35 lines
690 B
Markdown
35 lines
690 B
Markdown
|
|
|
||
|
|
---
|
||
|
|
license: apache-2.0
|
||
|
|
---
|
||
|
|
# CarbonVillain-v4-Sakura-Solar-Slerp
|
||
|
|
|
||
|
|
This is the model for CarbonVillain-v4-Sakura-Solar-Slerp. I used [mergekit](https://github.com/cg123/mergekit) to merge models.
|
||
|
|
|
||
|
|
# Yaml Config
|
||
|
|
|
||
|
|
```yaml
|
||
|
|
|
||
|
|
slices:
|
||
|
|
- sources:
|
||
|
|
- model: jeonsworld/CarbonVillain-en-10.7B-v4
|
||
|
|
layer_range: [0, 48]
|
||
|
|
- model: kyujinpy/Sakura-SOLAR-Instruct
|
||
|
|
layer_range: [0, 48]
|
||
|
|
|
||
|
|
merge_method: slerp
|
||
|
|
base_model: upstage/SOLAR-10.7B-Instruct-v1.0
|
||
|
|
|
||
|
|
parameters:
|
||
|
|
t:
|
||
|
|
- filter: self_attn
|
||
|
|
value: [0, 0.5, 0.3, 0.7, 1]
|
||
|
|
- filter: mlp
|
||
|
|
value: [1, 0.5, 0.7, 0.3, 0]
|
||
|
|
- value: 0.5 # fallback for rest of tensors
|
||
|
|
tokenizer_source: union
|
||
|
|
|
||
|
|
dtype: bfloat16
|
||
|
|
|
||
|
|
```
|