add README
Browse files- .README.md.swp +0 -0
- .mergelist.yaml.swp +0 -0
- README.md +62 -0
.README.md.swp
ADDED
Binary file (12.3 kB). View file
|
|
.mergelist.yaml.swp
ADDED
Binary file (12.3 kB). View file
|
|
README.md
ADDED
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
pipeline_tag: text-generation
|
6 |
+
inference: false
|
7 |
+
tags:
|
8 |
+
- dare
|
9 |
+
- super mario merge
|
10 |
+
- pytorch
|
11 |
+
- solar
|
12 |
+
- merge
|
13 |
+
---
|
14 |
+
|
15 |
+
# solar megamerge 10.7b
|
16 |
+
|
17 |
+
The following models were merged with DARE using [https://github.com/martyn/safetensors-merge-supermario](https://github.com/martyn/safetensors-merge-supermario)
|
18 |
+
|
19 |
+
## Mergelist
|
20 |
+
|
21 |
+
```
|
22 |
+
models:
|
23 |
+
- model: upstage/SOLAR-10.7B-v1.0
|
24 |
+
- model: upstage/SOLAR-10.7B-Instruct-v1.0
|
25 |
+
parameters:
|
26 |
+
weight: 0.20
|
27 |
+
density: 0.8
|
28 |
+
- model: kyujinpy/SOLAR-Platypus-10.7B-v1
|
29 |
+
parameters:
|
30 |
+
weight: 0.19
|
31 |
+
density: 0.75
|
32 |
+
- model: We-Want-GPU/SOLAR-10.7B-orca-alpaca-gpt4-math
|
33 |
+
parameters:
|
34 |
+
weight: 0.18
|
35 |
+
density: 0.75
|
36 |
+
- model: maywell/Synatra-10.7B-v0.4
|
37 |
+
parameters:
|
38 |
+
weight: 0.18
|
39 |
+
density: 0.7
|
40 |
+
- model: kyujinpy/SOLAR-Platypus-10.7B-v2
|
41 |
+
parameters:
|
42 |
+
weight: 0.17
|
43 |
+
density: 0.7
|
44 |
+
- model: Sao10K/Frostwind-10.7B-v1
|
45 |
+
parameters:
|
46 |
+
weight: 0.16
|
47 |
+
density: 0.65
|
48 |
+
- model: rishiraj/meow
|
49 |
+
parameters:
|
50 |
+
weight: 0.15
|
51 |
+
density: 0.6
|
52 |
+
```
|
53 |
+
|
54 |
+
## Merge command
|
55 |
+
|
56 |
+
```
|
57 |
+
python3 hf_merge.py mergelist.yaml solar-1
|
58 |
+
```
|
59 |
+
|
60 |
+
### Notes
|
61 |
+
|
62 |
+
* in the yaml: `p=weight` and `lambda=1/density`
|