OddTheGreat
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -1,50 +1,23 @@
|
|
1 |
-
---
|
2 |
-
base_model:
|
3 |
-
- OddTheGreat/Malevolent_12B.v2
|
4 |
-
- OddTheGreat/Pretender-12b
|
5 |
-
library_name: transformers
|
6 |
-
tags:
|
7 |
-
- mergekit
|
8 |
-
- merge
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
This is a merge of pre-trained language models
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
* [OddTheGreat/Pretender-12b](https://huggingface.co/OddTheGreat/Pretender-12b)
|
25 |
-
|
26 |
-
### Configuration
|
27 |
-
|
28 |
-
The following YAML configuration was used to produce this model:
|
29 |
-
|
30 |
-
```yaml
|
31 |
-
slices:
|
32 |
-
- sources:
|
33 |
-
- model: OddTheGreat/Malevolent_12B.v2
|
34 |
-
layer_range: [0, 40]
|
35 |
-
- model: OddTheGreat/Pretender-12b
|
36 |
-
layer_range: [0, 40]
|
37 |
-
merge_method: slerp
|
38 |
-
base_model: OddTheGreat/Malevolent_12B.v2
|
39 |
-
parameters:
|
40 |
-
t:
|
41 |
-
- filter: self_attn
|
42 |
-
value: [0.65, 0.7, 0.6, 0.5, 0.75]
|
43 |
-
- filter: mlp
|
44 |
-
value: [0.35, 0.3, 0.4, 0.5, 0.25]
|
45 |
-
- value: 0.5
|
46 |
-
dtype: float16
|
47 |
-
chat_template: "chatml"
|
48 |
-
tokenizer:
|
49 |
-
source: union
|
50 |
-
```
|
|
|
1 |
+
---
|
2 |
+
base_model:
|
3 |
+
- OddTheGreat/Malevolent_12B.v2
|
4 |
+
- OddTheGreat/Pretender-12b
|
5 |
+
library_name: transformers
|
6 |
+
tags:
|
7 |
+
- mergekit
|
8 |
+
- merge
|
9 |
+
language:
|
10 |
+
- ru
|
11 |
+
- en
|
12 |
+
---
|
13 |
+
This is a merge of pre-trained language models.
|
14 |
+
|
15 |
+
Due to mergekit space malfunctions, this merge was created locally.
|
16 |
+
|
17 |
+
Main goal of this merge is to combine interesting output of "Pretender" and dark creativity of "Malevolent"
|
18 |
+
|
19 |
+
Use with ChatML
|
20 |
+
|
21 |
+
Tested on T1.01
|
22 |
+
|
23 |
+
Perfect both on russian and english.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|