Overview

Locutusque/StockQwen-2.5-7B is a 7.6 billion parameter merged language model built upon the Qwen2.5 architecture. It was created using the Model Stock merge method, combining several specialized Qwen2.5 models into a single, more versatile foundation.

Key Capabilities

This model integrates the capabilities of its constituent models, which include:

• General Instruction Following: Derived from Qwen/Qwen2.5-7B-Instruct.
• Code Generation: Enhanced by Qwen/Qwen2.5-Coder-7B-Instruct.
• Mathematical Reasoning: Benefiting from Qwen/Qwen2.5-Math-7B-Instruct.

Merge Details

The merge process used Qwen/Qwen2.5-7B as the base model and incorporated the instruction-tuned, coder, and math-focused variants. This strategic merge aims to create a robust model with a broad skill set, suitable for a wide range of applications after further fine-tuning. The model supports a substantial context length of 131072 tokens, allowing for processing extensive inputs.

Usage Notes

As a merged model, Locutusque/StockQwen-2.5-7B is intended as a strong base that requires fine-tuning to achieve optimal performance for specific downstream tasks. It supports multiple languages, including Chinese, English, French, Spanish, Portuguese, German, Italian, Russian, Japanese, Korean, Vietnamese, Thai, and Arabic.