Datasets:
Tasks:
Visual Question Answering
Formats:
parquet
Languages:
English
Size:
10K - 100K
ArXiv:
Tags:
medical
License:
Commit
·
a8687a2
1
Parent(s):
6f1f4be
Update README.md
Browse files
README.md
CHANGED
|
@@ -56,8 +56,8 @@ Each instance consists of an image-question-answer triplet.
|
|
| 56 |
```
|
| 57 |
### Data Fields
|
| 58 |
- `'image'`: the image referenced by the question-answer pair.
|
| 59 |
-
- `'question'`: the
|
| 60 |
-
- `'answer'`: the
|
| 61 |
|
| 62 |
### Data Splits
|
| 63 |
The dataset is split into training, validation and test. The split is provided directly by the authors.
|
|
|
|
| 56 |
```
|
| 57 |
### Data Fields
|
| 58 |
- `'image'`: the image referenced by the question-answer pair.
|
| 59 |
+
- `'question'`: the question about the image.
|
| 60 |
+
- `'answer'`: the expected answer.
|
| 61 |
|
| 62 |
### Data Splits
|
| 63 |
The dataset is split into training, validation and test. The split is provided directly by the authors.
|