VaeModel: VaeModel

In gongx030/seatac: ATAC-seq V-plot analysis

VaeModel

Description

Build a VAE model for V-plot of multiple ATAC-seq datasets. This model takes the stacked V-plots of the same genomic regions as the input.

Usage

VaeModel(
  x,
  latent_dim = 10L,
  filters0 = 128L,
  filters = 32L,
  kernel_size = 3L,
  downsample_layers = 4L,
  upsample_layers = 4L,
  strides = c(2L, 2L),
  momentum = 0.8,
  rate = 0.1,
  name = NULL
)

Arguments

x	a Vplots or a VplotsList object
latent_dim	Latent dimension (default: 10L)
filters0	Filter size after the latent layer (default: 128L)
filters	Initial filter size of convolution layer (default: 32L)
kernel_size	Kernel size in convolution and deconvolution layers (default: 3L)
downsample_layers	Downsample layers (default: 4L)
upsample_layers	Upsample layers (default: 4L)
strides	Convolution strides
momentum	Momentum in BatchNormalization layer (default: 0.8)
rate	Dropout rate (default: 0.1)
name	Model name

Value

a VaeModel object

gongx030/seatac documentation built on April 15, 2023, 5:53 a.m.