README.md

---
title: Tabular App Space
emoji: 🏆
colorFrom: pink
colorTo: red
sdk: gradio
sdk_version: 5.47.0
app_file: app.py
pinned: false
license: mit
short_description: A simple user-facing interface for the tabular model
---

# Flower Color Predictor using AutoGluon 

This repository contains a trained `TabularPredictor` from the AutoGluon library, which was trained to classify flower colors based on their physical dimensions.

## Dataset

The model was trained on the `scottymcgee/flowers` dataset, using the synthetic (`augmented`) split for training and the original (`original`) split for final evaluation.

## Model

The model was borrowed from `https://huggingface.co/its-zion-18/flowers-tabular-autolguon-predictor` model.

## Evaluation Results

The final performance of the best model on the original dataset is as follows:

- **Accuracy**: `1.0000`
- **Weighted F1**: `1.0000`

## Potential Errors

Based on the accuracy being so high, I assume there may be data leakage.
Since the augmented data was created directly from the original data (by adding noise or small variations), 
the model wasn't learning to generalize to new information. It was simply memorizing the patterns it had already been shown. 
This could have led to overfitting, where a model learns the training data so well that it fails to perform on new, unseen data.

## Contact
Feel free to contact me for any questions or concerns: aslann@cmu.edu