import torch from torch.utils.data import DataLoader, TensorDataset from utils import extract_colors, preprocess_data def create_dataloader(N: int = 1e8, **kwargs): rgb_tensor = torch.rand((int(N), 3), dtype=torch.float32) rgb_tensor = preprocess_data(rgb_tensor) # Creating a dataset and data loader dataset = TensorDataset(rgb_tensor, torch.zeros(len(rgb_tensor))) train_dataloader = DataLoader(dataset, **kwargs) return train_dataloader def create_gray_supplement(N: int = 50): linear_space = torch.linspace(0, 1, N) gray_tensor = linear_space.unsqueeze(1).repeat(1, 3) gray_tensor = preprocess_data(gray_tensor) return [(gray_tensor[i], f"gray{i/N:2.4f}") for i in range(len(gray_tensor))] def create_named_dataloader(N: int = 0, **kwargs): rgb_tensor, xkcd_color_names = extract_colors() rgb_tensor = preprocess_data(rgb_tensor) # Creating a dataset with RGB values and their corresponding color names dataset_with_names = [ (rgb_tensor[i], xkcd_color_names[i].replace("xkcd:", "")) for i in range(len(rgb_tensor)) ] if N > 0: dataset_with_names += create_gray_supplement(N) train_dataloader_with_names = DataLoader(dataset_with_names, **kwargs) return train_dataloader_with_names if __name__ == "__main__": batch_size = 4 train_dataloader = create_dataloader(N=1e6, batch_size=batch_size, shuffle=True) print(len(train_dataloader.dataset)) train_dataloader_with_names = create_named_dataloader( batch_size=batch_size, shuffle=True ) # Extract a sample from the DataLoader sample_data = next(iter(train_dataloader)) # Sample RGB values and their corresponding dummy labels sample_rgb_values, _ = sample_data print(sample_rgb_values) # Extract a sample from the new DataLoader sample_data_with_names = next(iter(train_dataloader_with_names)) # Sample RGB values and their corresponding color names sample_rgb_values_with_names, sample_color_names = sample_data_with_names print(sample_rgb_values_with_names, sample_color_names)