[PyTorch]Read CIFAR10

#1

I don’t have CIFAR10 data locally, and then I download one, but it doesn’t download automatically, and the read fails.

import torchvision as tv
import torchvision.transforms as transforms
from torchvision.transforms import ToPILImage
import torch as t

transform = transforms.Compose([
    transforms.ToTensor()  
    ,transforms.Normalize((0.5,0.5,0.5),(0.5,0.5,0.5))  
])
trainset = tv.datasets.CIFAR10(
    root = './data',
    train = True,
    download = True, 
    transform = transform
)
trainloader = t.utils.data.DataLoader(
            trainset,
            batch_size = 4,
            shuffle = True,
            num_workers=2
)

---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-38-355678cb95ef> in <module>
      1 trainloader = t.utils.data.DataLoader(
----> 2             trainset,
      3             batch_size = 4,
      4             shuffle = True,
      5             num_workers=2

NameError: name 'trainset' is not defined

Best

#2

What do these lines of code return:

trainset = tv.datasets.CIFAR10(
    root = './data',
    train = True,
    download = True, 
    transform = transform
)

Do you get any error trying to download the data?

#3

The code doesn’t return anything.
Bset

#4

In that case trainset should be properly initialized.
What does print(trainset) afterwards print out?

#5

Got nothing and changed nothing

trainset = tv.datasets.CIFAR10(
    root = './data',
    train = True,
    download = True, 
    transform = transform
)
print(trainset)
#6

print(trainset) prints nothing out? :face_with_raised_eyebrow:
What does print(len(trainset)) output?

#7

Oh, I still didn’t get anything, but I copied the code to another block of code, and it’s done.
Sincerely thank you for helping me again.
Best

#8

Maybe the jupyter notebook have something wrong?

#9

Might be. I’ve seen all kind of weird restarts etc. with notebooks.
So if something doesn’t seem to right, I would recommend to run the script in a terminal, which should work.

#10

Thank you again. I’ll take your advice