Metalhead.jl
Metalhead.jl copied to clipboard
Computer vision models for Flux
Metalhead
Metalhead.jl provides standard machine learning vision models for use with Flux.jl. The architectures in this package make use of pure Flux layers, and they represent the best-practices for creating modules like residual blocks, inception blocks, etc. in Flux. Metalhead also provides some building blocks for more complex models in the Layers module.
Installation
]add Metalhead
Available models
Model Name | Function | Pre-trained? |
---|---|---|
VGG | VGG |
Y (w/o BN) |
ResNet | ResNet |
Y |
WideResNet | WideResNet |
Y |
GoogLeNet | GoogLeNet |
N |
Inception-v3 | Inceptionv3 |
N |
Inception-v4 | Inceptionv4 |
N |
InceptionResNet-v2 | Inceptionv3 |
N |
SqueezeNet | SqueezeNet |
Y |
DenseNet | DenseNet |
N |
ResNeXt | ResNeXt |
Y |
MobileNetv1 | MobileNetv1 |
N |
MobileNetv2 | MobileNetv2 |
N |
MobileNetv3 | MobileNetv3 |
N |
EfficientNet | EfficientNet |
N |
MLPMixer | MLPMixer |
N |
ResMLP | ResMLP |
N |
gMLP | gMLP |
N |
ViT | ViT |
N |
ConvNeXt | ConvNeXt |
N |
ConvMixer | ConvMixer |
N |
To contribute new models, see our contributing docs.
Getting Started
You can find the Metalhead.jl getting started guide here.