controlnet training resize inputs to multiple of 8

Author: williamberman

examples/controlnet/train_controlnet.py

@@ -25,6 +25,7 @@
 import torch
 import torch.nn.functional as F
 import torch.utils.checkpoint
+import torchvision.transforms.functional as TF
 import transformers
 from accelerate import Accelerator
 from accelerate.logging import get_logger
@@ -604,9 +605,22 @@ def tokenize_captions(examples, is_train=True):
         )
         return inputs.input_ids
 
+    def resize_multiple_8(image):
+        height = image.height
+        width = image.width
+
+        # round down to nearest multiple of 8
+        height = (height // 8) * 8
+        width = (width // 8) * 8
+
+        image = TF.resize(image, (height, width), transforms.InterpolationMode.BILINEAR)
+
+        return image
+
     image_transforms = transforms.Compose(
         [
             transforms.Resize(args.resolution, interpolation=transforms.InterpolationMode.BILINEAR),
+            resize_multiple_8,
             transforms.ToTensor(),
             transforms.Normalize([0.5], [0.5]),
         ]
@@ -615,6 +629,7 @@ def tokenize_captions(examples, is_train=True):
     conditioning_image_transforms = transforms.Compose(
         [
             transforms.Resize(args.resolution, interpolation=transforms.InterpolationMode.BILINEAR),
+            resize_multiple_8,
             transforms.ToTensor(),
         ]
     )