troyskie/Beewatch-hive-classifier

BeeWatch Hive Classifier

Overview

BeeWatch Hive Classifier is an AI model that classifies beehive sounds into three categories:

• Healthy
• Warning
• Low Activity

The model is designed to assist beekeepers in monitoring hive health using audio recordings.

---

Model Details

• Model: ASTForAudioClassification
• Base Model: MIT/ast-finetuned-audioset-10-10-0.4593
• Framework: Hugging Face Transformers
• Task: Audio Classification

---

Labels

Label	Description
Healthy	Hive is functioning normally.
Warning	Hive exhibits unusual activity that may require inspection.
Low Activity	Hive activity is significantly reduced.

---

Dataset

The model was fine-tuned using a custom beehive audio dataset containing WAV recordings collected from different hive conditions.

Classes:

• Healthy
• Warning
• Low Activity

---

Usage

from transformers import pipeline

classifier = pipeline(
    "audio-classification",
    model="troyskie/Beewatch-hive-classifier"
)

result = classifier("sample.wav")
print(result)

---

Example Output

[
 {'label': 'Healthy', 'score': 0.96},
 {'label': 'Warning', 'score': 0.03},
 {'label': 'Low Activity', 'score': 0.01}
]

---

Limitations

• Performance depends on audio quality.
• Excessive environmental noise may reduce accuracy.
• The model should assist, not replace, manual hive inspection.

---

Intended Use

This model is intended for research and educational purposes in smart beekeeping and hive health monitoring.