Update README.md
Browse files
README.md
CHANGED
@@ -1,9 +1,14 @@
|
|
1 |
---
|
2 |
library_name: peft
|
|
|
|
|
3 |
---
|
4 |
-
|
|
|
5 |
|
6 |
-
|
7 |
|
|
|
8 |
|
9 |
-
-
|
|
|
|
1 |
---
|
2 |
library_name: peft
|
3 |
+
language:
|
4 |
+
- ru
|
5 |
---
|
6 |
+
Use in the same way as IlyaGusev/saiga2_7b_lora.
|
7 |
+
WARNING! Load tokenizer as AutoTokenizer.from_pretrained(model_path, use_fast=True)
|
8 |
|
9 |
+
Up to 60% faster generation and 35% training with HF because of different tokenizer.
|
10 |
|
11 |
+
## Model description
|
12 |
|
13 |
+
Instruction version (Saiga datasets) of Russian adaptation of LLaMa-2-7B by replacing the tokenizer.
|
14 |
+
Paper: Tikhomirov M.M., Chernyshev D.I., Impact of Tokenization on LLaMa Russian Adaptation (will be soon)
|