does/clevr-math', 'multihop',
|
75 |
+
download_config=dl_config)
|
76 |
+
|
77 |
+
|
78 |
+
model_path = "openai/clip-vit-base-patch32"
|
79 |
+
extractor = CLIPProcessor.from_pretrained(model_path)
|
80 |
+
def transform_tokenize(e):
|
81 |
+
e['image'] = [image.convert('RGB') for image in e['image']]
|
82 |
+
return extractor(text=e['question'],
|
83 |
+
images=e['image'],
|
84 |
+
padding=True)
|
85 |
+
|
86 |
+
dataset = dataset.map(transform_tokenize,
|
87 |
+
batched=True,
|
88 |
+
num_proc=8,
|
89 |
+
padding='max_length')
|
90 |
+
|
91 |
+
dataset_subtraction = dataset.filter(lambda e:
|
92 |
+
e['template'].startswith('subtraction'), num_proc=4)
|
93 |
+
```
|
94 |
+
|
95 |
|
96 |
### Supported Tasks and Leaderboards
|
97 |
|
98 |
+
Leaderboard will be announced at a later date.
|
99 |
|
100 |
### Languages
|
101 |
|
|
|
105 |
|
106 |
### Data Instances
|
107 |
|
108 |
+
* `general` containing the default version with multihop questions in train and test
|
109 |
+
* `multihop` containing multihop questions only in test data to test generalisation of reasoning
|
110 |
+
|
111 |
|
112 |
### Data Fields
|
113 |
|
|
|
114 |
|
115 |
```
|
116 |
features = datasets.Features(
|
|
|
126 |
|
127 |
### Data Splits
|
128 |
|
129 |
+
train/val/test
|
130 |
|
131 |
## Dataset Creation
|
132 |
|
133 |
+
Data is generated using code provided with the CLEVR-dataset, using blender and templates constructed by the dataset curators.
|
|
|
|
|
|
|
|
|
134 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
135 |
|
136 |
## Considerations for Using the Data
|
137 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
138 |
### Other Known Limitations
|
139 |
|
140 |
[More Information Needed]
|
|
|
143 |
|
144 |
### Dataset Curators
|
145 |
|
|
|
146 |
Adam Dahlgren Lindström - [email protected]
|
147 |
|
148 |
### Licensing Information
|