---
dataset_info:
  features:
  - name: full_colour
    dtype: image
  - name: sketch
    dtype: image
  splits:
  - name: train
    num_bytes: 170204480.0
    num_examples: 922
  download_size: 170225532
  dataset_size: 170204480.0
---
# Dataset Card for "one_piece_dataset"

<h1> This dataset contains 922 images taken from the one piece anime, with each row containing the coloured image and the sketch one. </h1>
<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>