ltx23_vocoder

LTX-2.3 vocoder stage

Description

Mel spectrogram [B, C, T, M] -> waveform [B, out_channels, samples]. Channel and mel dims are flattened into conv channels; each upsample stage halves the channel count and averages three parallel ResNet branches.

Usage

ltx23_vocoder(in_channels = 128L, hidden_channels = 1536L, out_channels = 2L,
              upsample_kernel_sizes = c(11L, 4L, 4L, 4L, 4L, 4L),
              upsample_factors = c(5L, 2L, 2L, 2L, 2L, 2L),
              resnet_kernel_sizes = c(3L, 7L, 11L),
              resnet_dilations = list(c(1L, 3L, 5L), c(1L, 3L, 5L), c(1L, 3L, 5L)),
              antialias_ratio = 2L, antialias_kernel_size = 12L,
              final_bias = FALSE)

Arguments

  • in_channels: Integer. Flattened input channels (C * mel bins / 1).
  • hidden_channels: Integer.
  • out_channels: Integer.
  • resnet_kernel_sizes: Integer vector.
  • resnet_dilations: List of integer vectors.
  • final_bias: Logical.
  • upsample_kernel_sizes,upsample_factors: Integer vectors.
  • antialias_ratio,antialias_kernel_size: Integers.