---
library_name: pytorch
pipeline_tag: text-classification
tags:
- personality-prediction
- big-five
- essays
---

# Essays Big Five Personality Predictor

This repository contains a PyTorch deployment bundle trained in
`AFLPS-ML-Psycholinguistics` to predict Big Five personality dimensions from
essay text.

## Files

- `pytorch_model.bin`: PyTorch bundle containing model weights, TF-IDF
  vocabulary, IDF weights, threshold, and metrics.
- `vectorizers/essays_tfidf.npz`: portable TF-IDF artifact exported from the
  preprocessing environment.
- `deployment.py`: model architecture and preprocessing helpers.
- `predict.py`: command-line batch and single-text inference script.
- `requirements.txt`: minimal inference dependencies.

## Local Inference

```bash
pip install -r requirements.txt
python predict.py pytorch_model.bin --text "Paste a new essay here."
```

For a CSV with a `text` column:

```bash
python predict.py pytorch_model.bin --input essays.csv --output predictions.csv
```

## Output

The model returns probabilities and binary predictions for:

- `O`: Openness
- `C`: Conscientiousness
- `E`: Extraversion
- `A`: Agreeableness
- `N`: Neuroticism

Original local source repo upload target: `Lily-Trinh/PersonalityTraitsFromText`.