models¶

mfai.pytorch.models.load_from_settings_file(model_name, in_channels, out_channels, settings_path, input_shape)[source]¶

Instanciate a model from a settings file with Schema validation.

Return type:

Parameters:

model_name (str)
in_channels (int)
out_channels (int)
settings_path (Path)
input_shape (tuple[int, ...])

`archesweather`
`base`	Interface contract for our models.
`clip`	Implementation of CLIP (Contrastive Langage-Image Pre-training) model.
`deeplabv3`
`gan_dgmr`	Submodules of DGMR model.
`half_unet`
`identity`
`llms`
`nlam`	Graph Neural Network architectures adapted from https://github.com/mllam/neural-lam.
`pangu`
`resnet`
`segformer`	SegFormer adapted from https://github.com/lucidrains/segformer-pytorch.
`swinunetr`
`unet`	pytorch models wrapped for DSM/LabIA projects.
`unetrpp`	UnetR++ Vision transformer based on: "Shaker et al., Adapted from https://github.com/Amshaker/unetr_plus_plus Added 2d support and Bilinear interpolation for upsampling.
`utils`
`vit`	VIT adapted from Lucidrain's repo https://github.com/lucidrains/vit-pytorch.
`weather_projector`	This model takes weather/2d inputs (batch, features, height, width) and produces tokens for multimodal language models.