Return the current range width.

Encode a given input tensor with the specified sample rate at the given bandwidth. The RVQ encode method sets the appropriate number of quanti

Parameters ---------- x : torch.tensor Input wavs. Shape: (batch, channels, timesteps). n_q : in

Residual vector quantization on the given input tensor. Args: x (torch.Tensor): Input tensor. n_q (int): Number of qua

Parameters ---------- x : torch.tensor Input wavs. Shape should be (batch, channels, timesteps). layers

Pad for a convolution to make sure that the last window is full. Extra padding is added at the end. This is required to ensure that we can rebuild

Sync grad for buffers. If average is False, broadcast instead of averaging.

Simpler alternative to DistributedDataParallel, that doesn't rely on any black magic. For simple models it can also be as fast. Just call