Sorawiz commited on
Commit
8db4887
·
verified ·
1 Parent(s): 975fff9

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +112 -3
README.md CHANGED
@@ -6,13 +6,52 @@ base_model:
6
  - TheDrummer/Cydonia-24B-v2.1
7
  - Sorawiz/MistralSmall-Creative-24B
8
  - ReadyArt/Forgotten-Safeword-24B-V2.2
 
 
 
 
 
9
  library_name: transformers
10
  tags:
11
  - mergekit
12
  - merge
13
 
14
  ---
15
- # merge
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
16
 
17
  This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
18
 
@@ -28,13 +67,83 @@ The following models were included in the merge:
28
  * [TroyDoesAI/BlackSheep-24B](https://huggingface.co/TroyDoesAI/BlackSheep-24B)
29
  * [OddTheGreat/Apparatus_24B](https://huggingface.co/OddTheGreat/Apparatus_24B)
30
  * [TheDrummer/Cydonia-24B-v2.1](https://huggingface.co/TheDrummer/Cydonia-24B-v2.1)
31
- * [Sorawiz/MistralSmall-Creative-24B](https://huggingface.co/Sorawiz/MistralSmall-Creative-24B)
32
 
33
  ### Configuration
34
 
35
  The following YAML configuration was used to produce this model:
36
 
37
  ```yaml
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
38
  models:
39
  - model: ReadyArt/Forgotten-Safeword-24B-V2.2
40
  - model: OddTheGreat/Apparatus_24B
@@ -45,7 +154,7 @@ models:
45
  parameters:
46
  density: 1.00
47
  weight: 1.00
48
- - model: Sorawiz/MistralSmall-Creative-24B
49
  parameters:
50
  density: 0.50
51
  weight: 0.50
 
6
  - TheDrummer/Cydonia-24B-v2.1
7
  - Sorawiz/MistralSmall-Creative-24B
8
  - ReadyArt/Forgotten-Safeword-24B-V2.2
9
+ - ReadyArt/Forgotten-Safeword-24B-V2.0
10
+ - trashpanda-org/MS-24B-Mullein-v1-lora
11
+ - ReadyArt/Forgotten-Abomination-24B-V2.2
12
+ - allura-org/MS3-24B-Roselily-Creative
13
+ - Nohobby/MS3-Tantum-24B-v0.1
14
  library_name: transformers
15
  tags:
16
  - mergekit
17
  - merge
18
 
19
  ---
20
+
21
+ A highly conversational AI model that excels in dialogue but may become incoherent over long responses without a well-crafted system prompt.
22
+
23
+ # Chat Template
24
+
25
+ Mistral Instruct
26
+
27
+ ```
28
+ {{ if .System }}<|im_start|>system
29
+ {{ .System }}<|im_end|>
30
+ {{ end }}{{ if .Prompt }}<|im_start|>user
31
+ {{ .Prompt }}<|im_end|>
32
+ {{ end }}<|im_start|>assistant
33
+ {{ .Response }}<|im_end|>
34
+ ```
35
+
36
+ ChatML
37
+
38
+ ```
39
+ {{ if .System }}<|im_start|>system
40
+ {{ .System }}<|im_end|>
41
+ {{ end }}{{ if .Prompt }}<|im_start|>user
42
+ {{ .Prompt }}<|im_end|>
43
+ {{ end }}<|im_start|>assistant
44
+ {{ .Response }}{{ if .Response }}<|im_end|>{{ end }}
45
+ ```
46
+
47
+ # GGUF
48
+
49
+ Thank you [mradermacher](https://huggingface.co/mradermacher) for creating the GGUF versions of this model.
50
+
51
+ * Static quants - [mradermacher/MistralSmall-Creative-24B-Realist-GGUF](https://huggingface.co/mradermacher/MistralSmall-Creative-24B-Realist-GGUF)
52
+ * Imatrix quants - [mradermacher/MistralSmall-Creative-24B-Realist-i1-GGUF](https://huggingface.co/mradermacher/MistralSmall-Creative-24B-Realist-i1-GGUF)
53
+
54
+ # Merge
55
 
56
  This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
57
 
 
67
  * [TroyDoesAI/BlackSheep-24B](https://huggingface.co/TroyDoesAI/BlackSheep-24B)
68
  * [OddTheGreat/Apparatus_24B](https://huggingface.co/OddTheGreat/Apparatus_24B)
69
  * [TheDrummer/Cydonia-24B-v2.1](https://huggingface.co/TheDrummer/Cydonia-24B-v2.1)
70
+ * [Sorawiz/MistralSmall-Creative-24B-Stock]
71
 
72
  ### Configuration
73
 
74
  The following YAML configuration was used to produce this model:
75
 
76
  ```yaml
77
+ name: Sorawiz/MS-Creative-24B-Test-A
78
+ merge_method: dare_ties
79
+ base_model: ReadyArt/Forgotten-Safeword-24B-V2.2
80
+ models:
81
+ - model: ReadyArt/Forgotten-Safeword-24B-V2.2
82
+ parameters:
83
+ weight: 0.05
84
+ - model: ReadyArt/Forgotten-Abomination-24B-V2.2
85
+ parameters:
86
+ weight: 0.20
87
+ - model: OddTheGreat/Apparatus_24B
88
+ parameters:
89
+ weight: 0.20
90
+ - model: Darkknight535/WinterEngine-24B-Instruct
91
+ parameters:
92
+ weight: 0.15
93
+ - model: ReadyArt/Forgotten-Safeword-24B-V2.0+trashpanda-org/MS-24B-Mullein-v1-lora
94
+ parameters:
95
+ weight: 0.15
96
+ - model: allura-org/MS3-24B-Roselily-Creative
97
+ parameters:
98
+ weight: 0.15
99
+ - model: TroyDoesAI/BlackSheep-24B
100
+ parameters:
101
+ weight: 0.10
102
+ parameters:
103
+ density: 0.79
104
+ tokenizer:
105
+ source: union
106
+ chat_template: auto
107
+ ---
108
+ name: Sorawiz/MS-Creative-24B-Test-B
109
+ models:
110
+ - model: ReadyArt/Forgotten-Abomination-24B-V2.2
111
+ - model: OddTheGreat/Apparatus_24B
112
+ parameters:
113
+ density: 1.00
114
+ weight: 1.00
115
+ - model: TroyDoesAI/BlackSheep-24B
116
+ parameters:
117
+ density: 1.00
118
+ weight: 1.00
119
+ - model: Darkknight535/WinterEngine-24B-Instruct
120
+ parameters:
121
+ density: 1.00
122
+ weight: 1.00
123
+ - model: allura-org/MS3-24B-Roselily-Creative
124
+ parameters:
125
+ density: 0.70
126
+ weight: 0.50
127
+ - model: Nohobby/MS3-Tantum-24B-v0.1
128
+ parameters:
129
+ density: 0.70
130
+ weight: 0.50
131
+ merge_method: ties
132
+ base_model: ReadyArt/Forgotten-Abomination-24B-V2.2
133
+ parameters:
134
+ normalize: true
135
+ dtype: float32
136
+ ---
137
+ name: Sorawiz/MistralSmall-Creative-24B-Stock
138
+ models:
139
+ - model: Sorawiz/MS-Creative-24B-Test-A
140
+ - model: Sorawiz/MS-Creative-24B-Test-B
141
+ merge_method: model_stock
142
+ base_model: TheDrummer/Cydonia-24B-v2.1
143
+ parameters:
144
+ filter_wise: false
145
+ dtype: float32
146
+ ---
147
  models:
148
  - model: ReadyArt/Forgotten-Safeword-24B-V2.2
149
  - model: OddTheGreat/Apparatus_24B
 
154
  parameters:
155
  density: 1.00
156
  weight: 1.00
157
+ - model: Sorawiz/MistralSmall-Creative-24B-Stock
158
  parameters:
159
  density: 0.50
160
  weight: 0.50