marianbasti
commited on
Commit
•
1f96ecd
1
Parent(s):
d97af6b
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,20 @@
|
|
1 |
---
|
2 |
license: llama2
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: llama2
|
3 |
+
datasets:
|
4 |
+
- bertin-project/alpaca-spanish
|
5 |
+
language:
|
6 |
+
- es
|
7 |
+
library_name: transformers
|
8 |
---
|
9 |
+
## Llama 2-13b-alpaca-spanish LoRA
|
10 |
+
This is a LoRA for Llama 2 13B trained on a translated [alpaca dataset](https://huggingface.co/datasets/bertin-project/alpaca-spanish) on an attempt to improve spanish performance of the Llama-2 foundation model with a conversational focus.
|
11 |
+
|
12 |
+
Base model used was [The Bloke's Llama-2-13B-fp16](https://huggingface.co/TheBloke/Llama-2-13B-fp16) trained in 4bit precision with an added padding token.
|
13 |
+
|
14 |
+
| Training parameteres | |
|
15 |
+
| ----------- | ----------- |
|
16 |
+
| LoRA scale | 2 |
|
17 |
+
| Epochs | 0.75 |
|
18 |
+
| Learning Rate| 2e-5 |
|
19 |
+
| Warmup Steps| 100 |
|
20 |
+
| Loss | 1.07 |
|