| Property | Value |
|---|---|
| Model Size | 8B parameters |
| License | cc-by-nc-4.0 |
| Authors | NeverSleep, Undi, IkariDev |
| Architecture | Llama-3 |
What is Llama-3-Lumimaid-8B-v0.1?
Llama-3-Lumimaid-8B-v0.1 is a specialized language model built on the Llama-3 architecture, designed to provide balanced conversational capabilities with a 40/60 split between general and role-playing interactions. The model incorporates multiple training datasets including Aesir, NoRobots, and the Luminae dataset, creating a versatile conversational AI system.
Implementation Details
The model utilizes the Llama3 prompting format and is trained on a carefully curated mix of datasets. It's implemented in FP16 format and builds upon several base models including the initial LumiMaid 8B Finetune, Undi95/Llama-3-Unholy-8B-e4, and Undi95/Llama-3-LewdPlay-8B.
- Optimized for 8B parameter architecture
- Implements Llama3 prompting format
- Trained on diverse datasets including Aesir, NoRobots, and LimaRP
- Features balanced content ratio for versatile applications
Core Capabilities
- Advanced conversational abilities
- Role-playing optimization
- Context understanding up to 8k tokens
- Balanced general and specialized interactions
Frequently Asked Questions
Q: What makes this model unique?
This model's uniqueness lies in its balanced approach to content generation, utilizing a specific 40/60 ratio for general versus specialized interactions, along with the integration of the new Luminae dataset from Ikari.
Q: What are the recommended use cases?
The model is designed for conversational AI applications with a focus on role-playing scenarios, while maintaining capability for general interactions. It's particularly suited for applications requiring both creative and structured responses.