netdur/gemma-3-4b-radiology
netdur/gemma-3-4b-radiology is a 4.3 billion parameter Gemma 3 model fine-tuned by netdur for generating radiology reports from ultrasound images. This model leverages the Gemma architecture and was trained using the Unsloth library on the Radiology_mini dataset. It is specifically optimized for interpreting medical imaging and producing descriptive textual reports, making it suitable for specialized medical AI applications.
Loading preview...
Model Overview
netdur/gemma-3-4b-radiology is a specialized 4.3 billion parameter Gemma 3 model, fine-tuned to generate detailed radiology reports from ultrasound images. Developed by netdur, this model utilizes the Unsloth library for efficient fine-tuning and is provided in GGUF format, compatible with llama.cpp for local inference.
Key Capabilities
- Radiology Report Generation: Specifically trained to interpret ultrasound images and produce structured textual reports.
- Medical Imaging Analysis: Designed for tasks requiring descriptive analysis of medical visuals.
- Gemma Architecture: Built upon the Gemma 3 4B base model, leveraging its foundational language understanding.
- Optimized for Inference: Provided in GGUF format, enabling efficient execution with
llama.cppand GPU acceleration (e.g., Metal on macOS).
Good For
- Automated Radiology Reporting: Assisting radiologists by generating initial drafts or summaries of ultrasound findings.
- Medical AI Development: Serving as a base model for further research and development in medical image-to-text tasks.
- Educational Tools: Demonstrating AI capabilities in medical diagnostics and report generation.
This model was fine-tuned using the unsloth/Radiology_mini dataset, focusing on accurate and descriptive output for medical professionals.