|
--- |
|
license: apache-2.0 |
|
tags: |
|
- merge |
|
- mergekit |
|
- lazymergekit |
|
metrics: |
|
- code_eval |
|
- accuracy |
|
--- |
|
|
|
# Mnemosyne-7B |
|
|
|
Mnemosyne-7B is an experimental large language model (LLM) created by merging several pre-trained models designed for informative and educational purposes. It combines the strengths of these models with the hope of achieving a highly informative and comprehensive LLM. |
|
|
|
GGUF: https://huggingface.co/mradermacher/Mnemosyne-7B-GGUF |
|
|
|
### Important Note: |
|
|
|
This is an experimental model, and its performance and capabilities are not guaranteed. Further testing and evaluation are required to assess its effectiveness. |
|
|
|
|
|
## 🧩 Configuration |
|
|
|
```yaml |
|
models: |
|
- model: MaziyarPanahi/Mistral-7B-Instruct-KhanAcademy-v0.2 |
|
- model: openbmb/Eurus-7b-kto |
|
- model: Weyaxi/Newton-7B |
|
merge_method: model_stock |
|
base_model: mistralai/Mistral-7B-Instruct-v0.2 |
|
dtype: bfloat16 |
|
|
|
``` |
|
|
|
Mnemosyne-7B is a merge of the following models using [mergekit](https://github.com/cg123/mergekit): |