|
from functools import partial |
|
|
|
import jax |
|
import numpy as np |
|
|
|
|
|
def repeat_vmap(fun, in_axes=None): |
|
if in_axes is None: |
|
in_axes = [0] |
|
for axes in in_axes: |
|
fun = jax.vmap(fun, in_axes=axes) |
|
return fun |
|
|
|
|
|
def make_grid(patch_size: int | tuple[int, int]): |
|
if isinstance(patch_size, int): |
|
patch_size = (patch_size, patch_size) |
|
offset_h, offset_w = 1 / (2 * np.array(patch_size)) |
|
space_h = np.linspace(-0.5 + offset_h, 0.5 - offset_h, patch_size[0]) |
|
space_w = np.linspace(-0.5 + offset_w, 0.5 - offset_w, patch_size[1]) |
|
return np.stack(np.meshgrid(space_h, space_w, indexing='ij'), axis=-1) |
|
|
|
|
|
def interpolate_grid(coords, grid, order=0): |
|
""" |
|
Args: |
|
coords: Tensor de shape (B, H, W, 2) ou (H, W, 2) |
|
grid: Tensor de shape (B, H', W', C) |
|
""" |
|
|
|
if coords.ndim == 3: |
|
coords = coords[np.newaxis, ...] |
|
|
|
|
|
assert coords.ndim == 4, f"Dimensões inválidas para coords: {coords.shape}" |
|
assert grid.ndim == 4, f"Dimensões inválidas para grid: {grid.shape}" |
|
|
|
|
|
try: |
|
coords = coords.transpose((0, 3, 1, 2)) |
|
except ValueError as e: |
|
raise ValueError(f"Falha na transposição: {coords.shape} → (0,3,1,2)") from e |
|
|
|
|
|
coords = coords.at[:, 0].set(coords[:, 0] * grid.shape[-3] + (grid.shape[-3] - 1) / 2) |
|
coords = coords.at[:, 1].set(coords[:, 1] * grid.shape[-2] + (grid.shape[-2] - 1) / 2) |
|
|
|
|
|
map_coordinates = partial(jax.scipy.ndimage.map_coordinates, |
|
order=order, |
|
mode='nearest') |
|
|
|
return jax.vmap( |
|
jax.vmap( |
|
map_coordinates, |
|
in_axes=(2, None), |
|
out_axes=2 |
|
) |
|
)(grid, coords) |