Anemia Model Builder Lab

This repository contains an end-to-end experimentation workflow for anemia detection using palm images, including baseline deep learning training and advanced ensemble-style modeling.

Repository Structure

1. notebooks/core/01_palm_anemia_detection_convnext.ipynb
   Main baseline pipeline using TensorFlow + ConvNeXtTiny on palm image data.

2. notebooks/advanced/02_multimodal_ensemble_anemia.ipynb
   Advanced workflow with multi-model and tabular-fusion style experimentation using FastAI components.

3. notebooks/archive/03_experiments_fastai_timm.ipynb
   Archived trial notebook with intermediate experiments and setup iterations.

Quick Start

1. Open the core notebook to run the primary training and evaluation flow.
2. Continue with the advanced notebook for improved/extended modeling ideas.
3. Use archive only for reference and experiment history.

Notes

• Notebooks were originally developed in Colab-style environments.
• Update dataset and model paths before running in a new environment.

Project Structure

• Minor Project (this repo): model development and experimentation
• Major Project: full application and deployment

Major repo: https://github.com/Rithik-sudo/anemia-detection-fusion