A-MNIST / A-MNIST.py
gorar's picture
Update the dataset version
da9a9d6
raw
history blame
2.95 kB
"""Augmented MNIST Data Set"""
import struct
import numpy as np
import datasets
from datasets.tasks import ImageClassification
_DESCRIPTION = """\
The dataset is built on top of MNIST.
It consists from 130K of images in 10 classes - 120K training and 10K test samples.
The training set was augmented with additional 60K images.
"""
_URLS = {
"train_images": "data/train-images-idx3-ubyte.gz",
"train_labels": "data/train-labels-idx1-ubyte.gz",
"test_images": "data/t10k-images-idx3-ubyte.gz",
"test_labels": "data/t10k-labels-idx1-ubyte.gz",
}
class AMNIST(datasets.GeneratorBasedBuilder):
"""A-MNIST Data Set"""
BUILDER_CONFIGS = [
datasets.BuilderConfig(
name="amnist",
version=datasets.Version("1.1.0"),
description=_DESCRIPTION,
)
]
def _info(self):
return datasets.DatasetInfo(
description=_DESCRIPTION,
features=datasets.Features(
{
"image": datasets.Image(),
"label": datasets.features.ClassLabel(names=["0", "1", "2", "3", "4", "5", "6", "7", "8", "9"]),
}
),
supervised_keys=("image", "label"),
task_templates=[
ImageClassification(
image_column="image",
label_column="label",
)
],
)
def _split_generators(self, dl_manager):
urls_to_download = _URLS
downloaded_files = dl_manager.download_and_extract(urls_to_download)
return [
datasets.SplitGenerator(
name=datasets.Split.TRAIN,
gen_kwargs={
"filepath": [downloaded_files["train_images"],
downloaded_files["train_labels"]],
"split": "train",
},
),
datasets.SplitGenerator(
name=datasets.Split.TEST,
gen_kwargs={
"filepath": [downloaded_files["test_images"],
downloaded_files["test_labels"]],
"split": "test",
},
),
]
def _generate_examples(self, filepath, split):
"""This function returns the examples in the raw form."""
# Images
with open(filepath[0], "rb") as f:
# First 16 bytes contain some metadata
_ = f.read(4)
size = struct.unpack(">I", f.read(4))[0]
_ = f.read(8)
images = np.frombuffer(f.read(), dtype=np.uint8).reshape(size, 28, 28)
# Labels
with open(filepath[1], "rb") as f:
# First 8 bytes contain some metadata
_ = f.read(8)
labels = np.frombuffer(f.read(), dtype=np.uint8)
for idx in range(size):
yield idx, {"image": images[idx], "label": str(labels[idx])}