Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
- ar
|
| 6 |
+
- tl
|
| 7 |
+
- es
|
| 8 |
+
- ru
|
| 9 |
+
- ko
|
| 10 |
+
- ja
|
| 11 |
+
- ms
|
| 12 |
+
- id
|
| 13 |
+
- he
|
| 14 |
+
- hr
|
| 15 |
+
- pl
|
| 16 |
+
- pt
|
| 17 |
+
- af
|
| 18 |
+
- am
|
| 19 |
+
- az
|
| 20 |
+
- et
|
| 21 |
+
- eo
|
| 22 |
+
- th
|
| 23 |
+
- te
|
| 24 |
+
- ta
|
| 25 |
+
- tr
|
| 26 |
+
- uk
|
| 27 |
+
- ur
|
| 28 |
+
- uz
|
| 29 |
+
- vi
|
| 30 |
+
- ro
|
| 31 |
+
- km
|
| 32 |
+
- da
|
| 33 |
+
- de
|
| 34 |
+
- fi
|
| 35 |
+
- fa
|
| 36 |
+
- fr
|
| 37 |
+
- ga
|
| 38 |
+
- hu
|
| 39 |
+
- ky
|
| 40 |
+
- sa
|
| 41 |
+
tags:
|
| 42 |
+
- multilingual
|
| 43 |
+
- bert
|
| 44 |
+
- roberta
|
| 45 |
+
- xlmr
|
| 46 |
+
- bm
|
| 47 |
+
---
|
| 48 |
+
## Model type: Transformer-based masked language model
|
| 49 |
+
## Training data: No additional pretraining, merges two existing models
|
| 50 |
+
## Languages: 100+ languages
|
| 51 |
+
# Architecture:
|
| 52 |
+
- Base architectures:
|
| 53 |
+
- XLM-RoBERTa base (multilingual)
|
| 54 |
+
- BERT base cased (multilingual)
|
| 55 |
+
|
| 56 |
+
## Custom merging technique to combine weights from both base models into one unified model
|
| 57 |
+
|
| 58 |
+
# Mask token: <mask>
|