Librarian Bot: Add language metadata for dataset
#2
by
librarian-bot
- opened
README.md
CHANGED
|
@@ -1,10 +1,12 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
|
|
|
| 2 |
task_categories:
|
| 3 |
- zero-shot-classification
|
| 4 |
- text-retrieval
|
| 5 |
- image-to-text
|
| 6 |
- text-to-image
|
| 7 |
-
license: cc-by-4.0
|
| 8 |
---
|
| 9 |
This is an experimental augmentation of about 3 million synthetic captions from Recap-Datacomp-1B. This dataset includes about 2 million multilingual captions.
|
| 10 |
It attempts to balance for gender stereotypes, added race, union membership, and religion to a subsample. It also includes some permutations of sentence orders, and modificaitons of the number of items ("Two" is changed to "Three", "Four", etc.)
|
|
|
|
| 1 |
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
license: cc-by-4.0
|
| 5 |
task_categories:
|
| 6 |
- zero-shot-classification
|
| 7 |
- text-retrieval
|
| 8 |
- image-to-text
|
| 9 |
- text-to-image
|
|
|
|
| 10 |
---
|
| 11 |
This is an experimental augmentation of about 3 million synthetic captions from Recap-Datacomp-1B. This dataset includes about 2 million multilingual captions.
|
| 12 |
It attempts to balance for gender stereotypes, added race, union membership, and religion to a subsample. It also includes some permutations of sentence orders, and modificaitons of the number of items ("Two" is changed to "Three", "Four", etc.)
|