KAkston/Catholic-Phi3-Mini

TEXT GENERATIONConcurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:4kPublished:May 17, 2025License:cc0-1.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

KAkston/Catholic-Phi3-Mini is a 4 billion parameter language model, fully fine-tuned by Hershel Kysar of Gray's Creek Media, based on Microsoft's Phi-3 Mini architecture. This model is specifically retrained on public domain Catholic datasets to provide concise, factual answers on Catholic teachings, Bible verses, and prayers. It is optimized as an educational tool for Catholic contexts, offering specialized knowledge in this domain.

Loading preview...

Catholic-Phi3-Mini: Specialized for Catholic Education

Catholic-Phi3-Mini is a 4 billion parameter language model, developed by Hershel Kysar of Gray's Creek Media, through the full fine-tuning of Microsoft's Phi-3 Mini. This model is uniquely specialized to address questions pertaining to Catholic teachings, Bible verses, prayers, and general Catholic knowledge.

Key Capabilities & Features

  • Specialized Knowledge: Provides factual and concise responses on Catholic doctrine, drawing from public domain texts like the Douay-Rheims Bible and pre-1923 Catechism.
  • Educational Focus: Designed as an educational tool for Catholic students, educators, and families.
  • Concise Output: Optimized for short, factual answers, typically 1-2 sentences, as enforced by its fine-tuning and recommended prompt structure.
  • Open License: Released under the CC0 1.0 Universal Public Domain Dedication, allowing for unrestricted use, modification, and distribution, including commercial applications.

Differentiators & Use Cases

Unlike general-purpose LLMs, Catholic-Phi3-Mini is explicitly tailored to reflect Catholic perspectives, making it ideal for specific religious educational applications. It is built for environments where accurate, contextually appropriate information within a Catholic framework is paramount. The model requires a GPU with CUDA support and at least 8 GB VRAM for efficient inference. Users should note its optimization for short, factual answers and its potential limitations with complex reasoning or non-Catholic viewpoints.