Update README.md

README.md

@@ -15,4 +15,32 @@ dataset_info:
 # Dataset Card for "one_piece_dataset"
 
 <h1> This dataset contains 922 images taken from the one piece anime (128x128x3), with each row containing the coloured image and the sketch one. </h1>
-<img alt='example images' src="./combined.png"/ > 
+<img alt='example images' src="./combined.png"/ > 
+
+<h2> Example Setup (the images are not normalized) </h2>
+<code>
+
+  
+    transform = Compose([
+                transforms.Resize((128, 128)),
+                transforms.ToTensor(),
+                transforms.Lambda(lambda t: (t * 2) - 1)
+    ])
+    
+    def train():
+
+    def transforms(examples):
+        examples["sketch_pixel_values"] = [transform(image.convert("RGB"))  for image in examples["sketch"]]
+        examples["full_colour_pixel_values"] = [ transform(image.convert("RGB")) for image in  examples["full_colour"]]
+        del examples["sketch"]
+        del examples["full_colour"]
+
+        return examples
+
+
+    dataset = load_dataset("pawlo2013/one_piece_dataset", split="train")
+
+    transformed_full_colour_dataset = dataset.with_transform(transforms)
+
+    dataloader = DataLoader(transformed_full_colour_dataset, batch_size=16, shuffle=True, num_workers=0)
+</code>