Update README.md
Browse files
README.md
CHANGED
@@ -10,9 +10,29 @@ tags:
|
|
10 |
- merge
|
11 |
|
12 |
---
|
13 |
-
|
14 |
|
15 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
16 |
|
17 |
## Merge Details
|
18 |
### Merge Method
|
|
|
10 |
- merge
|
11 |
|
12 |
---
|
13 |
+
## Siskin 0.1
|
14 |
|
15 |
+
![Хе-хе, сискин](https://qu.ax/WFju.png)
|
16 |
+
|
17 |
+
[The 0.2 version writes more — check it out here](https://huggingface.co/Nohobby/ThisWontWork4)
|
18 |
+
|
19 |
+
### Overview
|
20 |
+
|
21 |
+
Somewhat experimental merge of some Nemo models. Consists mostly of human data, so there should be very little gptisms/claudisms. Models work with about ~28k context and could maybe be stretched further.
|
22 |
+
|
23 |
+
Note: Both v0.1 and v0.2 can write in the first person despite the initial message. Specify the narrative style somewhere in the prompt to fix this if you don't like it.
|
24 |
+
|
25 |
+
### Prompt template: Mistral
|
26 |
+
|
27 |
+
```
|
28 |
+
<s>[INST] {input} [/INST] {output}</s>
|
29 |
+
```
|
30 |
+
|
31 |
+
### Quants
|
32 |
+
|
33 |
+
Static: [Q8](https://huggingface.co/Nohobby/ThisWontWork3-Q8_0-GGUF/resolve/main/thiswontwork3-q8_0.gguf?download=true) [Q6](https://huggingface.co/Nohobby/ThisWontWork3-Q6_K-GGUF/resolve/main/thiswontwork3-q6_k.gguf?download=true) [Q5_K_M](https://huggingface.co/Nohobby/ThisWontWork3-Q5_K_M-GGUF/resolve/main/thiswontwork3-q5_k_m.gguf?download=true)
|
34 |
+
|
35 |
+
(More coming soon)
|
36 |
|
37 |
## Merge Details
|
38 |
### Merge Method
|