scgen.SCGENVAE¶
- class scgen.SCGENVAE(n_input, n_hidden=800, n_latent=10, n_layers=2, dropout_rate=0.1, log_variational=False, latent_distribution='normal', use_batch_norm='both', use_layer_norm='none', kl_weight=5e-05)[source]¶
Variational auto-encoder model.
- Parameters
- n_input
int Number of input genes
- n_hidden
int(default:800) Number of nodes per hidden layer
- n_latent
int(default:10) Dimensionality of the latent space
- n_layers
int(default:2) Number of hidden layers used for encoder and decoder NNs
- dropout_rate
float(default:0.1) Dropout rate for neural networks
- use_layer_norm {‘encoder’, ‘decoder’, ‘none’, ‘both’}
Literal[‘encoder’, ‘decoder’, ‘none’, ‘both’] (default:'none') Whether to use layer norm in layers
- kl_weight
float(default:5e-05) Weight for kl divergence
- n_input
Methods
generative(z)Runs the generative model.
get_reconstruction_loss(x, px)- rtype
inference(x)High level inference method.
loss(tensors, inference_outputs, ...)Compute the loss for a minibatch of data.
sample(tensors[, n_samples])Generate observation samples from the posterior predictive distribution.