---
license: mit
tags:
- medical
- pneumonia
- chest-xray
- explainable-ai
- pytorch
---

# Pneumonia Detection Model

Trained ResNet-18 model for pneumonia detection from chest X-rays with explainable AI features.

## Model Details

- **Architecture**: ResNet-18 with transfer learning
- **Task**: Binary classification (Normal vs Pneumonia)
- **Input**: 224x224 RGB chest X-ray images
- **Training Data**: Kaggle Chest X-Ray Pneumonia Dataset

## Files

- `models/pneumonia_classifier.pth`: Trained model weights (~87 MB)
- `embeddings/embeddings.npy`: Feature embeddings for similarity search
- `embeddings/similarity_index.faiss`: FAISS index for fast retrieval
- `embeddings/labels.npy`: Training labels
- `embeddings/paths.pkl`: Training image paths

## Usage

```python
from huggingface_hub import hf_hub_download

# Download model
model_path = hf_hub_download(
    repo_id="Ollyvern/xray-pneumonia-model",
    filename="models/pneumonia_classifier.pth"
)
```

## Streamlit App

This model is used in the X-Ray Transparency Lab Streamlit app with multi-view explainable AI.

**⚠️ Disclaimer**: This is a research/educational model and should NOT be used for clinical diagnosis.