summaryrefslogtreecommitdiff
path: root/src/text_recognizer/networks
diff options
context:
space:
mode:
Diffstat (limited to 'src/text_recognizer/networks')
-rw-r--r--src/text_recognizer/networks/misc.py5
1 files changed, 3 insertions, 2 deletions
diff --git a/src/text_recognizer/networks/misc.py b/src/text_recognizer/networks/misc.py
index 9440f9d..2fbab8f 100644
--- a/src/text_recognizer/networks/misc.py
+++ b/src/text_recognizer/networks/misc.py
@@ -21,8 +21,9 @@ def sliding_window(
"""
unfold = Unfold(kernel_size=patch_size, stride=stride)
- patches = unfold(images)
+ # Preform the slidning window, unsqueeze as the channel dimesion is lost.
+ patches = unfold(images).unsqueeze(1)
patches = rearrange(
- patches, "b (h w) c -> b c h w", h=patch_size[0], w=patch_size[1]
+ patches, "b c (h w) t -> b t c h w", h=patch_size[0], w=patch_size[1]
)
return patches