Update README.md
Browse files
README.md
CHANGED
|
@@ -18,6 +18,7 @@ language:
|
|
| 18 |
- es
|
| 19 |
metrics:
|
| 20 |
- recall
|
|
|
|
| 21 |
---
|
| 22 |
|
| 23 |
# CCRK: Improving the Consistency in Cross-Lingual Cross-Modal Retrieval with 1-to-K Contrastive Learning
|
|
@@ -282,4 +283,4 @@ About code, Our project is based on [CCLM](https://github.com/zengyan-97/CCLM).
|
|
| 282 |
|
| 283 |
About pretraining datasets, `zh,ja,de,fr,cs` texts in `cc3m` are translated by [UC2](https://github.com/zmykevin/UC2) while `zh,ja,de,fr,cs` texts in `sbu/coco/vg` are translated by [CCLM](https://github.com/zengyan-97/CCLM). For other languages `id,es,ru,tr`, we use `m2m_100_1.2B` model developed by [Meta AI](https://ai.facebook.com/research/) and [EasyNMT](https://github.com/UKPLab/EasyNMT) as a tool to translate all datasets from English.
|
| 284 |
|
| 285 |
-
Thanks for their great jobs!
|
|
|
|
| 18 |
- es
|
| 19 |
metrics:
|
| 20 |
- recall
|
| 21 |
+
pipeline_tag: visual-question-answering
|
| 22 |
---
|
| 23 |
|
| 24 |
# CCRK: Improving the Consistency in Cross-Lingual Cross-Modal Retrieval with 1-to-K Contrastive Learning
|
|
|
|
| 283 |
|
| 284 |
About pretraining datasets, `zh,ja,de,fr,cs` texts in `cc3m` are translated by [UC2](https://github.com/zmykevin/UC2) while `zh,ja,de,fr,cs` texts in `sbu/coco/vg` are translated by [CCLM](https://github.com/zengyan-97/CCLM). For other languages `id,es,ru,tr`, we use `m2m_100_1.2B` model developed by [Meta AI](https://ai.facebook.com/research/) and [EasyNMT](https://github.com/UKPLab/EasyNMT) as a tool to translate all datasets from English.
|
| 285 |
|
| 286 |
+
Thanks for their great jobs!
|