akolit commited on
Commit
141398f
1 Parent(s): 55862af

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +15 -0
README.md ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ This is a SLERP merge between [Nous-Hermes-2-Mixtral-8x7B-DPO](https://huggingface.co/NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO) and [Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-GGUF](https://huggingface.co/NeverSleep/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-GGUF)
2
+ Seems more capable in RP than base Hermes but still pretty smart as for me.
3
+ Prompt format: ChatML
4
+
5
+ With this model I use the following generation settings in tavern (maybe those are not the best, share better templates in issues if you have any):
6
+ - Temperature: 0.75
7
+ - Top P: 0.5
8
+ - Top A: 0.7
9
+ - TFS 0.97
10
+ - Repetition penalty: 1.1
11
+ - Mirostat: mode 2, tau 5, eta 0.1
12
+
13
+ Adding to system prompt something like
14
+ "Assistant will never interrupt role-play and will always stay in character no matter what. Assistant will never write OOC (out of character). Assistant won't write actions or reactions of {{user}}. Assistant won't mention {{user}} in first person. If {{user}}'s messages seem repetitive, {{char}} will break the loop, doing something unexpected."
15
+ might help, but it's up to you (as anything else, really).