scaden_train: Trains a Scaden model
In PelzKo/immunedeconv2: Second generation methods for immune cell deconvolution

scaden_train

R Documentation

Trains a Scaden model

Description

Trains a Scaden model on pre-processed training data.

Usage

scaden_train(
  h5ad_processed,
  temp_dir = NULL,
  batch_size = 128,
  learning_rate = 1e-04,
  model_path = NULL,
  steps = 5000,
  verbose = FALSE
)

Arguments

`h5ad_processed`	pre-processed training data
`temp_dir`	The temporary directory to use for the computations (optional)
`batch_size`	Batch size to use for training. Default: 128
`learning_rate`	Learning rate used for training. Default: 0.0001
`model_path`	Path to store the model in
`steps`	Number of training steps
`verbose`	Whether to produce an output on the console. (default: false)