rajnish612/ast-heart-murmur-classifier

Heart Murmur AST Classifier

This model is a fine-tuned Audio Spectrogram Transformer (AST) for heart sound classification.

Classes

Label ID	Class
0	Absent
1	Murmur
2	Unknown

Base Model

• MIT/ast-finetuned-audioset-10-10-0.4593

Dataset

Fine-tuned on the CirCor DigiScope Phonocardiogram Dataset.

Performance

• Validation Accuracy: 86.2%
• Validation Loss: 0.4385

Usage

from transformers import ASTForAudioClassification

model = ASTForAudioClassification.from_pretrained(
    "rajnish612/ast-heart-murmur-classifier"
)

Feature Extractor

from transformers import AutoFeatureExtractor

feature_extractor = AutoFeatureExtractor.from_pretrained(
    "MIT/ast-finetuned-audioset-10-10-0.4593"
)

Prediction

inputs = feature_extractor(
    audio,
    sampling_rate=16000,
    return_tensors="pt"
)

outputs = model(**inputs)
prediction = outputs.logits.argmax(-1).item()

Labels

labels = {
    0: "Absent",
    1: "Murmur",
    2: "Unknown"
}

Intended Use

This model is intended for research and educational purposes in heart sound analysis and murmur detection.

Limitations

This model is not a medical device and should not be used for clinical diagnosis.