DeiT-Small RGB Carbohydrate Regression - Stage 2 Fine-Tuned

This model predicts dish-level carbohydrate content from overhead RGB images from Nutrition5K.

Model version

This repository contains the stage 2 fine-tuned version of the model. The main checkpoint is:

  • best_model_stage2.keras

Dataset

  • Nutrition5K
  • Overhead RGB only
  • Train / validation / test split from prepared metadata CSV

Architecture

  • DeiT-Small distilled backbone pretrained on ImageNet
  • Stage 1: frozen backbone + regression head training
  • Stage 2: partial unfreezing of later transformer layers
  • Regression head with linear output for carbohydrate prediction

Target

  • total_carb
Downloads last month
68
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support