|
--- |
|
license: llama3 |
|
tags: |
|
- merge |
|
- mergekit |
|
- lazymergekit |
|
--- |
|
|
|
# L3-NotCrazy-8B |
|
|
|
L3-NotCrazy-8B is a language model designed to combine the strengths of three existing models: a better reasoning model, a roleplay creative model, and a super precise instruction following model. This combination aims to provide a more versatile and capable language model for various tasks, including: |
|
|
|
* Complex reasoning and problem-solving: Leveraging the better reasoning model's capabilities. |
|
* Creative and engaging role-playing: Utilizing the roleplay creative model's strengths. |
|
* Precise and accurate instruction following: Benefiting from the super precise instruction following model's abilities. |
|
|
|
L3-NotCrazy-8B is a merge of the following models using [mergekit](https://github.com/cg123/mergekit): |
|
|
|
## 🧩 Configuration |
|
|
|
```yaml |
|
models: |
|
- model: Casual-Autopsy/L3-Umbral-Mind-RP-v2.0-8B |
|
- model: turboderp/llama3-turbcat-instruct-8b |
|
- model: refuelai/Llama-3-Refueled |
|
merge_method: model_stock |
|
base_model: turboderp/llama3-turbcat-instruct-8b |
|
dtype: bfloat16 |
|
|
|
|
|
``` |
