Sweet vs Savory Classifier
This model classifies photographs of food items as sweet (0) or savory (1). It fine-tunes a pretrained ResNet-50 with AutoGluon MultiModal.
Inputs and outputs
Supply a pandas DataFrame with an image column containing local image-file paths. The predictor reads the pixels; labels and provenance fields are not inputs. predict() returns 0 or 1, and predict_proba() returns class probabilities.
| Label | Meaning |
|---|---|
0 |
Sweet |
1 |
Savory |
Data and training
Source Dataset: kwongnon/2026-24679-image-dataset.
| Training setting | Recorded value |
|---|---|
| Framework | AutoGluon MultiModal 1.6.1 |
| Image backbone | timm resnet50, pretrained, fine-tuned for two classes |
| Preset and branch | medium_quality; model.names = ["timm_image"] |
| Selection metric | Validation accuracy |
| Fit budget and seed | 300 seconds; 24679 |
| Recorded environment | Linux; Python 3.13.15; PyTorch 2.11.0+cpu |
The notebook reports 327.53 seconds of training for the final ResNet-50 model with a final validation accuracy of 0.7143.
Evaluation
| Test metric | Value |
|---|---|
| Test Accuracy | 0.3750 |
| Weighted F1 | 0.3849 |