Basic Visual Synthesis

Basic Visual Synthesis is a project that performs image analysis and generates detailed commentary by integrating YOLO-based object detection with Anything LLM (Deepseek R114B) using a RAG (Retrieval Augmented Generation) approach. The pipeline detects objects in an image, retrieves additional context from a built-in knowledge base, and then leverages the LLM to produce comprehensive commentary on the image.

Project Description: "This project aims to perform image analysis and generate detailed commentary by integrating YOLO-based object detection with the Deepseek LLM using a RAG (Retrieval Augmented Generation) approach."

Table of Contents

• Features
• Installation & Requirements

Features

• Object Detection: Utilizes Ultralytics YOLO for detecting objects in images.
• Context Retrieval: Provides additional context via a built-in knowledge base.
• RAG Approach: Integrates Anything LLM for detailed commentary generation. with Deepseek R114B to generate detailed commentary using a Retrieval Augmented Generation method.
• Colored Logging: Uses Colorama with Python's logging module for enhanced, colored log output.

Installation & Requirements

1. Python Version: Tested with Python 3.8 and above.

2. Clone the Repository:

   git clone https://github.com/oaslananka/BasicVisualSynthesis.git
   cd BasicVisualSynthesis

3. Create a Virtual Environment (Optional but Recommended):

   python -m venv .venv
   source .venv/bin/activate
   pip install ultralytics colorama requests
   python VisualSynthesisRAG.py

4. Install Required Packages:

   pip install ultralytics colorama requests

5. Configure API Keys and Model Paths:

   • Update the API_KEY and BASE_URL constants in the code as needed.