Casual-Autopsy
commited on
Commit
•
0baeb3b
1
Parent(s):
3763388
Update README.md
Browse files
README.md
CHANGED
@@ -1,42 +1,176 @@
|
|
1 |
---
|
2 |
-
base_model:
|
3 |
-
- Casual-Autopsy/L3-Umbral-Mind-RP-v3-8B
|
4 |
-
- Casual-Autopsy/Llama-3-Deppressed-Therapist-8B
|
5 |
-
library_name: transformers
|
6 |
tags:
|
7 |
-
- mergekit
|
8 |
- merge
|
9 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
---
|
|
|
|
|
|
|
|
|
11 |
# merge
|
12 |
|
13 |
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
14 |
|
15 |
## Merge Details
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
16 |
### Merge Method
|
17 |
|
18 |
-
This model was merged using
|
19 |
|
20 |
### Models Merged
|
21 |
|
22 |
The following models were included in the merge:
|
23 |
-
* [Casual-Autopsy/Llama-3-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
|
25 |
-
|
26 |
|
27 |
-
|
28 |
|
29 |
```yaml
|
30 |
slices:
|
31 |
- sources:
|
32 |
-
- model:
|
33 |
layer_range: [0, 32]
|
34 |
-
|
|
|
|
|
35 |
layer_range: [0, 32]
|
36 |
parameters:
|
37 |
-
weight: 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
38 |
merge_method: task_arithmetic
|
39 |
-
base_model:
|
|
|
40 |
dtype: bfloat16
|
|
|
41 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
42 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
|
|
|
|
|
|
|
|
2 |
tags:
|
|
|
3 |
- merge
|
4 |
+
- mergekit
|
5 |
+
- lazymergekit
|
6 |
+
- not-for-all-audiences
|
7 |
+
- nsfw
|
8 |
+
- rp
|
9 |
+
- roleplay
|
10 |
+
- role-play
|
11 |
+
license: llama3
|
12 |
+
language:
|
13 |
+
- en
|
14 |
+
library_name: transformers
|
15 |
+
pipeline_tag: text-generation
|
16 |
+
base_model:
|
17 |
+
- Sao10K/L3-8B-Stheno-v3.2
|
18 |
+
- Hastagaras/Jamet-8B-L3-MK.V-Blackroot
|
19 |
+
- grimjim/Llama-3-Oasis-v1-OAS-8B
|
20 |
+
- Casual-Autopsy/SOVL-MopeyMule-8B
|
21 |
+
- Casual-Autopsy/MopeyMule-Blackroot-8B
|
22 |
+
- ResplendentAI/Theory_of_Mind_Llama3
|
23 |
+
- ResplendentAI/RP_Format_QuoteAsterisk_Llama3
|
24 |
+
- ResplendentAI/Smarts_Llama3
|
25 |
+
- Casual-Autopsy/L3-Umbral-Mind-RP-v3-8B
|
26 |
+
- Casual-Autopsy/Llama-3-Depressed-Therapist-8B
|
27 |
---
|
28 |
+
|
29 |
+
<img src="https://huggingface.co/Casual-Autopsy/L3-Umbral-Mind-RP-v3-8B/resolve/main/63073798_p0_master1200.jpg" style="display: block; margin: auto;">
|
30 |
+
Image by ろ47
|
31 |
+
|
32 |
# merge
|
33 |
|
34 |
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
35 |
|
36 |
## Merge Details
|
37 |
+
|
38 |
+
The goal of this merge was to make an RP model better suited for role-plays with heavy themes such as but not limited to:
|
39 |
+
- Mental illness
|
40 |
+
- Self-harm
|
41 |
+
- Trauma
|
42 |
+
- Suicide
|
43 |
+
|
44 |
+
I hated how RP models tended to be overly positive and hopeful with role-plays involving such themes,
|
45 |
+
but thanks to [failspy/Llama-3-8B-Instruct-MopeyMule](https://huggingface.co/failspy/Llama-3-8B-Instruct-MopeyMule) this problem has been lessened considerably.
|
46 |
+
|
47 |
+
If you're an enjoyer of savior/reverse savior type role-plays like myself, then this bot is for you.
|
48 |
+
|
49 |
+
### Quants
|
50 |
+
|
51 |
+
|
52 |
+
|
53 |
### Merge Method
|
54 |
|
55 |
+
This model was merged using several Task Arithmetic merges and then tied together with a Model Stock merge, followed by another Task Arithmetic merge with a model containing psychology data.
|
56 |
|
57 |
### Models Merged
|
58 |
|
59 |
The following models were included in the merge:
|
60 |
+
* [Casual-Autopsy/Llama-3-Depressed-Therapist-8B](Casual-Autopsy/Llama-3-Depressed-Therapist-8B)
|
61 |
+
* [Casual-Autopsy/L3-Umbral-Mind-RP-v0.3-8B](https://huggingface.co/Casual-Autopsy/L3-Umbral-Mind-RP-v0.3-8B)
|
62 |
+
* Casual-Autopsy/Umbral-v3-1 + [ResplendentAI/Theory_of_Mind_Llama3](https://huggingface.co/ResplendentAI/Theory_of_Mind_Llama3)
|
63 |
+
* [Sao10K/L3-8B-Stheno-v3.2](https://huggingface.co/Sao10K/L3-8B-Stheno-v3.2)
|
64 |
+
* [Casual-Autopsy/SOVL-MopeyMule-8B](https://huggingface.co/Casual-Autopsy/SOVL-MopeyMule-8B)
|
65 |
+
* [Casual-Autopsy/MopeyMule-Blackroot-8B](https://huggingface.co/Casual-Autopsy/MopeyMule-Blackroot-8B)
|
66 |
+
|
67 |
+
* Casual-Autopsy/Umbral-v3-2 + [ResplendentAI/Smarts_Llama3](https://huggingface.co/ResplendentAI/Smarts_Llama3)
|
68 |
+
* [Hastagaras/Jamet-8B-L3-MK.V-Blackroot](https://huggingface.co/Hastagaras/Jamet-8B-L3-MK.V-Blackroot)
|
69 |
+
* [Casual-Autopsy/SOVL-MopeyMule-8B](https://huggingface.co/Casual-Autopsy/SOVL-MopeyMule-8B)
|
70 |
+
* [Casual-Autopsy/MopeyMule-Blackroot-8B](https://huggingface.co/Casual-Autopsy/MopeyMule-Blackroot-8B)
|
71 |
+
|
72 |
+
* Casual-Autopsy/Umbral-v3-3 + [ResplendentAI/RP_Format_QuoteAsterisk_Llama3](https://huggingface.co/ResplendentAI/RP_Format_QuoteAsterisk_Llama3)
|
73 |
+
* [grimjim/Llama-3-Oasis-v1-OAS-8B](https://huggingface.co/grimjim/Llama-3-Oasis-v1-OAS-8B)
|
74 |
+
* [Casual-Autopsy/SOVL-MopeyMule-8B](https://huggingface.co/Casual-Autopsy/SOVL-MopeyMule-8B)
|
75 |
+
* [Casual-Autopsy/MopeyMule-Blackroot-8B](https://huggingface.co/Casual-Autopsy/MopeyMule-Blackroot-8B)
|
76 |
+
|
77 |
+
## Secret Sauce
|
78 |
|
79 |
+
The following YAML configurations were used to produce this model:
|
80 |
|
81 |
+
### Umbral-v3-1
|
82 |
|
83 |
```yaml
|
84 |
slices:
|
85 |
- sources:
|
86 |
+
- model: Sao10K/L3-8B-Stheno-v3.2
|
87 |
layer_range: [0, 32]
|
88 |
+
parameters:
|
89 |
+
weight: 0.65
|
90 |
+
- model: Casual-Autopsy/SOVL-MopeyMule-8B
|
91 |
layer_range: [0, 32]
|
92 |
parameters:
|
93 |
+
weight: 0.25
|
94 |
+
- model: Casual-Autopsy/MopeyMule-Blackroot-8B
|
95 |
+
layer_range: [0, 32]
|
96 |
+
parameters:
|
97 |
+
weight: 0.1
|
98 |
+
merge_method: task_arithmetic
|
99 |
+
base_model: Sao10K/L3-8B-Stheno-v3.2
|
100 |
+
normalize: False
|
101 |
+
dtype: bfloat16
|
102 |
+
```
|
103 |
+
|
104 |
+
### Umbral-v3-2
|
105 |
+
|
106 |
+
```yaml
|
107 |
+
slices:
|
108 |
+
- sources:
|
109 |
+
- model: Hastagaras/Jamet-8B-L3-MK.V-Blackroot
|
110 |
+
layer_range: [0, 32]
|
111 |
+
parameters:
|
112 |
+
weight: 0.75
|
113 |
+
- model: Casual-Autopsy/SOVL-MopeyMule-8B
|
114 |
+
layer_range: [0, 32]
|
115 |
+
parameters:
|
116 |
+
weight: 0.15
|
117 |
+
- model: Casual-Autopsy/MopeyMule-Blackroot-8B
|
118 |
+
layer_range: [0, 32]
|
119 |
+
parameters:
|
120 |
+
weight: 0.1
|
121 |
merge_method: task_arithmetic
|
122 |
+
base_model: Hastagaras/Jamet-8B-L3-MK.V-Blackroot
|
123 |
+
normalize: False
|
124 |
dtype: bfloat16
|
125 |
+
```
|
126 |
|
127 |
+
### Umbral-v3-3
|
128 |
+
|
129 |
+
```yaml
|
130 |
+
slices:
|
131 |
+
- sources:
|
132 |
+
- model: grimjim/Llama-3-Oasis-v1-OAS-8B
|
133 |
+
layer_range: [0, 32]
|
134 |
+
parameters:
|
135 |
+
weight: 0.55
|
136 |
+
- model: Casual-Autopsy/SOVL-MopeyMule-8B
|
137 |
+
layer_range: [0, 32]
|
138 |
+
parameters:
|
139 |
+
weight: 0.35
|
140 |
+
- model: Casual-Autopsy/MopeyMule-Blackroot-8B
|
141 |
+
layer_range: [0, 32]
|
142 |
+
parameters:
|
143 |
+
weight: 0.1
|
144 |
+
merge_method: task_arithmetic
|
145 |
+
base_model: grimjim/Llama-3-Oasis-v1-OAS-8B
|
146 |
+
normalize: False
|
147 |
+
dtype: bfloat16
|
148 |
```
|
149 |
+
|
150 |
+
### L3-Umbral-Mind-RP-v0.3-8B
|
151 |
+
|
152 |
+
```yaml
|
153 |
+
models:
|
154 |
+
- model: Casual-Autopsy/Umbral-v3-1+ResplendentAI/Theory_of_Mind_Llama3
|
155 |
+
- model: Casual-Autopsy/Umbral-v3-2+ResplendentAI/Smarts_Llama3
|
156 |
+
- model: Casual-Autopsy/Umbral-v3-3+ResplendentAI/RP_Format_QuoteAsterisk_Llama3
|
157 |
+
merge_method: model_stock
|
158 |
+
base_model: Casual-Autopsy/Umbral-v3-1
|
159 |
+
dtype: bfloat16
|
160 |
+
```
|
161 |
+
|
162 |
+
### L3-Umbral-Mind-RP-v0.6.2-8B
|
163 |
+
|
164 |
+
```yaml
|
165 |
+
slices:
|
166 |
+
- sources:
|
167 |
+
- model: Casual-Autopsy/L3-Umbral-Mind-RP-v0.3-8B
|
168 |
+
layer_range: [0, 32]
|
169 |
+
- model: Casual-Autopsy/Llama-3-Depressed-Therapist-8B
|
170 |
+
layer_range: [0, 32]
|
171 |
+
parameters:
|
172 |
+
weight: 0.2
|
173 |
+
merge_method: task_arithmetic
|
174 |
+
base_model: Casual-Autopsy/L3-Umbral-Mind-RP-v0.3-8B
|
175 |
+
dtype: bfloat16
|
176 |
+
```
|