Commit
•
3455559
1
Parent(s):
2eb8615
Update README.md
Browse files
README.md
CHANGED
|
@@ -42,7 +42,7 @@ The SEA-LION-BERT model is built on the MosaicBERT architecture and has a vocabu
|
|
| 42 |
|
| 43 |
For tokenization, the model employs our custom SEABPETokenizer, which is specially tailored for SEA languages, ensuring optimal model performance.
|
| 44 |
|
| 45 |
-
The training data for SEA-LION-BERT encompasses
|
| 46 |
|
| 47 |
- **Developed by:** Products Pillar, AI Singapore
|
| 48 |
- **Funded by:** Singapore NRF
|
|
|
|
| 42 |
|
| 43 |
For tokenization, the model employs our custom SEABPETokenizer, which is specially tailored for SEA languages, ensuring optimal model performance.
|
| 44 |
|
| 45 |
+
The training data for SEA-LION-BERT encompasses 790B tokens.
|
| 46 |
|
| 47 |
- **Developed by:** Products Pillar, AI Singapore
|
| 48 |
- **Funded by:** Singapore NRF
|