Datasets:
Formats:
parquet
Size:
1K - 10K
Tags:
vision-language
multimodal
benchmarking
low-resource-languages
cross-lingual-evaluation
long-text-grounding
License:
Update the YAML front matter in README.md
Browse files
README.md
CHANGED
|
@@ -25,16 +25,15 @@ configs:
|
|
| 25 |
- split: train
|
| 26 |
path: "train/metadata.parquet"
|
| 27 |
dataset_info:
|
| 28 |
-
config_name: main
|
| 29 |
features:
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
splits:
|
| 39 |
- name: train
|
| 40 |
num_examples: 3415
|
|
|
|
| 25 |
- split: train
|
| 26 |
path: "train/metadata.parquet"
|
| 27 |
dataset_info:
|
|
|
|
| 28 |
features:
|
| 29 |
+
- name: id
|
| 30 |
+
dtype: string
|
| 31 |
+
- name: image
|
| 32 |
+
dtype: image
|
| 33 |
+
- name: text
|
| 34 |
+
dtype: string
|
| 35 |
+
- name: language
|
| 36 |
+
dtype: string
|
| 37 |
splits:
|
| 38 |
- name: train
|
| 39 |
num_examples: 3415
|