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.
]add Metalhead| Model Name | Function | Pre-trained? |
|---|---|---|
| VGG-11 | VGG |
N |
| VGG-11 (w/ BN) | VGG |
N |
| VGG-13 | VGG |
N |
| VGG-13 (w/ BN) | VGG |
N |
| VGG-16 | VGG |
N |
| VGG-16 (w/ BN) | VGG |
N |
| VGG-19 | VGG |
N |
| VGG-19 (w/ BN) | VGG |
N |
| ResNet-18 | ResNet |
N |
| ResNet-34 | ResNet |
N |
| ResNet-50 | ResNet |
N |
| ResNet-101 | ResNet |
N |
| ResNet-152 | ResNet |
N |
| GoogLeNet | GoogLeNet |
N |
| Inception-v3 | Inception3 |
N |
| SqueezeNet | SqueezeNet |
N |
| DenseNet-121 | DenseNet |
N |
| DenseNet-161 | DenseNet |
N |
| DenseNet-169 | DenseNet |
N |
| DenseNet-201 | DenseNet |
N |
| ResNeXt-50 | ResNeXt |
N |
| ResNeXt-101 | ResNeXt |
N |
| ResNeXt-152 | ResNeXt |
N |
| MobileNetv2 | MobileNetv2 |
N |
| MobileNetv3 | MobileNetv3 |
N |
| MLPMixer | MLPMixer |
N |
You can find the Metalhead.jl getting started guide here.