---
license: apache-2.0
base_model:
  - Qwen/Qwen2.5-Coder-7B-Instruct
  - lightblue/Karasu-DPO-7B
tags:
  - merge
  - mergekit
  - dare_ties
  - japanese
  - coding
---

# DARE-TIES Merged Model (Ratio: 0.3)

This is a merged model created using the DARE_TIES method with mergekit.

## Base Models
- **Qwen/Qwen2.5-Coder-7B-Instruct** (Weight: 0.7)
- **lightblue/Karasu-DPO-7B** (Weight: 0.3)

## Merge Method
- **Method**: DARE_TIES
- **Density**: 0.5
- **Data Type**: bfloat16

## Purpose
This model aims to enhance Japanese code generation capabilities while maintaining English coding performance.

## Usage
```python
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("noirchan/DARE-TIES-Qwen2.5-Coder-Karasu-0.3")
model = AutoModelForCausalLM.from_pretrained("noirchan/DARE-TIES-Qwen2.5-Coder-Karasu-0.3")
```

## Evaluation
This model is part of a systematic evaluation of different merge ratios to find the optimal balance between Japanese language capabilities and code generation performance.