Fizzarolli
commited on
Commit
•
144ba56
1
Parent(s):
3f5a614
Update README.md
Browse files
README.md
CHANGED
@@ -1,2 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
# Dory 12b
|
2 |
redone instruct finetune of mistral nemo 12b
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
datasets:
|
4 |
+
- BeaverAI/Nemo-Inst-Tune-ds
|
5 |
+
language:
|
6 |
+
- en
|
7 |
+
library_name: transformers
|
8 |
+
---
|
9 |
# Dory 12b
|
10 |
redone instruct finetune of mistral nemo 12b
|
11 |
+
|
12 |
+
![image/gif](https://cdn-uploads.huggingface.co/production/uploads/634262af8d8089ebaefd410e/tP5kiebGrg8DJZbeBEc3Q.gif)
|
13 |
+
|
14 |
+
## Prompting
|
15 |
+
alpaca-like:
|
16 |
+
```
|
17 |
+
### System:
|
18 |
+
[Optional system prompt]
|
19 |
+
|
20 |
+
### Instruction:
|
21 |
+
[Query]
|
22 |
+
|
23 |
+
### Response:
|
24 |
+
[Response]<EOT>
|
25 |
+
|
26 |
+
### Instruction:
|
27 |
+
[...]
|
28 |
+
```
|
29 |
+
|
30 |
+
## Training details
|
31 |
+
Rank 64 QDoRA, trained on primarily Claude and Gemma 2 multiturn data (it's midnight and I'll probably write more details tomorrow)
|