*My post explains Oxford 102 Flower.
Flowers102() can use Oxford 102 Flower dataset as shown below:
*Memos:
- The 1st argument is
root(Required-Type:strorpathlib.Path). *An absolute or relative path is possible. - The 2nd argument is
split(Optional-Default:"train"-Type:str). *"train"(1,020 images),"val"(1,020 images) or"test"(6,149 images) can be set to it. - The 3rd argument is
transform(Optional-Default:None-Type:callable). - The 4th argument is
target_transform(Optional-Default:None-Type:callable). - The 5th argument is
download(Optional-Default:False-Type:bool):
*Memos:
- If it's
True, the dataset is downloaded from the internet and extracted(unzipped) toroot. - If it's
Trueand the dataset is already downloaded, it's extracted. - If it's
Trueand the dataset is already downloaded and extracted, nothing happens. - It should be
Falseif the dataset is already downloaded and extracted because it's faster. - You can manually download and extract the dataset(
102flowers.tgzwithimagelabels.matandsetid.matfffrom here todata/flowers-102/.
- If it's
- About the label from the categories(classes) for the train and validation image indices, 0 is 0~9, 1 is 10~19, 2 is 20~29, 3 is 30~39, 4 is 40~49, 5 is 50~59, 6 is 60~69, 7 is 70~79, 8 is 80~89, 9 is 90~99, etc.
- About the label from the categories(classes) for the test image indices, 0 is 0~19, 1 is 20~59, 2 is 60~79, 3 is 80~115, 4 is 116~160, 5 is 161~185, 6 is 186~205, 7 is 206~270, 8 is 271~296, 9 is 297~321, etc.
from torchvision.datasets import Flowers102
train_data = Flowers102(
root="data"
)
train_data = Flowers102(
root="data",
split="train",
transform=None,
target_transform=None,
download=False
)
val_data = Flowers102(
root="data",
split="val"
)
test_data = Flowers102(
root="data",
split="test"
)
len(train_data), len(val_data), len(test_data)
# (1020, 1020, 6149)
train_data
# Dataset Flowers102
# Number of datapoints: 1020
# Root location: data
# split=train
train_data.root
# 'data'
train_data._split
# 'train'
print(train_data.transform)
# None
print(train_data.target_transform)
# None
train_data.download
# <bound method Flowers102.download of Dataset Flowers102
# Number of datapoints: 1020
# Root location: data
# split=train>
len(set(train_data._labels)), train_data._labels
# (102,
# [0, 0, 0, ..., 1, ..., 2, ..., 3, ..., 4, ..., 5, ..., 6, ..., 101])
train_data[0]
# (<PIL.Image.Image image mode=RGB size=754x500>, 0)
train_data[1]
# (<PIL.Image.Image image mode=RGB size=624x500>, 0)
train_data[2]
# (<PIL.Image.Image image mode=RGB size=667x500>, 0)
train_data[10]
# (<PIL.Image.Image image mode=RGB size=500x682>, 1)
train_data[20]
# (<PIL.Image.Image image mode=RGB size=667x500>, 2)
val_data[0]
# (<PIL.Image.Image image mode=RGB size=606x500>, 0)
val_data[1]
# (<PIL.Image.Image image mode=RGB size=667x500>, 0)
val_data[2]
# (<PIL.Image.Image image mode=RGB size=500x628>, 0)
val_data[10]
# (<PIL.Image.Image image mode=RGB size=500x766>, 1)
val_data[20]
# (<PIL.Image.Image image mode=RGB size=624x500>, 2)
test_data[0]
# (<PIL.Image.Image image mode=RGB size=523x500>, 0)
test_data[1]
# (<PIL.Image.Image image mode=RGB size=666x500>, 0)
test_data[2]
# (<PIL.Image.Image image mode=RGB size=595x500>, 0)
test_data[20]
# (<PIL.Image.Image image mode=RGB size=500x578>, 1)
test_data[60]
# (<PIL.Image.Image image mode=RGB size=500x625>, 2)
import matplotlib.pyplot as plt
def show_images(data, ims, main_title=None):
plt.figure(figsize=(10, 5))
plt.suptitle(t=main_title, y=1.0, fontsize=14)
for i, j in enumerate(ims, start=1):
plt.subplot(2, 5, i)
im, lab = data[j]
plt.imshow(X=im)
plt.title(label=lab)
plt.tight_layout()
plt.show()
train_ims = (0, 1, 2, 10, 20, 30, 40, 50, 60, 70)
val_ims = (0, 1, 2, 10, 20, 30, 40, 50, 60, 70)
test_ims = (0, 1, 2, 20, 60, 80, 116, 161, 186, 206)
show_images(data=train_data, ims=train_ims, main_title="train_data")
show_images(data=train_data, ims=val_ims, main_title="val_data")
show_images(data=test_data, ims=test_ims, main_title="test_data")
