Home-3B-v3-GGUF
| Property | Value |
|---|---|
| Base Model | StableLM-3B-Zephyr |
| License | STABILITY AI NON-COMMERCIAL RESEARCH COMMUNITY LICENSE |
| Author | acon96 |
| Accuracy | 97.11% on JSON function calling |
What is Home-3B-v3-GGUF?
Home-3B-v3-GGUF is a specialized AI model fine-tuned for home automation control and basic question-answering capabilities. Built upon StableLM-3B-Zephyr, this model represents a significant advancement in home device control through natural language processing, featuring enhanced accuracy and multi-language support.
Implementation Details
The model was trained using LoRA with rank=64 and alpha=128, targeting specific projection modules. It's optimized for low-resource environments through GGUF quantization, making it suitable for devices like Raspberry Pis running Home Assistant.
- Custom curated dataset for function calling training
- Zephyr prompt format compatibility
- Quantized for efficient resource usage
- Multi-personality support implementation
Core Capabilities
- Home device control through natural language
- Function calling with 97.11% accuracy
- Multi-language support (German, Spanish, French)
- Basic Q&A and logical reasoning
- Home Assistant service integration
- Support for various entity types (lights, fans, vacuums, timers, todos)
Frequently Asked Questions
Q: What makes this model unique?
This model combines home automation control with natural language processing in a lightweight package, featuring high accuracy in function calling and support for multiple languages, all while being optimized for resource-constrained environments.
Q: What are the recommended use cases?
The model is ideal for home automation systems, particularly Home Assistant installations. It excels at controlling smart home devices through natural language commands and can handle basic Q&A tasks. It's particularly suitable for deployment on low-resource hardware like Raspberry Pis.
