Downloads the CIFAR10 training dataset or fetches it locally.

Returns a tuple of format:

{{images_binary, images_type, images_shape},
 {labels_binary, labels_type, labels_shape}}

If you want to one-hot encode the labels, you can:

labels_binary
|> Nx.from_binary(labels_type)
|> Nx.new_axis(-1)
|> Nx.equal(Nx.tensor(Enum.to_list(0..9)))

options

Options.

• :base_url - Dataset base URL.

  Defaults to "https://www.cs.toronto.edu/~kriz/"

• :dataset_file - Dataset filename.

  Defaults to "cifar-10-binary.tar.gz"

• :cache_dir - Cache directory.

  Defaults to System.tmp_dir!()

examples

Examples

iex> Scidata.CIFAR10.download()
{{<<59, 43, 50, 68, 98, 119, 139, 145, 149, 149, 131, 125, 142, 144, 137, 129,
137, 134, 124, 139, 139, 133, 136, 139, 152, 163, 168, 159, 158, 158, 152,
148, 16, 0, 18, 51, 88, 120, 128, 127, 126, 116, 106, 101, 105, 113, 109,
112, ...>>, {:u, 8}, {50000, 3, 32, 32}},
{<<6, 9, 9, 4, 1, 1, 2, 7, 8, 3, 4, 7, 7, 2, 9, 9, 9, 3, 2, 6, 4, 3, 6, 6, 2,
    6, 3, 5, 4, 0, 0, 9, 1, 3, 4, 0, 3, 7, 3, 3, 5, 2, 2, 7, 1, 1, 1, ...>>,
  {:u, 8}, {50000}}}

Downloads the CIFAR10 test dataset or fetches it locally.

Accepts the same options as download/1.