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
1from transformers import pipeline
2
3classifier = pipeline(
4 "text-classification",
5 model="cloudwoowoo/zooguide-bert-animal-fact-assistant",
6 tokenizer="cloudwoowoo/zooguide-bert-animal-fact-assistant"
7)
8
9classifier("Tell me about an elephant with a trunk.")
