test inputs are in the data/ folder including: data.txt, labels.txt and peaks.txt
- peak count matrix: (required) for training model, peaks(rows) x cells(columns). e.g. data/data.txt
- ground truth labels: (optional) for comparing predicted clustering assignments. e.g. data/labels.txt
outputs are saved in default folder output/ or user-specified folder including:
- model.pt: trained SCALE model
- feature.txt: latent feature of input data
- cluster_assignments.txt: predicted cluster assignments by k-means on latent feature
- imputed_data.txt: imputed data reconstructed from latent feature via decoder of SCALE