README.md

---
library_name: transformers
tags:
- mergekit
- merge
new_version: Lunzima/NQLSG-Qwen2.5-14B-MegaFusion-v3
---
# About

Merged on a local 512GB memory AMD EPYC 7773X server.

# NQLSG-Qwen2.5-14B-MegaFusion

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 [SCE](https://arxiv.org/abs/2408.07990) merge method using /root/LLM/NQLSG-Qwen2.5-14B-Base as a base.

### Models Merged

The following models were included in the merge:
* /root/LLM/NQLSG-Qwen2.5-14B-Base6
* /root/LLM/NQLSG-Qwen2.5-14B-Base3
* /root/LLM/NQLSG-Qwen2.5-14B-Base4
* /root/LLM/NQLSG-Qwen2.5-14B-Base5
* /root/LLM/NQLSG-Qwen2.5-14B-Base2

### Configuration

The following YAML configuration was used to produce this model:

```yaml
base_model: /root/LLM/NQLSG-Qwen2.5-14B-Base
chat_template: auto
dtype: bfloat16
merge_method: sce
parameters:
  int8_mask: 1.0
slices:
- sources:
  - layer_range: [0, 48]
    model: /root/LLM/NQLSG-Qwen2.5-14B-Base
  - layer_range: [0, 48]
    model: /root/LLM/NQLSG-Qwen2.5-14B-Base2
  - layer_range: [0, 48]
    model: /root/LLM/NQLSG-Qwen2.5-14B-Base3
  - layer_range: [0, 48]
    model: /root/LLM/NQLSG-Qwen2.5-14B-Base4
  - layer_range: [0, 48]
    model: /root/LLM/NQLSG-Qwen2.5-14B-Base5
  - layer_range: [0, 48]
    model: /root/LLM/NQLSG-Qwen2.5-14B-Base6
tokenizer: {}
```