Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- ru
|
| 4 |
+
- kbd
|
| 5 |
+
tags:
|
| 6 |
+
- translation
|
| 7 |
+
datasets:
|
| 8 |
+
- anzorq/kbd-ru-1.67M-temp
|
| 9 |
+
- 17753 Russian-Kabardian pairs of text
|
| 10 |
+
widget:
|
| 11 |
+
- text: "ru->kbd: Я иду домой."
|
| 12 |
+
example_title: "Я иду домой."
|
| 13 |
+
- text: "ru->kbd: Сколько у тебя яблок?"
|
| 14 |
+
example_title: "Сколько у тебя яблок?"
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
[google/t5-v1_1-small](google/t5-v1_1-small) model, pretrained on [anzorq/kbd-ru-1.67M-temp](anzorq/kbd-ru-1.67M-temp) and fine-tuned on **17753** Russian-Kabardian pairs of text.
|
| 18 |
+
|
| 19 |
+
kbd text uses custom latin script for optimization reasons.
|
| 20 |
+
|
| 21 |
+
Translation input should start with '**ru->kbd:** '.
|
| 22 |
+
|
| 23 |
+
**Tokenizer**: T5 sentencepiece, char, cased.
|