Commit
·
20ff549
1
Parent(s):
89c5980
Update README.md
Browse files
README.md
CHANGED
|
@@ -15,12 +15,7 @@ language:
|
|
| 15 |
tags:
|
| 16 |
- erzya
|
| 17 |
- mordovian
|
| 18 |
-
-
|
| 19 |
-
- pretraining
|
| 20 |
-
- embeddings
|
| 21 |
-
- masked-lm
|
| 22 |
-
- feature-extraction
|
| 23 |
-
- sentence-similarity
|
| 24 |
license: cc-by-sa-4.0
|
| 25 |
datasets:
|
| 26 |
- slone/myv_ru_2022
|
|
@@ -31,11 +26,11 @@ This a model to translate texts from the Erzya language (`myv`, cyrillic script)
|
|
| 31 |
|
| 32 |
It is described in the paper "The first neural machine translation system for the Erzya language".
|
| 33 |
|
| 34 |
-
This model is based on [facebook/mbart-large-50](https://huggingface.co/facebook/mbart-large-50)
|
| 35 |
- Added an extra language token `myv_XX` and 19K new BPE tokens for the Erzya language;
|
| 36 |
- Fine-tuned to translate to Erzya: first from Russian, then from all 11 languages.
|
| 37 |
|
| 38 |
-
The following code can be used to run translation using the model
|
| 39 |
|
| 40 |
```Python
|
| 41 |
from transformers import MBartForConditionalGeneration, MBart50Tokenizer
|
|
|
|
| 15 |
tags:
|
| 16 |
- erzya
|
| 17 |
- mordovian
|
| 18 |
+
- translation
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
license: cc-by-sa-4.0
|
| 20 |
datasets:
|
| 21 |
- slone/myv_ru_2022
|
|
|
|
| 26 |
|
| 27 |
It is described in the paper "The first neural machine translation system for the Erzya language".
|
| 28 |
|
| 29 |
+
This model is based on [facebook/mbart-large-50](https://huggingface.co/facebook/mbart-large-50), but with updated vocabulary and checkpoint:
|
| 30 |
- Added an extra language token `myv_XX` and 19K new BPE tokens for the Erzya language;
|
| 31 |
- Fine-tuned to translate to Erzya: first from Russian, then from all 11 languages.
|
| 32 |
|
| 33 |
+
The following code can be used to run translation using the model:
|
| 34 |
|
| 35 |
```Python
|
| 36 |
from transformers import MBartForConditionalGeneration, MBart50Tokenizer
|