TheMrguiller
commited on
Commit
•
7d96081
1
Parent(s):
4ef22cc
Update README.md
Browse files
README.md
CHANGED
@@ -13,8 +13,52 @@ dataset_info:
|
|
13 |
num_bytes: 343036247.4
|
14 |
num_examples: 990
|
15 |
download_size: 1709055735
|
16 |
-
dataset_size: 1715181237
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
---
|
18 |
# Dataset Card for "BilbaoCaptions"
|
19 |
|
20 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
num_bytes: 343036247.4
|
14 |
num_examples: 990
|
15 |
download_size: 1709055735
|
16 |
+
dataset_size: 1715181237
|
17 |
+
language:
|
18 |
+
- en
|
19 |
+
tags:
|
20 |
+
- code
|
21 |
+
size_categories:
|
22 |
+
- 100B<n<1T
|
23 |
---
|
24 |
# Dataset Card for "BilbaoCaptions"
|
25 |
|
26 |
+
## Dataset Description
|
27 |
+
|
28 |
+
- **Homepage:** https://github.com/TheMrguiller/MUCSI_Modal
|
29 |
+
- **Repository:** https://github.com/TheMrguiller/MUCSI_Modal
|
30 |
+
- **Paper:** It is a follow up of the Flamingo model paper
|
31 |
+
- **Leaderboard:**
|
32 |
+
- **Point of Contact:** https://github.com/TheMrguiller/MUCSI_Modal
|
33 |
+
|
34 |
+
### Dataset Summary
|
35 |
+
|
36 |
+
This dataset was collected for a proyect for a master degree in Computation and Intelligent System from University of Deusto. It was done by students and recolected from webpages famous in the Basque Country: Deia and Getimages.
|
37 |
+
### Supported Tasks and Leaderboards
|
38 |
+
|
39 |
+
The dataset is prepared to used it for visual question-answering.
|
40 |
+
|
41 |
+
### Languages
|
42 |
+
|
43 |
+
The dataset is in english.
|
44 |
+
|
45 |
+
## Dataset Structure
|
46 |
+
|
47 |
+
### Data Fields
|
48 |
+
|
49 |
+
- `Caption`: This field has the description of the image.
|
50 |
+
- `Image`: This field has the image corresponding to the description.
|
51 |
+
|
52 |
+
### Data Splits
|
53 |
+
|
54 |
+
The dataset is split in 80% train and 20% test.
|
55 |
+
|
56 |
+
## Considerations for Using the Data
|
57 |
+
|
58 |
+
The dataset has some flaws regarding to the descriptions. The descriptions sometimes are to specific for a captioning task. There are also to many futbol match data, so it isnt to well balanced. There are also some description that are to generic.
|
59 |
+
|
60 |
+
## Additional Information
|
61 |
+
|
62 |
+
### Dataset Curators
|
63 |
+
|
64 |
+
The curators of this dataset where the students from the Masters degree in Computation and Inteligent Systems from University of Deusto.
|