Mae-Model-MedMNIST-Predictor

Hali5/Mae-Model-MedMNIST-Predictor

Mae-Model-MedMNIST-Predictor is a image classification model from Hali5. Use it when you need a label for an image. The card lists the license as mit.

Using Masked Autoencoder model pretraining to Predict Medical Imaging Results.

Downloads · 30 days

Access

Public

Updated Sep 7, 2026

Repo size

9 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.pt7 GB · 100%

At a glance

Task: Image Classification
License: mit
Access: Public
Created: Aug 17, 2026
Updated: Sep 7, 2026
SHA: d1626f3d

Spaces

Hali5/MAE-PathMNIST-Predictor-Demo

Model card

From the Hugging Face model README

MAE-Model-MedMNIST-Predictor

Using Masked Autoencoder model pretraining to Predict Medical Imaging Results.

For Code of the Model Architecture and Training:

Github Model Code

For Live Demo:

Hugging Face Model Space Demo

Results:

The best predictor variant was the cross attention pooler with 95.11% test accuracy.

This test accuracy beats the best ResNet (ResNet-50 (28)) model accuracy from the official MedMNIST v2 dataset benchmark by 4%.

Here are the variants and their test accuracies and AUCs:

Variant	Accuracy	AUC
Benchmark ResNet-50 (28)	0.911	0.990
Linear Probe (GAP)	0.9415	0.9942
Cross Attention Pooler	0.9511	0.9942
Cross Attention Hybrid	0.9433	0.9954

Task: Image Classification
License: mit
Created: Aug 17, 2026
Updated: Sep 7, 2026