Access Chainer dataset using indexer

dataset_indexer_exp.py

import numpy as np
import chainer
from chainer.datasets import TupleDataset

train, test = chainer.datasets.get_mnist()

# 1. Extract data easily without explicitly call concat_examples
imgs, labels = train.ix[:, :]
#imgs, labels = train.ix[:]  # same

print('imgs[2]', imgs[2].shape)
#print('imgs[2]', imgs[2])
print('labels[2]', labels[2])

# 2. Extract only labels, without extracting unnecessary imgs data
labels = train.ix[:, 1]
print('labels, len', len(labels))
print('labels', labels[:10])

# 3. Extract index whose label is 3, only for first 1000 dataset.
label3_index = np.argwhere(train.ix[:1000, 1] == 3)
print('label3_index', label3_index)

# Get image whose label is 3
imgs3 = train.ix[label3_index, 0]
print('imgs3', imgs3.shape)

# 4. Extract images whose label is 5 (Test for bool index access at axis=0)
imgs5 = train.ix[train.ix[:, 1] == 5, 0]
print('imgs5', imgs5.shape)

# 5. Test for bool index access at axis=0
#labels2 = train.ix[train.ix[:, 1] == 2, [False, True]]
#print('labels2', len(labels2), labels2)


tuple_dataset = TupleDataset([0, 1, 2], [0, 1, 4])
targets = tuple_dataset.ix[:, 1]
print('targets', targets)