Asherah_7B Overview

Asherah_7B is a 7 billion parameter language model developed by ResplendentAI. Its creation involved merging four distinct models from Sanji Watsuki, followed by fine-tuning on two proprietary datasets, Gnosis and Synthetic Soul, both authored by the developer. This unique training approach aims to imbue the model with specific characteristics derived from its diverse foundational components and custom data.

Key Capabilities

• Standard LLM Functionality: Asherah_7B can be utilized as a general-purpose large language model for various text-based tasks.
• Multimodal Support (GGUF): The model offers multimodal capabilities specifically for users leveraging the GGUF format. This functionality requires an external mmproj file, which is provided within the model repository for convenience.

Usage Notes

• For multimodal use with GGUF, users are directed to use a compatible mmproj file, such as the one available from cjpais/llava-1.6-mistral-7b-gguf.
• The developer recommends using tools like Koboldcpp for optimal multimodal GGUF experience.
• While multimodal features are currently limited to GGUF users, the model remains fully functional as a text-only LLM for all users.