---
license: mit
datasets:
- aryan27/geometry-cot
language:
- en
base_model:
- TinyLlama/TinyLlama-1.1B-Chat-v1.0
tags:
- geometry
---


# MiniLLM Geometry Engine

MiniLLM Geometry Engine is a specialized, fine-tuned version of the TinyLlama 1.1B language model, optimized for generating geometry-related functions for a geometry engine. By fine-tuning TinyLlama 1.1B on a custom Geometry Chain-of-Thought (CoT) dataset, MiniLLM Geometry Engine excels at producing accurate and efficient mathematical functions tailored for geometric computations.

## Key Features
- **Base Model**: TinyLlama 1.1B, a lightweight and efficient language model.
- **Fine-Tuning**: Trained on a Geometry CoT dataset, enabling step-by-step reasoning for geometry problems.
- **Output**: Generates precise geometry functions (e.g., calculating areas, volumes, distances, or intersections) in a format compatible with geometry engines.
- **Applications**: Ideal for educational tools, CAD software, game development, and computational geometry tasks.
- **Efficiency**: Optimized for low-resource environments, balancing performance and computational cost.

## Capabilities
MiniLLM Geometry Engine can interpret geometry-related queries and output functional code or mathematical expressions. For example, given a prompt like "Generate a function to calculate the area of a triangle," it produces executable code or formulas with clear reasoning, leveraging its CoT fine-tuning to ensure logical accuracy.

## Technical Details
- **Dataset**: Custom Geometry CoT dataset, including problems on 2D/3D shapes, trigonometry, and coordinate geometry, with step-by-step solutions.
- **Architecture**: Inherits TinyLlama's transformer-based structure, fine-tuned to prioritize geometric reasoning.
- **Output Format**: Produces functions in Python or pseudocode, compatible with geometry engine APIs.
- **Performance**: Enhanced accuracy on geometry tasks compared to the base TinyLlama model, with minimal latency.

MiniLLM Geometry Engine is a powerful, compact solution for developers and researchers needing reliable geometry function generation in resource-constrained environments.