Ei z S SK rS SKJr S SKJrJrJrJr S SK r
S SK J r SSK
JrJrJr SSKJr " S S \5 rg)
N)Path)AnyCallableOptionalUnion)Image )check_integrity download_urlverify_str_arg)
VisionDatasetc
^ \ rS rSrSr/ SQ/ SQ/ SQS.r SS\\\4 S \S
\ \
S \ \
S \ S
S4 U 4S jjjr S\
S
\\\4 4S jrS
\
4S jrS
\ 4S jrSS jrS
\4S jrSrU =r$ )SVHN aD `SVHN `_ Dataset.
Note: The SVHN dataset assigns the label `10` to the digit `0`. However, in this Dataset,
we assign the label `0` to the digit `0` to be compatible with PyTorch loss functions which
expect the class labels to be in the range `[0, C-1]`
.. warning::
This class needs `scipy `_ to load data from `.mat` format.
Args:
root (str or ``pathlib.Path``): Root directory of the dataset where the data is stored.
split (string): One of {'train', 'test', 'extra'}.
Accordingly dataset is selected. 'extra' is Extra training set.
transform (callable, optional): A function/transform that takes in a PIL image
and returns a transformed version. E.g, ``transforms.RandomCrop``
target_transform (callable, optional): A function/transform that takes in the
target and transforms it.
download (bool, optional): If true, downloads the dataset from the internet and
puts it in root directory. If dataset is already downloaded, it is not
downloaded again.
)z6http://ufldl.stanford.edu/housenumbers/train_32x32.matztrain_32x32.mat e26dedcc434d2e4c54c9b2d4a06d8373)z5http://ufldl.stanford.edu/housenumbers/test_32x32.matztest_32x32.mat eb5a983be6a315427106f1b164d9cef3)z6http://ufldl.stanford.edu/housenumbers/extra_32x32.matzextra_32x32.mat a93ce644f1a588dc4d68dda5feec44a7)traintestextraNrootsplit transformtarget_transformdownloadreturnc : > [ TU ] XUS9 [ US[ U R R 5 5 5 U l U R U S U l U R U S U l U R U S U l U( a U R 5 U R 5 ( d [ S5 eSS K
Jn UR [ R" R% U R& U R 5 5 nUS U l US R+ [, R. 5 R1 5 U l [, R4 " U R2 U R2 S :H S5 [, R6 " U R( S
5 U l g ) N)r r r r r zHDataset not found or corrupted. You can use download=True to download itXy
) r r r ) super__init__r tuple
split_listkeysr urlfilenamefile_md5r _check_integrity RuntimeErrorscipy.ioioloadmatospathjoinr dataastypenpint64squeezelabelsplace transpose) selfr r r r r sio
loaded_mat __class__s V/home/mande/repo/quber/.venv/lib/python3.13/site-packages/torchvision/datasets/svhn.pyr$
SVHN.__init__6 s1 EUV#E7E$//:N:N:P4QR
??5)!, .q1
.q1
MMO$$&& ij j [[ dii !GH
sO
!o,,RXX6>>@
dkkR /3LLL9 indexc " U R U [ U R U 5 p2[ R " [
R " US5 5 nU R b U R U5 nU R b U R U5 nX#4$ )zn
Args:
index (int): Index
Returns:
tuple: (image, target) where target is index of the target class.
)r r r ) r3 intr8 r fromarrayr5 r: r r )r; rB imgtargets r? __getitem__SVHN.__getitem__^ sy ii&DKK,>(?V oobll3 :; >> %..%C ,**62F{rA c , [ U R 5 $ )N)lenr3 r; s r? __len__ SVHN.__len__t s 499~ rA c U R nU R U R S n[ R R XR 5 n[ X25 $ Nr )r r& r r0 r1 r2 r) r
)r; r md5fpaths r? r+ SVHN._check_integrityw sC yyoodjj)!, T==1 u**rA c U R U R S n[ U R U R U R
U5 g rP )r& r r r( r r) )r; rQ s r? r
SVHN.download} s3 oodjj)!,TXXtyy$--=rA c : SR " S0 U R D6$ )NzSplit: {split} )format__dict__rL s r?
extra_reprSVHN.extra_repr s &&7 77rA )r3 r* r) r8 r r( )r NNF)r N)__name__
__module__ __qualname____firstlineno____doc__r& r strr r r boolr$ rD r% r rH rM r+ r rZ __static_attributes__
__classcell__)r> s @r? r r s 0
J* (,/3
&:CI &: &: H% &:
#8, &:
&:
&: &:P sCx , +$ + >8C 8 8rA r )os.pathr0 pathlibr typingr r r r numpyr5 PILr utilsr
r r visionr
r rW rA r? rl s. 1 1 @ @ !v8= v8rA