Update README.md
Browse files
README.md
CHANGED
|
@@ -18,7 +18,18 @@ tags:
|
|
| 18 |
|
| 19 |
### Model Description
|
| 20 |
|
| 21 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
|
| 23 |
|
| 24 |
|
|
|
|
| 18 |
|
| 19 |
### Model Description
|
| 20 |
|
| 21 |
+
LoRA finetuned Llama-2-7B-Chat on the Alpaca dataset.
|
| 22 |
+
|
| 23 |
+
Base model : https://huggingface.co/TheBloke/Llama-2-7B-Chat-fp16
|
| 24 |
+
|
| 25 |
+
https://github.com/rahuln32/LLMs-Finetuning-Safety-Drift-Fixes/tree/main/llama2
|
| 26 |
+
|
| 27 |
+
Params :
|
| 28 |
+
|
| 29 |
+
Batch size : 64
|
| 30 |
+
Gradient accumulation steps : 1
|
| 31 |
+
LR : 5e-5
|
| 32 |
+
Epochs : 1
|
| 33 |
|
| 34 |
|
| 35 |
|