A
1
Reddit: https://www.reddit.com/r/LocalLLaMA/
Discord: https://discord.gg/Y8H8uUtxc3
2
For inference I use KoboldCPP (latest version as they come out)
3
I create a .bat file for each model, and run it. Here's a .bat example for Mixtral 8x7b model:

title koboldcpp
:start
koboldcpp ^
--model mixtral-8x7b-instruct-v0.1.Q4_K_M.gguf ^
--contextsize 4096 ^
--usecublas ^
--gpulayers 6 ^
--threads 9
pause
goto start
4
When the GUI opens up, here's how I configure KoboldCPP settings.
First I change "quick presets" to [Default], then I move temperature and rep penalty to minimum and deactivate all samplers for the most deterministic output I can get.
5
I change the "Instract Tag Preset" (or "Start/End Sequence") to whatever prompt template is listed on the model's Huggingface page.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Some good benchmarks:
36
https://oobabooga.github.io/benchmark.html
37
https://dubesor.de/benchtable
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100