Commit
•
4378cf0
1
Parent(s):
2042cb9
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,6 +1,21 @@
|
|
| 1 |
---
|
| 2 |
library_name: peft
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
## Training procedure
|
| 5 |
|
| 6 |
|
|
@@ -18,4 +33,4 @@ The following `bitsandbytes` quantization config was used during training:
|
|
| 18 |
### Framework versions
|
| 19 |
|
| 20 |
|
| 21 |
-
- PEFT 0.5.0
|
|
|
|
| 1 |
---
|
| 2 |
library_name: peft
|
| 3 |
+
license: cc-by-nc-2.0
|
| 4 |
+
language:
|
| 5 |
+
- fr
|
| 6 |
+
- en
|
| 7 |
+
tags:
|
| 8 |
+
- pytorch
|
| 9 |
+
- llama
|
| 10 |
+
- code
|
| 11 |
---
|
| 12 |
+
|
| 13 |
+
## Aria 7B V3
|
| 14 |
+
|
| 15 |
+
We decided to build a V3 of Aria 7B based on Mistral instruct instead of LLAMA 2. The base model has been quantized with Qlora to reduce the model size and trained on a high quality french dataset.
|
| 16 |
+
|
| 17 |
+
## Base Model : Mistral-7B-Instruct-v0.1
|
| 18 |
+
|
| 19 |
## Training procedure
|
| 20 |
|
| 21 |
|
|
|
|
| 33 |
### Framework versions
|
| 34 |
|
| 35 |
|
| 36 |
+
- PEFT 0.5.0
|