MLX Image Models

Introduction

This repository contains implementations for image models in MLX with functions to load pretrained weights from PyTorch. Example training code is provided in mimm/scripts.

Installation

pip install mimm

Usage

from mimm import get_model, list_models

# List available models
print(list_models())

# Load a model and convert weights from torchvision
model = get_model('resnet50', pretrained=True)

Models

Classification

• [x] AlexNet
• [x] VGG + BatchNorm variants
• [x] Resnet
• [x] Wide Resnet
• [x] MobileNetV3
• [x] ViT
• [x] Swin Transformer
• [x] Swin v2

Acknowledgements

This repository is based on the following repositories:

• PyTorch Image Models
• Torchvision