OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

MEDVITV2

MEDVITV2 is an open-source medical-image classification model and research-code repository. It integrates KAN layers and Dilated Neighborhood Attention into a transformer architecture and provides pretrained weights, evaluation and training resources, and a Grad-CAM visualization tutorial.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

MEDVITV2 — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use MEDVITV2 when you need research code and model resources for medical-image classification experiments, particularly when evaluating across the project’s listed datasets or inspecting Grad-CAM visualizations.

Who it is for

  • Medical-imaging researchers comparing classification approaches.
  • Developers exploring model training or evaluation on medical-image datasets.
  • Practitioners who want to inspect model heatmaps using the provided Grad-CAM tutorial.

Documented uses

  • Train or evaluate MedViT models on medical datasets.
  • Start an evaluation workflow from the project’s Colab notebook.
  • Explore pretrained weights for model experiments.
  • Visualize model heatmaps with the repository’s Grad-CAM tutorial.
Before you choose

Requirements & limits

The repository is public and identifies its license as MIT. Its evaluation notebook is linked for opening in Google Colab.

  • The project provides both pretrained weights and code for training or evaluation.
  • The README specifies PyTorch 2.5.0 and natten 0.17.3, with CUDA 12.4-specific package references.
  • The evaluation tutorial is presented as supporting 17 medical datasets; the README also describes results on corrupted datasets.
  • A separate notebook is provided for Grad-CAM visualization.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare a local environment for MedViTV2

Install the repository’s documented dependencies before using its training or evaluation code.

Before you start

  • A local copy of the MEDVITV2 repository.
  • The README specifies PyTorch 2.5.0, torchvision 0.20.0, torchaudio 2.5.0 and natten 0.17.3, followed by the repository requirements.

Steps

  1. Clone the MEDVITV2 repository and enter its directory.

  2. Install the specified PyTorch, torchvision and torchaudio versions.

  3. Install natten 0.17.3 and the packages listed in requirements.txt.

Expected outcome

The README’s listed software dependencies are installed for the repository’s training and evaluation workflows.

Guide limitations

  • The README’s clone command contains malformed text, so the repository URL should be taken from the project page.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
Omid-Nejati/MedViTV2 · Official homepage title · retained in its source language
How we review sourcesSuggest a correction