ECG Arrhythmia Detection — CNN-BiLSTM
Multi-class arrhythmia classification from ECG signals trained on the MIT-BIH Arrhythmia Database.
Model Performance
| Class | F1 |
|---|---|
| N — Normal | 0.99 |
| L — LBBB | 1.00 |
| R — RBBB | 1.00 |
| V — PVC | 0.98 |
| A — APC | 0.91 |
| Macro avg | 0.98 |
Overall accuracy: 99%
Usage
from huggingface_hub import hf_hub_download
import torch, json, pickle
config = json.load(open(hf_hub_download("dheerajthuvara/ecg-arrhythmia-detection", "models/model_config.json")))
encoder = pickle.load(open(hf_hub_download("dheerajthuvara/ecg-arrhythmia-detection", "models/label_encoder.pkl"), "rb"))
# load model weights and run inference
Links
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support