Not-For-All-Audiences
rAIfle commited on
Commit
b95614d
·
verified ·
1 Parent(s): c450c02

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +69 -0
README.md ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-nc-4.0
3
+ tags:
4
+ - not-for-all-audiences
5
+ ---
6
+ Quantized using 200 samples of 8192 tokens from an RP-oriented [PIPPA](https://huggingface.co/datasets/royallab/PIPPA-cleaned) dataset.
7
+
8
+ Branches:
9
+ - `main` -- `measurement.json`
10
+ - `3.5b6h` -- 3.5bpw, 6bit lm_head
11
+ - `6b6h` -- 6bpw, 6bit lm_head
12
+
13
+
14
+ Requires ExllamaV2 version 0.0.12 and up.
15
+
16
+ Original model link: [Envoid/Verdict-8x7B](https://huggingface.co/Envoid/Verdict-8x7B)
17
+
18
+ Original model README below.
19
+
20
+ ***
21
+ # Warning this model may output mature or disturbing content
22
+ ![](https://files.catbox.moe/q2rrja.jpg)
23
+ ## Verdict-8x7B
24
+
25
+ This model has been through so many merge steps that I failed to keep track of due to constantly needing to shift drive space around that I will simply give accredition to all of its component models.
26
+
27
+ retrieval-bar/Mixtral-8x7B-v0.1_case-briefs (no longer available)
28
+
29
+ ## New version available at:
30
+
31
+ [retrieval-bar/Mixtral-8x7B-v0.1_case-briefs](https://huggingface.co/retrieval-bar/Mixtral-8x7B-v0.1_case-briefs)
32
+
33
+ [mistralai/Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1)
34
+
35
+ [mistralai/Mixtral-8x7B-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-v0.1)
36
+
37
+ [Envoid/Augmentasanguis-8x7B](https://huggingface.co/Envoid/Augmentasanguis-8x7B)
38
+
39
+ [Envoid/CATA-8x7B](https://huggingface.co/Envoid/CATA-8x7B)
40
+
41
+ [crestf411/daybreak-mixtral-8x7b-hf](https://huggingface.co/crestf411/daybreak-mixtral-8x7b-hf)
42
+
43
+ [IBI-CAAI/MELT-Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/IBI-CAAI/MELT-Mixtral-8x7B-Instruct-v0.1)
44
+
45
+ unreleased experimental model that gets mentioned quite frequently
46
+
47
+ ### Many of the models were included multiple times in multiple intermediate steps
48
+
49
+ using a combination of slerp, linear, dare_ties and task_arithmetic
50
+
51
+ ## Model was tested in q8 gguf form (not included due to being from an old pull of llama.cpp that I'm too lazy to update)
52
+
53
+ # Warning: This models alignment is heavily diminished.
54
+
55
+ It responds to mixtral instruct formatting
56
+
57
+ # Example:
58
+ ```
59
+ [INST]Write me a haiku about violence.[/INST]
60
+ Fist clenched in anger,
61
+ Echoes of pain in the air,
62
+ Silence after storm.
63
+ ```
64
+
65
+ It can be a bit unruly in roleplay- sometimes to its advantage, sometimes to its detriment.
66
+
67
+ It's a perfectly usable instruct model and good for creative writing via instruct- its diminished alignment allowing it to respond to requests for more nuanced scences.
68
+
69
+ All-in-all I'm happy enough with the results that I plan to use it as a daily driver for a while to see what cracks show from long term use.