Update README.md
Browse files
README.md
CHANGED
@@ -1,6 +1,16 @@
|
|
1 |
## Model Info
|
2 |
Merge of my [VicUnlocked-alpaca-half-30b LoRA](https://huggingface.co/Aeala/VicUnlocked-alpaca-half-30b-LoRA)
|
3 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
## Benchmarks
|
5 |
wikitext2: 4.372413635253906
|
6 |
ptb-new: 24.69171714782715
|
|
|
1 |
## Model Info
|
2 |
Merge of my [VicUnlocked-alpaca-half-30b LoRA](https://huggingface.co/Aeala/VicUnlocked-alpaca-half-30b-LoRA)
|
3 |
|
4 |
+
**Important Note**: While this is trained on a cleaned ShareGPT dataset like Vicuna used, this was trained in the *Alpaca* format, so prompting should be something like:
|
5 |
+
|
6 |
+
```
|
7 |
+
### Instruction:
|
8 |
+
|
9 |
+
<prompt> (without the <>)
|
10 |
+
|
11 |
+
### Response:
|
12 |
+
```
|
13 |
+
|
14 |
## Benchmarks
|
15 |
wikitext2: 4.372413635253906
|
16 |
ptb-new: 24.69171714782715
|