zooguide-bert-animal-fact-assistant

ZooGuide-BERT Animal Fact Assistant

Model Description

This model is a fine-tuned distilbert-base-uncased text classifier for a small animal education assistant. It predicts which animal a user is asking about, then the application returns a curated short animal fact.

Task

Multi-class text classification.

Animal Classes

dog, cat, elephant, lion, tiger, horse, butterfly, spider, chicken, sheep

Dataset

The dataset is a custom zoo animal fact dataset created in the project notebook. Each example contains a short animal question, animal clue, or classroom-style sentence, paired with the correct animal label.

Training Details

• Base model: distilbert-base-uncased
• Training approach: Fine-tuning
• Epochs: 5
• Learning rate: 3e-5
• Evaluation metrics: accuracy and macro F1
• Hardware: Google Colab GPU recommended

Evaluation Results

Final evaluation results after running the notebook:

• Test accuracy: 0.9667
• Test macro F1: 0.9663

Intended Uses

This model is intended for educational demonstrations, classroom activities, museum-style animal learning tools, and beginner AI projects.

Limitations

This model is trained on a small custom dataset. It should not be used for scientific animal identification, wildlife safety, or expert biology tasks. It may perform poorly on animals outside the ten project classes.

How to Use

from transformers import pipeline

classifier = pipeline(
    "text-classification",
    model="cloudwoowoo/zooguide-bert-animal-fact-assistant",
    tokenizer="cloudwoowoo/zooguide-bert-animal-fact-assistant"
)

classifier("Tell me about an elephant with a trunk.")