| """Utility functions for diffusion models.""" |
| from __future__ import annotations |
|
|
| import jax.numpy as jnp |
|
|
|
|
| def extract_and_expand(arr: jnp.ndarray, t_index: jnp.ndarray, ndim: int) -> jnp.ndarray: |
| """Extract values from a 1D array and expand. |
| |
| This function is not jittable. |
| |
| Args: |
| arr: 1D of shape (num_timesteps, ). |
| t_index: storing index values < self.num_timesteps, |
| shape (batch, ) or has ndim dimension. |
| ndim: number of dimensions for an array of shape (batch, ...). |
| |
| Returns: |
| Expanded array of shape (batch, ...), expanded axes have dim 1. |
| """ |
| if arr.ndim != 1: |
| raise ValueError(f"arr must be 1D, got {arr.ndim}D.") |
| x = arr[t_index] |
| return expand(x, ndim) |
|
|
|
|
| def expand(x: jnp.ndarray, ndim: int) -> jnp.ndarray: |
| """Expand. |
| |
| This function is not jittable. |
| |
| Args: |
| x: a 1D or nD array. |
| ndim: number of dimensions as output. |
| |
| Returns: |
| Expanded array, expanded axes have dim 1. |
| """ |
| if x.ndim == 1: |
| return jnp.expand_dims(x, axis=tuple(range(1, ndim))) |
| if x.ndim == ndim: |
| return x |
| raise ValueError(f"t_index must be 1D or {ndim}D, got {x.ndim}D.") |
|
|
