Multi resolution STFT loss module. See [Yamamoto et al., 2019](https://arxiv.org/abs/1910.11480) Args: fft_sizes (list): List of FFT

Randomly invert the phase of a signal

Base implementation of a pattern over a sequence with multiple codebooks. The codebook pattern consists in a layout, defining for each sequence s

STFT magnitude loss module. See [Arik et al., 2018](https://arxiv.org/abs/1808.06719) and [Engel et al., 2020](https://arxiv.org/abs/2001.046

Aggregates temporal dimension with attention. Also used with pos emb as global aggregation in both streams.

Embeds scalar timesteps into vector representations.

Wrapper for training a conditional audio diffusion model.

FIR pre-emphasis filtering module. Args: filter_type (str): Shape of the desired FIR filter ("hp", "fd", "aw"). Default: "hp" coe

This class serves three puposes: 1. Renames the class to MotionFormer. 2. Downloads the cfg from the original repo and patche

A module that applies multiple conditioners to an input dictionary based on the keys Args: conditioners: a dictionary of conditioner

Conditioner that takes a list of floats, normalizes them for a given range, and returns a list of embeddings

A conditioner that turns text into phonemes and embeds them using a lookup table Only works for English text Args: output_dim: t

A conditioner that uses a pretransform's encoder for conditioning Args: pretransform: an instantiated pretransform to use for condit