tia-mk2 / README.md
Dogge's picture
Upload folder using huggingface_hub
bfb5ae4 verified
---
base_model:
- Sao10K/L3-70B-Euryale-v2.1
- Bllossom/llama-3-Korean-Bllossom-70B
library_name: transformers
tags:
- mergekit
- merge
---
# merge
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
## Merge Details
### Merge Method
This model was merged using the SLERP merge method.
### Models Merged
The following models were included in the merge:
* [Sao10K/L3-70B-Euryale-v2.1](https://huggingface.co/Sao10K/L3-70B-Euryale-v2.1)
* [Bllossom/llama-3-Korean-Bllossom-70B](https://huggingface.co/Bllossom/llama-3-Korean-Bllossom-70B)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
slices:
- sources:
- model: Sao10K/L3-70B-Euryale-v2.1
layer_range: [0, 70] # 전체 레이어
- model: Bllossom/llama-3-Korean-Bllossom-70B
layer_range: [0, 70] # 전체 레이어
merge_method: slerp
base_model: Sao10K/L3-70B-Euryale-v2.1 # 영문 모델을 기반으로
parameters:
t:
- filter: self_attn
value: [0.5, 0.3, 0.7, 0.9, 1] # 가중치 조정
- filter: mlp
value: [1, 0.5, 0.6, 0.4, 0.2]
- value: 0.5 # 기본값
dtype: bfloat16
```