We have released AstroSage-LLaMA-3.1-8B (de Haan et al. 2024), our flagship 8B parameter model for astronomy, now available in GGUF format for efficient local deployment.

The model is available on Hugging Face in two quantized versions:

• AstroSage-8B-BF16.gguf: BFloat16 precision for maximum accuracy
• AstroSage-8B-Q8_0.gguf: 8-bit quantized for more efficient deployment

Key Features:

• CPU-friendly inference on standard laptops
• Compatible with llama.cpp ecosystem
• Maintains core capabilities of the original model while optimizing for local deployment

For detailed installation instructions, usage examples, and comprehensive documentation, please visit our Hugging Face model page.