---
license: apache-2.0
tags:
- medical-imaging
- pathology
- pytorch
- resnet
- image-classification
datasets:
- medmnist/pathmnist
metrics:
- accuracy
model-index:
- name: ODE PathMNIST Classifier
  results:
  - task:
      type: image-classification
      name: Colorectal Pathology Classification
    dataset:
      name: PathMNIST
      type: medmnist/pathmnist
    metrics:
    - type: accuracy
      value: 0.9596
      name: Validation Accuracy
---

# ODE PathMNIST Classifier — Medical Pathology AI

**By Llewellyn Systems Inc** | Part of [ODE Science](https://www.llewellynsystems.com/products/ode-science)

## Model Description
ResNet18-based classifier for colorectal polyp histology images. 9-class classification on PathMNIST benchmark.

## Performance
- **95.96% validation accuracy** on PathMNIST (9 classes, ~100K images)
- Trained on NVIDIA Tesla T4 GPU
- ResNet18 backbone with custom classification head

## Classes
1. Adipose | 2. Background | 3. Debris | 4. Lymphocytes | 5. Mucus
6. Smooth Muscle | 7. Normal Colon Mucosa | 8. Cancer Stroma | 9. Colorectal Adenocarcinoma

## About ODE Science
AI-powered scientific research platform featuring molecular modeling, medical imaging, protein analysis, and quantum computing.

[Website](https://www.llewellynsystems.com) | [ODE Science](https://www.llewellynsystems.com/products/ode-science)