fix spacing

sparsecoding/data/datasets/field.py

@@ -22,7 +22,7 @@ class FieldDataset(Dataset):
     patch_size : int
         Side length of patches for sparse dictionary learning.
     stride : int, optional
-        Stride for sampling patches. If not specified, set to `patch_size` 
+        Stride for sampling patches. If not specified, set to `patch_size`
         (non-overlapping patches).
     """
 

sparsecoding/data/datasets/vanhateren.py

@@ -0,0 +1,57 @@
+import os
+
+from scipy.io import loadmat
+import torch
+from torch.utils.data import Dataset
+
+from sparsecoding.data.transforms.patch import patchify
+
+
+class VanHaterenDataset(Dataset):
+    """Dataset used in Olshausen & Field (1996).
+
+    Paper:
+        https://courses.cs.washington.edu/courses/cse528/11sp/Olshausen-nature-paper.pdf
+        Emergence of simple-cell receptive field properties
+        by learning a sparse code for natural images.
+
+    Parameters
+    ----------
+    root : str
+        Location to download the dataset to.
+    patch_size : int
+        Side length of patches for sparse dictionary learning.
+    """
+
+    B = 10
+    C = 1
+    H = 512
+    W = 512
+
+    def __init__(
+        self,
+        root: str,
+        patch_size: int = 8,
+    ):
+        self.P = patch_size
+
+        root = os.path.expanduser(root)
+        os.system(f"mkdir -p {root}")
+        if not os.path.exists(f"{root}/field.mat"):
+            os.system("wget https://rctn.org/bruno/sparsenet/IMAGES.mat")
+            os.system(f"mv IMAGES.mat {root}/field.mat")
+
+        self.images = torch.tensor(loadmat(f"{root}/field.mat")["IMAGES"])  # [H, W, B]
+        assert self.images.shape == (self.H, self.W, self.B)
+
+        self.images = torch.permute(self.images, (2, 0, 1))  # [B, H, W]
+        self.images = torch.reshape(self.images, (self.B, self.C, self.H, self.W))  # [B, C, H, W]
+
+        self.patches = patchify(patch_size, self.images)  # [B, N, C, P, P]
+        self.patches = torch.reshape(self.patches, (-1, self.C, self.P, self.P))  # [B*N, C, P, P]
+
+    def __len__(self):
+        return self.patches.shape[0]
+
+    def __getitem__(self, idx):
+        return self.patches[idx]

sparsecoding/data/transforms/patch.py

@@ -80,8 +80,8 @@ def patchify(
             C is the number of channels,
             H is the image height,
             W is the image width.
-    stride : int, optional 
-        Stride between patches in pixel space. If not specified, set to 
+    stride : int, optional
+        Stride between patches in pixel space. If not specified, set to
         `patch_size` (non-overlapping patches).
 
     Returns