---
title: "Discussion on \"Building a Fine-Tuned Nutri-Score Classifier.\" | Hashnode"
description: "Discussion on \"Building a Fine-Tuned Nutri-Score Classifier.\". A complete and honest account of the data audit, including all the problems and imperfections.\nGithub Link: https://github.com/vishsec/nutri-score-classifier\nBefore writing a single line of model trai"
url: "https://hashnode.com/posts/building-a-fine-tuned-nutri-score-classifier/6a69be668184976f1c4345fc"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Building a Fine-Tuned Nutri-Score Classifier.](https://vishsec.hashnode.dev/building-a-fine-tuned-nutri-score-classifier?utm_source=hashnode&utm_medium=feed)

A complete and honest account of the data audit, including all the problems and imperfections. Github Link: https://github.com/vishsec/nutri-score-classifier Before writing a single line of model trai

Jul 29 · 29 min read · vishsec.hashnode.dev  
Updated Aug 2  
Read the article

-   [#ai](https://hashnode.com/tag/ai)
-   [#food](https://hashnode.com/tag/food)
-   [#fine-tuning-models](https://hashnode.com/tag/fine-tuning-models)

## Written by

[Vishwajit Avarsekar](https://hashnode.com/@vishsec)

@vishsec · Bangalore,India

  
[View profile](https://hashnode.com/@vishsec)

Software Engineer

## Responses 1

Markdown supported · ⌘↵ to post

[Matt Hanson](https://hashnode.com/@matthanson)

[1d ago](https://hashnode.com/posts/building-a-fine-tuned-nutri-score-classifier/6a69be668184976f1c4345fc/comment/6af32537320cba0b78611d58)

Really enjoyed the honesty of this one — especially the data audit before any modeling. I research fast-food menus, so I live in the same problem from the other side: nutrition panels that look complete until you actually sample the rows, and grades that quietly depend on which formula version was used. Your point about the beverage rules and the India eval gap is the part that stuck with me. A milkshake and a bottle of water are both drinks, but they are not scored anything like each other, and a model that misses that will mislead people. Thanks for publishing the audit, not just the accuracy number.