Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,31 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
task_categories:
|
| 4 |
+
- question-answering
|
| 5 |
+
language:
|
| 6 |
+
- or
|
| 7 |
+
- hi
|
| 8 |
+
tags:
|
| 9 |
+
- agriculture
|
| 10 |
+
- videoqna
|
| 11 |
+
- videos
|
| 12 |
+
size_categories:
|
| 13 |
+
- 1K<n<10K
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
The dataset is in XLS format with multiple sheets named for different languages. The dataset is primarily used for training and ground truth of answers that can be generated for agriculture related queries from the videos.
|
| 17 |
+
|
| 18 |
+
Each sheet has list of video urls (youtube links) and the question that can be asked, corresponding answers that can be generated from the videos, source of information in the answer and time stamps.
|
| 19 |
+
|
| 20 |
+
The sources of information could be:
|
| 21 |
+
|
| 22 |
+
Transcript: based on what one hears
|
| 23 |
+
Object: Based on an object shown
|
| 24 |
+
Scene description: based on what is described
|
| 25 |
+
Text overlay: based on text over lay shown in video
|
| 26 |
+
Corresponding time stamps are also provided.
|
| 27 |
+
|
| 28 |
+
The videos are in the following languages:
|
| 29 |
+
|
| 30 |
+
Hindi
|
| 31 |
+
Oriya
|