US Attractions Classifier
This model classifies text descriptions of US tourist attractions as Natural or Human-made. It fine-tunes a pretrained distilbert-base-uncased model with AutoGluon MultiModal.
Inputs and outputs
Supply a pandas DataFrame with a text column containing attraction descriptions.
predict() returns Natural or Human-made.
| Label | Meaning |
|---|---|
0 |
Natural |
1 |
Human-made |
Data and training
Source: US Attractions Dataset.
| Training setting | Recorded value |
|---|---|
| Framework | AutoGluon MultiModal 1.6.1 |
| Text backbone | distilbert-base-uncased |
| Selection metric | Validation accuracy |
| Seed | 24679 |
| Max epochs | 5 |
| Learning rate | 2e-5 |
| Fine-tuning mode | norm_fit (62,978 / 66,364,418 trainable parameters) |
Evaluation
| Test metric | Value |
|---|---|
| Accuracy | 0.9333 |
| Weighted F1 | 0.9333 |