Update README.md
Browse files
README.md
CHANGED
@@ -50,4 +50,55 @@ configs:
|
|
50 |
path: data/fr-*
|
51 |
- split: ru
|
52 |
path: data/ru-*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
53 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
50 |
path: data/fr-*
|
51 |
- split: ru
|
52 |
path: data/ru-*
|
53 |
+
license: mit
|
54 |
+
task_categories:
|
55 |
+
- visual-question-answering
|
56 |
+
language:
|
57 |
+
- ar
|
58 |
+
- en
|
59 |
+
- es
|
60 |
+
- fr
|
61 |
+
- ru
|
62 |
+
pretty_name: XVNLI
|
63 |
+
size_categories:
|
64 |
+
- 1K<n<10K
|
65 |
---
|
66 |
+
|
67 |
+
# XVNLI
|
68 |
+
|
69 |
+
### This is a copy from the original repo: https://github.com/e-bug/iglue
|
70 |
+
|
71 |
+
|
72 |
+
If you use this dataset, please cite the original authors:
|
73 |
+
```bibtex
|
74 |
+
@inproceedings{bugliarello-etal-2022-iglue,
|
75 |
+
title = {{IGLUE}: A Benchmark for Transfer Learning across Modalities, Tasks, and Languages},
|
76 |
+
author = {Bugliarello, Emanuele and Liu, Fangyu and Pfeiffer, Jonas and Reddy, Siva and Elliott, Desmond and Ponti, Edoardo Maria and Vuli{\'c}, Ivan},
|
77 |
+
booktitle = {Proceedings of the 39th International Conference on Machine Learning},
|
78 |
+
pages = {2370--2392},
|
79 |
+
year = {2022},
|
80 |
+
editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan},
|
81 |
+
volume = {162},
|
82 |
+
series = {Proceedings of Machine Learning Research},
|
83 |
+
month = {17--23 Jul},
|
84 |
+
publisher = {PMLR},
|
85 |
+
pdf = {https://proceedings.mlr.press/v162/bugliarello22a/bugliarello22a.pdf},
|
86 |
+
url = {https://proceedings.mlr.press/v162/bugliarello22a.html},
|
87 |
+
}
|
88 |
+
```
|
89 |
+
|
90 |
+
### How to read the image
|
91 |
+
Due to a [bug](https://github.com/huggingface/datasets/issues/4796), the images cannot be stored as PIL.Image.Images directly but need to be converted to dataset.Images-. Hence, to load them, this additional step is required:
|
92 |
+
|
93 |
+
```python
|
94 |
+
from datasets import Image, load_dataset
|
95 |
+
|
96 |
+
ds = load_dataset("floschne/xvnli", split="en")
|
97 |
+
ds.map(
|
98 |
+
lambda sample: {
|
99 |
+
"image_t": [Image().decode_example(img) for img in sample["image"]],
|
100 |
+
},
|
101 |
+
remove_columns=["image"],
|
102 |
+
).rename_columns({"image_t": "image"})
|
103 |
+
|
104 |
+
```
|