Update README.md
Browse files
README.md
CHANGED
@@ -6,20 +6,21 @@ tags:
|
|
6 |
- merge
|
7 |
|
8 |
---
|
9 |
-
#
|
10 |
|
11 |
-
This
|
12 |
-
|
13 |
-
## Merge Details
|
14 |
-
### Merge Method
|
15 |
-
|
16 |
-
This model was merged using the SLERP merge method.
|
17 |
|
18 |
### Models Merged
|
19 |
|
20 |
The following models were included in the merge:
|
21 |
-
|
22 |
-
|
|
|
|
|
|
|
|
|
|
|
23 |
|
24 |
### Configuration
|
25 |
|
@@ -28,12 +29,12 @@ The following YAML configuration was used to produce this model:
|
|
28 |
```yaml
|
29 |
slices:
|
30 |
- sources:
|
31 |
-
- model:
|
32 |
layer_range: [0, 80]
|
33 |
-
- model:
|
34 |
layer_range: [0, 80]
|
35 |
merge_method: slerp
|
36 |
-
base_model:
|
37 |
parameters:
|
38 |
t:
|
39 |
- filter: self_attn
|
|
|
6 |
- merge
|
7 |
|
8 |
---
|
9 |
+
# OrcaHermes-Mistral-70B
|
10 |
|
11 |
+
This model was created by SLERP Merging 2 Miqu Models trained on 2 high preforming datsets
|
12 |
+
Just an experiment have not seen much miqu slerps yet.
|
|
|
|
|
|
|
|
|
13 |
|
14 |
### Models Merged
|
15 |
|
16 |
The following models were included in the merge:
|
17 |
+
|
18 |
+
[Miqu-Openhermes](https://huggingface.co/alicecomfy/miqu-openhermes-full)
|
19 |
+
- Base Miqu Trained on [Openhermes](https://huggingface.co/datasets/teknium/OpenHermes-2.5)
|
20 |
+
|
21 |
+
[ShinojiResearch/Senku-70B-Full](https://huggingface.co/ShinojiResearch/Senku-70B-Full)
|
22 |
+
- Base Miqu Trained on [Slimorca](https://huggingface.co/datasets/Open-Orca/SlimOrca)
|
23 |
+
|
24 |
|
25 |
### Configuration
|
26 |
|
|
|
29 |
```yaml
|
30 |
slices:
|
31 |
- sources:
|
32 |
+
- model: local//path//to//Senku-70B-Full
|
33 |
layer_range: [0, 80]
|
34 |
+
- model: local//path//to//miqu-openhermes-full
|
35 |
layer_range: [0, 80]
|
36 |
merge_method: slerp
|
37 |
+
base_model: local//path//to//Senku-70B-Full
|
38 |
parameters:
|
39 |
t:
|
40 |
- filter: self_attn
|