Harmonic embedding

Summary: Implements the positional embedding of NeRF

Reviewed By: nikhilaravi

Differential Revision: D25684406

fbshipit-source-id: 9f3b657babacff48bd6a0497d7a859607ffa5f89

projects/nerf/nerf/harmonic_embedding.py

+# Copyright (c) Facebook, Inc. and its affiliates. All rights reserved.
+import torch
+
+
+class HarmonicEmbedding(torch.nn.Module):
+    def __init__(
+        self,
+        n_harmonic_functions: int = 6,
+        omega0: float = 1.0,
+        logspace: bool = True,
+    ):
+        """
+        Given an input tensor `x` of shape [minibatch, ... , dim],
+        the harmonic embedding layer converts each feature
+        in `x` into a series of harmonic features `embedding`
+        as follows:
+            ```
+            embedding[..., i*dim:(i+1)*dim] = [
+                sin(x[..., i]),
+                sin(f_1*x[..., i]),
+                sin(f_2*x[..., i]),
+                ...
+                sin(f_N * x[..., i]),
+                cos(x[..., i]),
+                cos(f_1*x[..., i]),
+                cos(f_2*x[..., i]),
+                ...
+                cos(f_N * x[..., i])
+            ]
+            ```
+        where N corresponds to `n_harmonic_functions`, and f_i is a scalar
+        denoting the i-th frequency of the harmonic embedding.
+
+        If `logspace==True`, the frequencies `[f_1, ..., f_N]` are
+        either powers of 2:
+            `f_1, ..., f_N = 2**torch.arange(n_harmonic_functions)`
+
+        If `logspace==False`, frequencies are linearly spaced  between
+        `1.0` and `2**(n_harmonic_functions-1)`:
+            `f_1, ..., f_N = torch.linspace(
+                1.0, 2**(n_harmonic_functions-1), n_harmonic_functions
+            )`
+
+        Note that `x` is also premultiplied by the base frequency `omega0`
+        before evaluting the harmonic functions.
+        """
+        super().__init__()
+
+        if logspace:
+            frequencies = 2.0 ** torch.arange(
+                n_harmonic_functions,
+                dtype=torch.float32,
+            )
+        else:
+            frequencies = torch.linspace(
+                1.0,
+                2.0 ** (n_harmonic_functions - 1),
+                n_harmonic_functions,
+                dtype=torch.float32,
+            )
+
+        self.register_buffer("_frequencies", omega0 * frequencies)
+
+    def forward(self, x: torch.Tensor):
+        """
+        Args:
+            x: tensor of shape [..., dim]
+        Returns:
+            embedding: a harmonic embedding of `x`
+                of shape [..., n_harmonic_functions * dim * 2]
+        """
+        embed = (x[..., None] * self._frequencies).view(*x.shape[:-1], -1)
+        return torch.cat((embed.sin(), embed.cos()), dim=-1)