Statuo commited on
Commit
ec4f07f
·
verified ·
1 Parent(s): ec6d1f1

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +35 -0
README.md CHANGED
@@ -11,6 +11,41 @@ model-index:
11
  results: []
12
  ---
13
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
14
  ![Kunou](https://huggingface.co/Sao10K/72B-Qwen2.5-Kunou-v1/resolve/main/knn.png)
15
 
16
  **Elder Sister Versions for Heavier Use!**
 
11
  results: []
12
  ---
13
 
14
+ Blessed be Sao who came down from the heavens to give us another 14b model. There's a 32b link but it's not up yet unfortunately. I'm sure when I post this it'll be up as a gotcha to me personally by God.
15
+
16
+ <br>
17
+
18
+ Anyway, before I get into the links you should know the following:
19
+ <br>
20
+ ChatML format seems to be the best fit here.
21
+ <br>
22
+ I disagree with Sao's sampler settings because the provided settings seem to have the settle into a kind of word-salad-y prose after a few messages. Try instead one of the following settings, or drop your Min P LOW. Like 0.02.
23
+ <br>
24
+ Min P: 0.02
25
+ <br>
26
+ Rep Penalty: 1.08
27
+ <br>
28
+ Presence Penalty: 1.1-1.11
29
+ <br>
30
+ OR
31
+ <br>
32
+ Min P: 0.02
33
+ <br>
34
+ XTC: 0.15, 0.3
35
+ <br>
36
+ DRY: 0.3, 1.75, 2
37
+
38
+ <br>
39
+ Raise the Min P slowly until you hit your sweet spot. Good luck, Tarnished.
40
+ <br>
41
+
42
+ [This is the EXL2 8bpw version of this model. You can find the original model here.](https://huggingface.co/Sao10K/14B-Qwen2.5-Kunou-v1)
43
+ <br>
44
+ [For the 6bpw version, go here.](https://huggingface.co/Statuo/Qwen-2.5-14b-Kunou-EXL2-6bpw)
45
+ <br>
46
+ [For the 4bpw version, go here.](https://huggingface.co/Statuo/Qwen-2.5-14b-Kunou-EXL2-4bpw)
47
+ ---
48
+
49
  ![Kunou](https://huggingface.co/Sao10K/72B-Qwen2.5-Kunou-v1/resolve/main/knn.png)
50
 
51
  **Elder Sister Versions for Heavier Use!**