SpatialCL is a workflow for representation learning of spots in spatial transcriptome data, utilizing contrastive learning techniques.
The file description is as follows:
- Model.py : The artichture of the Siamese network, you can change it here;
- Tranier.py : A Tranier Class, you can configure the training parameters here;
- Transform.py : Different ways of adding noise are defined;
- GeneAggr2Set : The different way of aggregating expression values of gene pathways is defined
- workflow.py : Training workflow & infer workflow
