Update README.md
Browse files
README.md
CHANGED
@@ -35,11 +35,11 @@ Finally, explain why you chose your specific theme.
|
|
35 |
|
36 |
* Base model: senseable/WestLake-7B-v2 based of Mistral-7B-v0.1
|
37 |
* Context size: **8192** (even though Mistral-7B is 32k, WestLake was trained with 8k, and using a larger context is likely to cause problems)
|
38 |
-
* Prompt format: in general, Mistral based models are able to understand many prompt formats, but the following produce the best results, and are recommended
|
39 |
-
- **
|
40 |
-
-
|
41 |
-
-
|
42 |
-
-
|
43 |
|
44 |
## Merge Details
|
45 |
|
|
|
35 |
|
36 |
* Base model: senseable/WestLake-7B-v2 based of Mistral-7B-v0.1
|
37 |
* Context size: **8192** (even though Mistral-7B is 32k, WestLake was trained with 8k, and using a larger context is likely to cause problems)
|
38 |
+
* Prompt format: in general, Mistral based models are able to understand many prompt formats, but the following produce the best results, and are recommended (in order of preference)
|
39 |
+
- **Alpaca** (reported by senseable as working better than ChatML, and confirmed by me)
|
40 |
+
- ChatML (used during WestLake training)
|
41 |
+
- Mistral Instruct (original format from Mistral-7B)
|
42 |
+
- Zephyr (variant of ChatML which I have found to sometimes produce better results)
|
43 |
|
44 |
## Merge Details
|
45 |
|