Overview

SakanaAI/EvoLLM-JP-A-v1-7B is a 7 billion parameter experimental general-purpose Japanese large language model developed by Sakana AI. This model is notable for its creation method: Evolutionary Model Merge. This technique combines the strengths of multiple existing models, specifically Shisa Gamma 7B v1, Arithmo2 Mistral 7B, and Abel 7B 002, to produce a new, optimized model. More details on this evolutionary merging process are available in their research paper and blog post.

Key Characteristics

• Model Type: Autoregressive Language Model
• Language Focus: Primarily Japanese
• Development Method: Evolutionary Model Merge, a novel approach to combining pre-trained models.
• License: Apache License, Version 2.0

Intended Use

This model is provided for research and development purposes only and should be considered an experimental prototype. It is not intended for commercial use or deployment in critical environments. Users should be aware of its experimental nature and use it at their own discretion, understanding that performance and outcomes are not guaranteed.