Commit
•
abb7600
1
Parent(s):
448af64
Correct instruct template (via xzyun)
Browse files
README.md
CHANGED
|
@@ -38,10 +38,10 @@ Thanks to Kearm for helping with setting up LF on that node and to Featherless f
|
|
| 38 |
Model responds to Gemma instruct formatting, exactly like it's base model.
|
| 39 |
|
| 40 |
```
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
|
| 45 |
```
|
| 46 |
|
| 47 |
**Training config**
|
|
|
|
| 38 |
Model responds to Gemma instruct formatting, exactly like it's base model.
|
| 39 |
|
| 40 |
```
|
| 41 |
+
<bos><start_of_turn>user
|
| 42 |
+
{user message}<end_of_turn>
|
| 43 |
+
<start_of_turn>model
|
| 44 |
+
{response}<end_of_turn><eos>
|
| 45 |
```
|
| 46 |
|
| 47 |
**Training config**
|