Build a Model
- mmit.create_model(encoder_name, decoder_name, classes, task='segmentation', encoder_cfg=None, decoder_cfg=None)
Build a model from an encoder and a decoder.
- Parameters:
encoder_name (
str) – The name of the encoder.decoder_name (
str) – The name of the decoder.classes (
int) – The number of classes.task (
str) – The task of the model.encoder_cfg (
Optional[dict]) – Keyword arguments for the encoder. Check the specific encoder docs for more info.decoder_cfg (
Optional[dict]) – Keyword arguments for the decoder. Check the specific decoder docs for more info.