| Property | Value |
|---|---|
| License | Non-commercial |
| Framework | PyTorch |
| Base Model | LLaMA 30B |
What is Alpacino30b?
Alpacino30b (Alpaca Integrated Narrative Optimization) is an advanced language model that combines three powerful components: Alpaca, Chain-of-Thought (CoT), and Storytelling capabilities. Built on the LLaMA 30B architecture, this model represents a sophisticated merge that enhances Alpaca's base capabilities while maintaining its reliable instruction format.
Implementation Details
The model implements a triple-merge architecture, utilizing LoRAs from multiple sources including ChanSung's Alpaca, magicgh's CoT, and GamerUntouch's Storytelling components. It's available in GGML format, making it accessible to a broader community of developers.
- Built on LLaMA 30B architecture
- Incorporates multiple specialized LoRAs
- Optimized for text generation and inference
- Compatible with Text-Generation-WebUI and KoboldAI
Core Capabilities
- Enhanced reasoning abilities through Chain-of-Thought integration
- Superior narrative generation and storytelling
- Text-based adventure game capabilities
- Verbose and detailed description generation
- Context-aware responses with up to 2048 tokens
Frequently Asked Questions
Q: What makes this model unique?
Alpacino30b's uniqueness lies in its specialized triple-merge architecture that combines reasoning, storytelling, and instruction-following capabilities while maintaining Alpaca's proven instruction format. This makes it particularly effective for creative writing and interactive narrative applications.
Q: What are the recommended use cases?
The model excels in text-based adventure games, creative writing, and interactive storytelling. Recommended settings include using the "Storywriter" preset with temperature at 1.1 or "Godlike" preset with 2048 context tokens and 680+ generation tokens.