Update README.md
Browse files
README.md
CHANGED
@@ -1,7 +1,14 @@
|
|
1 |
---
|
2 |
base_model:
|
3 |
-
-
|
4 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
- TheDrummer/Cydonia-24B-v2.1
|
6 |
library_name: transformers
|
7 |
tags:
|
@@ -9,26 +16,121 @@ tags:
|
|
9 |
- merge
|
10 |
|
11 |
---
|
12 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
|
14 |
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
15 |
|
16 |
## Merge Details
|
17 |
### Merge Method
|
18 |
|
19 |
-
This model was merged using the [Model Stock](https://arxiv.org/abs/2403.19522) merge method
|
20 |
|
21 |
### Models Merged
|
22 |
|
23 |
The following models were included in the merge:
|
24 |
-
* [
|
25 |
-
* [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
26 |
|
27 |
### Configuration
|
28 |
|
29 |
The following YAML configuration was used to produce this model:
|
30 |
|
31 |
```yaml
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
32 |
models:
|
33 |
- model: Sorawiz/MS-Creative-24B-Test-A
|
34 |
- model: Sorawiz/MS-Creative-24B-Test-B
|
|
|
1 |
---
|
2 |
base_model:
|
3 |
+
- ReadyArt/Forgotten-Safeword-24B-V2.2
|
4 |
+
- ReadyArt/Forgotten-Safeword-24B-V2.0
|
5 |
+
- trashpanda-org/MS-24B-Mullein-v1-lora
|
6 |
+
- ReadyArt/Forgotten-Abomination-24B-V2.2
|
7 |
+
- OddTheGreat/Apparatus_24B
|
8 |
+
- Darkknight535/WinterEngine-24B-Instruct
|
9 |
+
- allura-org/MS3-24B-Roselily-Creative
|
10 |
+
- TroyDoesAI/BlackSheep-24B
|
11 |
+
- Nohobby/MS3-Tantum-24B-v0.1
|
12 |
- TheDrummer/Cydonia-24B-v2.1
|
13 |
library_name: transformers
|
14 |
tags:
|
|
|
16 |
- merge
|
17 |
|
18 |
---
|
19 |
+
# Chat Template
|
20 |
+
|
21 |
+
Mistral Instruct
|
22 |
+
|
23 |
+
```
|
24 |
+
{{ if .System }}<|im_start|>system
|
25 |
+
{{ .System }}<|im_end|>
|
26 |
+
{{ end }}{{ if .Prompt }}<|im_start|>user
|
27 |
+
{{ .Prompt }}<|im_end|>
|
28 |
+
{{ end }}<|im_start|>assistant
|
29 |
+
{{ .Response }}<|im_end|>
|
30 |
+
```
|
31 |
+
|
32 |
+
ChatML
|
33 |
+
|
34 |
+
```
|
35 |
+
{{ if .System }}<|im_start|>system
|
36 |
+
{{ .System }}<|im_end|>
|
37 |
+
{{ end }}{{ if .Prompt }}<|im_start|>user
|
38 |
+
{{ .Prompt }}<|im_end|>
|
39 |
+
{{ end }}<|im_start|>assistant
|
40 |
+
{{ .Response }}{{ if .Response }}<|im_end|>{{ end }}
|
41 |
+
```
|
42 |
+
|
43 |
+
# GGUF
|
44 |
+
|
45 |
+
* Q6_K quant - [Sorawiz/MistralSmall-Creative-24B-Q6_K-GGUF](https://huggingface.co/Sorawiz/MistralSmall-Creative-24B-Q6_K-GGUF)
|
46 |
+
|
47 |
+
# MERGE
|
48 |
|
49 |
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
50 |
|
51 |
## Merge Details
|
52 |
### Merge Method
|
53 |
|
54 |
+
This model was merged using the [Model Stock](https://arxiv.org/abs/2403.19522) merge method.
|
55 |
|
56 |
### Models Merged
|
57 |
|
58 |
The following models were included in the merge:
|
59 |
+
* [ReadyArt/Forgotten-Safeword-24B-V2.2](https://huggingface.co/ReadyArt/Forgotten-Safeword-24B-V2.2)
|
60 |
+
* [ReadyArt/Forgotten-Safeword-24B-V2.0](https://huggingface.co/ReadyArt/Forgotten-Safeword-24B-V2.0) + [trashpanda-org/MS-24B-Mullein-v1-lora](https://huggingface.co/trashpanda-org/MS-24B-Mullein-v1-lora)
|
61 |
+
* [ReadyArt/Forgotten-Abomination-24B-V2.2](https://huggingface.co/ReadyArt/Forgotten-Abomination-24B-V2.2)
|
62 |
+
* [OddTheGreat/Apparatus_24B](https://huggingface.co/OddTheGreat/Apparatus_24B)
|
63 |
+
* [Darkknight535/WinterEngine-24B-Instruct](https://huggingface.co/Darkknight535/WinterEngine-24B-Instruct)
|
64 |
+
* [allura-org/MS3-24B-Roselily-Creative](https://huggingface.co/allura-org/MS3-24B-Roselily-Creative)
|
65 |
+
* [TroyDoesAI/BlackSheep-24B](https://huggingface.co/TroyDoesAI/BlackSheep-24B)
|
66 |
+
* [Nohobby/MS3-Tantum-24B-v0.1](https://huggingface.co/Nohobby/MS3-Tantum-24B-v0.1)
|
67 |
+
* [TheDrummer/Cydonia-24B-v2.1](https://huggingface.co/TheDrummer/Cydonia-24B-v2.1)
|
68 |
|
69 |
### Configuration
|
70 |
|
71 |
The following YAML configuration was used to produce this model:
|
72 |
|
73 |
```yaml
|
74 |
+
name: Sorawiz/MS-Creative-24B-Test-A
|
75 |
+
merge_method: dare_ties
|
76 |
+
base_model: ReadyArt/Forgotten-Safeword-24B-V2.2
|
77 |
+
models:
|
78 |
+
- model: ReadyArt/Forgotten-Safeword-24B-V2.2
|
79 |
+
parameters:
|
80 |
+
weight: 0.05
|
81 |
+
- model: ReadyArt/Forgotten-Abomination-24B-V2.2
|
82 |
+
parameters:
|
83 |
+
weight: 0.20
|
84 |
+
- model: OddTheGreat/Apparatus_24B
|
85 |
+
parameters:
|
86 |
+
weight: 0.20
|
87 |
+
- model: Darkknight535/WinterEngine-24B-Instruct
|
88 |
+
parameters:
|
89 |
+
weight: 0.15
|
90 |
+
- model: ReadyArt/Forgotten-Safeword-24B-V2.0+trashpanda-org/MS-24B-Mullein-v1-lora
|
91 |
+
parameters:
|
92 |
+
weight: 0.15
|
93 |
+
- model: allura-org/MS3-24B-Roselily-Creative
|
94 |
+
parameters:
|
95 |
+
weight: 0.15
|
96 |
+
- model: TroyDoesAI/BlackSheep-24B
|
97 |
+
parameters:
|
98 |
+
weight: 0.10
|
99 |
+
parameters:
|
100 |
+
density: 0.79
|
101 |
+
tokenizer:
|
102 |
+
source: union
|
103 |
+
chat_template: auto
|
104 |
+
---
|
105 |
+
name: Sorawiz/MS-Creative-24B-Test-B
|
106 |
+
models:
|
107 |
+
- model: ReadyArt/Forgotten-Abomination-24B-V2.2
|
108 |
+
- model: OddTheGreat/Apparatus_24B
|
109 |
+
parameters:
|
110 |
+
density: 1.00
|
111 |
+
weight: 1.00
|
112 |
+
- model: TroyDoesAI/BlackSheep-24B
|
113 |
+
parameters:
|
114 |
+
density: 1.00
|
115 |
+
weight: 1.00
|
116 |
+
- model: Darkknight535/WinterEngine-24B-Instruct
|
117 |
+
parameters:
|
118 |
+
density: 1.00
|
119 |
+
weight: 1.00
|
120 |
+
- model: allura-org/MS3-24B-Roselily-Creative
|
121 |
+
parameters:
|
122 |
+
density: 0.70
|
123 |
+
weight: 0.50
|
124 |
+
- model: Nohobby/MS3-Tantum-24B-v0.1
|
125 |
+
parameters:
|
126 |
+
density: 0.70
|
127 |
+
weight: 0.50
|
128 |
+
merge_method: ties
|
129 |
+
base_model: ReadyArt/Forgotten-Abomination-24B-V2.2
|
130 |
+
parameters:
|
131 |
+
normalize: true
|
132 |
+
dtype: float32
|
133 |
+
---
|
134 |
models:
|
135 |
- model: Sorawiz/MS-Creative-24B-Test-A
|
136 |
- model: Sorawiz/MS-Creative-24B-Test-B
|