EiR \ S SK r S SKrS SKJrJr S SKJr S SKJr S SKJ r J
r
J r J r S SK
rS SKrS SKJr SSKJrJr S S
KJr S S KJrJr S S KJr \ \R" \R" \ \R: \ \R: 4 r \ \R" \R" \ \R: 4 r S
r " S S\\5 r! " S S\!5 r" " S S\!5 r# " S S\!5 r$ " S S\!5 r% " S S\!5 r&S\'S\R: 4S jr(S\'S\ \R: \R: 4 4S jr)g) N)ABCabstractmethod)globPath)AnyCallableOptionalUnion)Image )
decode_png read_file )default_loader) _read_pfmverify_str_arg)
VisionDataset) KittiFlowSintelFlyingThings3D FlyingChairsHD1Kc 8 ^ \ rS rSrSrS\4S\\\4 S\ \
S\
\/\ 4 SS4U 4S jjjr S \S\\
R \R 4 4S
jr\S \4S j5 rS \S\\\4 4S
jrS\4S jrS\S\R0 R2 R4 4S jrSr U =r $ ) FlowDataset FNroot
transformsloaderreturnc T > [ TU ] US9 X l / U l / U l X0l g )N)r )super__init__r
_flow_list _image_list_loader)selfr r r __class__s _/home/mande/repo/quber/.venv/lib/python3.13/site-packages/torchvision/datasets/_optical_flow.pyr# FlowDataset.__init__$ s. d#$%',. file_namec $ U R U5 $ N)r& r' r, s r) _read_imgFlowDataset._read_img2 s ||I&&r+ c g r. r/ s r)
_read_flowFlowDataset._read_flow5 s
r+ indexc U R U R U S 5 nU R U R U S 5 nU R ( a7 U R U R U 5 nU R ( a Uu pEOS nOS =pEU R
b U R X#XE5 u p#pEU R ( d Ub X#XE4$ X#U4$ )Nr r )r0 r% r$ r4 _has_builtin_flow_maskr )r' r6 img1img2flowvalid_flow_masks r) __getitem__FlowDataset.__getitem__: s ~~d . .u 5a 89~~d . .u 5a 89 ????4??5#9:D**(,%o"&%) )D ?? &04D0b -D & &/*Et4 4t# #r+ c , [ U R 5 $ r. )lenr% )r' s r) __len__FlowDataset.__len__Q s 4##$$r+ vc \ [ R R R U /U- 5 $ r. )torchutilsdata
ConcatDataset)r' rC s r) __rmul__FlowDataset.__rmul__T s# {{ --tfqj99r+ )r$ r% r& r ) __name__
__module__ __qualname____firstlineno__r8 r r strr r
r r r# r rE Tensorr0 r r4 intT1T2r= rA rF rG rH rI __static_attributes__
__classcell__r( s @r) r r s #
*.'5 CI X& #$
'3 '5ell1J+K '
C
$ $r2v $.% %:# :%++"2"2"@"@ : :r+ r c
^ \ rS rSrSrSSS\4S\\\4 S\S\S \ \
S
\
\/\ 4 S S4 U 4S jjjr S
\
S \\\4 4U 4S jjrS\S \R$ 4S jrSrU =r$ )r X a: `Sintel `_ Dataset for optical flow.
The dataset is expected to have the following structure: ::
root
Sintel
testing
clean
scene_1
scene_2
...
final
scene_1
scene_2
...
training
clean
scene_1
scene_2
...
final
scene_1
scene_2
...
flow
scene_1
scene_2
...
Args:
root (str or ``pathlib.Path``): Root directory of the Sintel Dataset.
split (string, optional): The dataset split, either "train" (default) or "test"
pass_name (string, optional): The pass to use, either "clean" (default), "final", or "both". See link above for
details on the different passes.
transforms (callable, optional): A function/transform that takes in
``img1, img2, flow, valid_flow_mask`` and returns a transformed version.
``valid_flow_mask`` is expected for consistency with other datasets which
return a built-in valid mask, such as :class:`~torchvision.datasets.KittiFlow`.
loader (callable, optional): A function to load an image given its path.
By default, it uses PIL as its image loader, but users could also pass in
``torchvision.io.decode_image`` for decoding image data into tensors directly.
traincleanNr split pass_namer r r c : > [ T
U ] XUS9 [ USSS9 [ USSS9 US:X a SS /OU/n[ U5 S
- nUS - S - nU H nUS
:X a S OUnX- U- n [ R
" U 5 H n
[
[ [ X- S- 5 5 5 n [ [ U 5 S-
5 H" n U =R X XS- //-
sl M$ US
:X d Mj U =R [
[ [ Xz- S- 5 5 5 -
sl M M g )Nr r r r[ rY test valid_valuesr\ rZ finalbothre rZ rd r trainingr; rY *.pngr *.flo)
r" r# r r oslistdirsortedr rO ranger@ r% r$ )r' r r[ r\ r r passes flow_root split_dir
image_rootscene
image_listir( s r) r# Sintel.__init__ s d&Iug4EFy+ [ TU ] U5 $ a Return example at given index.
Args:
index(int): The index of the example to retrieve
Returns:
tuple: A 3-tuple with ``(img1, img2, flow)``.
The flow is a numpy array of shape (2, H, W) and the images are PIL images.
``flow`` is None if ``split="test"``.
If a valid flow mask is generated within the ``transforms`` parameter,
a 4-tuple with ``(img1, img2, flow, valid_flow_mask)`` is returned.
r" r= r' r6 r( s r) r= Sintel.__getitem__ w"5))r+ r, c [ U5 $ r. _read_flor/ s r) r4 Sintel._read_flow
##r+ r3 rK rL rM rN __doc__r r rO r r
r r r# rQ rR rS r= npndarrayr4 rT rU rV s @r) r r X s )\ )-'5
VCI V V V
X& V #$
V
V V8
*
*r2v
* $C $BJJ $ $r+ r c ^ \ rS rSrSrSrSS\4S\\\ 4 S\S\
\ S \ \/\ 4 S
S4
U 4S jjjr
S \S
\\\4 4U 4S
jjrS\S
\\R( \R( 4 4S jrSrU =r$ )r aw `KITTI `__ dataset for optical flow (2015).
The dataset is expected to have the following structure: ::
root
KittiFlow
testing
image_2
training
image_2
flow_occ
Args:
root (str or ``pathlib.Path``): Root directory of the KittiFlow Dataset.
split (string, optional): The dataset split, either "train" (default) or "test"
transforms (callable, optional): A function/transform that takes in
``img1, img2, flow, valid_flow_mask`` and returns a transformed version.
loader (callable, optional): A function to load an image given its path.
By default, it uses PIL as its image loader, but users could also pass in
``torchvision.io.decode_image`` for decoding image data into tensors directly.
TrY Nr r[ r r r c > [ T U ] XUS9 [ USSS9 [ U5 S- US- - n[ [ [
US- S- 5 5 5 n[ [ [
US- S - 5 5 5 nU( a U( d [ S
5 e[ XV5 H u pxU =R Xx//-
sl M US :X a) [ [ [
US - S- 5 5 5 U l
g g )
Nr^ r[ r_ ra r ingimage_2z*_10.pngz*_11.pngzZCould not find the Kitti flow images. Please make sure the directory structure is correct.rY flow_occ) r" r# r r rk r rO FileNotFoundErrorzipr% r$ )
r' r r[ r r images1images2r9 r: r( s
r) r# KittiFlow.__init__ s d&Iug4EFDzK'55=9c$"2Z"? @ABc$"2Z"? @ABg#l
g/JD $ . 0 G $T#dZ.?*.L*M%N ODO r+ r6 c " > [ TU ] U5 $ )a Return example at given index.
Args:
index(int): The index of the example to retrieve
Returns:
tuple: A 4-tuple with ``(img1, img2, flow, valid_flow_mask)``
where ``valid_flow_mask`` is a numpy boolean mask of shape (H, W)
indicating which flow values are valid. The flow is a numpy array of
shape (2, H, W) and the images are PIL images. ``flow`` and ``valid_flow_mask`` are None if
``split="test"``.
rw rx s r) r= KittiFlow.__getitem__ rz r+ r, c [ U5 $ r. )_read_16bits_png_with_flow_and_valid_maskr/ s r) r4 KittiFlow._read_flow
8CCr+ )r$ )rK rL rM rN r r8 r r rO r r
r r r# rQ rR rS r= tupler r r4 rT rU rV s @r) r r s , "
)-'5 PCI P P X& P
#$ P
P P4
*
*r2v
* DC DE"**bjj2H,I D Dr+ r c ^ \ rS rSrSrSS\\\4 S\S\\ SS4U 4S jjjr
S \ S\\ \
4 4U 4S
jjrS \S\R 4S jrS
rU =r$ )r a `FlyingChairs `_ Dataset for optical flow.
You will also need to download the FlyingChairs_train_val.txt file from the dataset page.
The dataset is expected to have the following structure: ::
root
FlyingChairs
data
00001_flow.flo
00001_img1.ppm
00001_img2.ppm
...
FlyingChairs_train_val.txt
Args:
root (str or ``pathlib.Path``): Root directory of the FlyingChairs Dataset.
split (string, optional): The dataset split, either "train" (default) or "val"
transforms (callable, optional): A function/transform that takes in
``img1, img2, flow, valid_flow_mask`` and returns a transformed version.
``valid_flow_mask`` is expected for consistency with other datasets which
return a built-in valid mask, such as :class:`~torchvision.datasets.KittiFlow`.
Nr r[ r r c > [ T
U ] XS9 [ USSS9 [ U5 S- n[ [ [
US- S- 5 5 5 n[ [ [
US- S- 5 5 5 nS n[ R R X- 5 ( d [ S
5 e[ R " [
X- 5 [ R S 9n[ [ U5 5 Hb nXx n US :X a U S
:X d US:X d M U S:X d M# U =R XX /-
sl U =R" USU- USU- S
- //-
sl Md g )N)r r r[ )rY valra r rG z*.ppmrh zFlyingChairs_train_val.txtzmThe FlyingChairs_train_val.txt file was not found - please download it from the dataset page (see docstring).)dtyperY r r r
)r" r# r r rk r rO ri pathexistsr r loadtxtint32rl r@ r$ r% ) r' r r[ r imagesflowssplit_file_name
split_listrs split_idr( s r) r# FlyingChairs.__init__ s&
d:ug4DEDzN*S !8 9:;tCv
7 89:6ww~~d 455#
ZZD$: ;288L
s5z"A!}H X]8WX=EH:- fQUmVAEAI5F%G$HH #r+ r6 c " > [ TU ] U5 $ )a Return example at given index.
Args:
index(int): The index of the example to retrieve
Returns:
tuple: A 3-tuple with ``(img1, img2, flow)``.
The flow is a numpy array of shape (2, H, W) and the images are PIL images.
``flow`` is None if ``split="val"``.
If a valid flow mask is generated within the ``transforms`` parameter,
a 4-tuple with ``(img1, img2, flow, valid_flow_mask)`` is returned.
rw rx s r) r= FlyingChairs.__getitem__* rz r+ r, c [ U5 $ r. r| r/ s r) r4 FlyingChairs._read_flow9 r r+ r3 )rY N)rK rL rM rN r r rO r r
r r# rQ rR rS r= r r r4 rT rU rV s @r) r r s| 2IU39 - Ic IQYZbQc Ios I I.
*
*r2v
* $C $BJJ $ $r+ r c ^ \ rS rSrSrSSSS\4S\\\4 S\S \S
\S \ \
S \
\/\ 4 S
S4U 4S jjjr S\
S
\\\4 4U 4S jjrS\S
\R$ 4S jrSrU =r$ )r i= a `FlyingThings3D `_ dataset for optical flow.
The dataset is expected to have the following structure: ::
root
FlyingThings3D
frames_cleanpass
TEST
TRAIN
frames_finalpass
TEST
TRAIN
optical_flow
TEST
TRAIN
Args:
root (str or ``pathlib.Path``): Root directory of the intel FlyingThings3D Dataset.
split (string, optional): The dataset split, either "train" (default) or "test"
pass_name (string, optional): The pass to use, either "clean" (default) or "final" or "both". See link above for
details on the different passes.
camera (string, optional): Which camera to return images from. Can be either "left" (default) or "right" or "both".
transforms (callable, optional): A function/transform that takes in
``img1, img2, flow, valid_flow_mask`` and returns a transformed version.
``valid_flow_mask`` is expected for consistency with other datasets which
return a built-in valid mask, such as :class:`~torchvision.datasets.KittiFlow`.
loader (callable, optional): A function to load an image given its path.
By default, it uses PIL as its image loader, but users could also pass in
``torchvision.io.decode_image`` for decoding image data into tensors directly.
rY rZ leftNr r[ r\ camerar r r c
4 >^^ [ TU ] XUS9 [ USSS9 UR 5 n[ USSS9 S/S/SS/S.U n[ TS S
S9 TS :X a S S
/OT/n[ U5 S- nSn [
R " XxU 5 GH u nmm[ [ [ X- U- S- 5 5 5 n
[ U4S jU
5 5 n
[ [ [ US- U- S- 5 5 5 n [ UU4S jU 5 5 n U
( a U ( d [ S5 e[ X5 H u p[ [ [ U S- 5 5 5 n[ [ [ U
S- 5 5 5 n[ [ U5 S-
5 H nTS:X a< U =R UU UUS- //-
sl U =R UU /-
sl ME TS:X d MM U =R UUS- UU //-
sl U =R UUS- /-
sl M M GM g )Nr^ r[ r_ ra r\ rc frames_cleanpassframes_finalpassr )r rightre re r r r ) into_future into_pastz*/*c 3 @ ># U H n[ U5 T- v M g 7fr. r ).0 image_dirr s r) *FlyingThings3D.__init__..z s U*YY& 8*s optical_flowc 3 F ># U H n[ U5 T- T- v M g 7fr. r )r flow_dirr directions r) r r } s! ]S\xtH~ 9F BS\s !zcCould not find the FlyingThings3D flow images. Please make sure the directory structure is correct.rg z*.pfmr r r )r" r# r upperr itertoolsproductrk r rO r r rl r@ r% r$ )r' r r[ r\ r r r rm cameras
directions
image_dirs flow_dirsr r r r rs r r( s ` @r) r# FlyingThings3D.__init__] s d&Iug4EF
y+? tC7(:$; <= s5zA~.A M1 ((fQiA -G,HH( E!H:5"k1 ((fQUmVAY-G,HH( E!a%L>9
/ (B -[r+ r6 c " > [ TU ] U5 $ rv rw rx s r) r= FlyingThings3D.__getitem__ rz r+ r, c [ U5 $ r. )r r/ s r) r4 FlyingThings3D._read_flow r r+ r3 r rV s @r) r r = s D )-'51:CI 1: 1: 1:
1: X&
1: #$1:
1: 1:f
*
*r2v
* $C $BJJ $ $r+ r c ^ \ rS rSrSrSrSS\4S\\\ 4 S\S\
\ S \ \/\ 4 S
S4
U 4S jjjr
S \S
\\R \R 4 4S
jrS\S
\\\4 4U 4S jjrSrU =r$ )r i ak `HD1K `__ dataset for optical flow.
The dataset is expected to have the following structure: ::
root
hd1k
hd1k_challenge
image_2
hd1k_flow_gt
flow_occ
hd1k_input
image_2
Args:
root (str or ``pathlib.Path``): Root directory of the HD1K Dataset.
split (string, optional): The dataset split, either "train" (default) or "test"
transforms (callable, optional): A function/transform that takes in
``img1, img2, flow, valid_flow_mask`` and returns a transformed version.
loader (callable, optional): A function to load an image given its path.
By default, it uses PIL as its image loader, but users could also pass in
``torchvision.io.decode_image`` for decoding image data into tensors directly.
TrY Nr r[ r r r c
> [ T
U ] XUS9 [ USSS9 [ U5 S- nUS:X a [ S5 H n[ [
[ US- S - US
S 3- 5 5 5 n[ [
[ US - S
- US
S 3- 5 5 5 n[ [ U5 S-
5 H: nU =R Xh /-
sl U =R Xx XxS- //-
sl
M< M Ow[ [
[ US- S
- S- 5 5 5 n [ [
[ US- S
- S- 5 5 5 n
[ X5 H u pU =R X//-
sl
M U R ( d [ S5 eg )Nr^ r[ r_ ra hd1krY $ hd1k_flow_gtr 06dz_*.png
hd1k_inputr r hd1k_challengez*10.pngz*11.pngzTCould not find the HD1K images. Please make sure the directory structure is correct.)
r" r# r r rl rk r rO r@ r$ r% r r )r' r r[ r r seq_idxr r rs r r image1image2r( s r) r#
HD1K.__init__ s d&Iug4EFDzF" G 9 tC~(=
(JPWX[}\bMc(c$d ef S ) [ TU ] U5 $ )a Return example at given index.
Args:
index(int): The index of the example to retrieve
Returns:
tuple: A 4-tuple with ``(img1, img2, flow, valid_flow_mask)`` where ``valid_flow_mask``
is a numpy boolean mask of shape (H, W)
indicating which flow values are valid. The flow is a numpy array of
shape (2, H, W) and the images are PIL images. ``flow`` and ``valid_flow_mask`` are None if
``split="test"``.
rw rx s r) r= HD1K.__getitem__ rz r+ r3 )rK rL rM rN r r8 r r rO r r
r r r# r r r r4 rQ rR rS r= rT rU rV s @r) r r s . "
)-'5 CI X&
#$
>DC DE"**bjj2H,I D
*
*r2v
*
*r+ r r, r c [ U S5 n[ R " USSS9R 5 nUS:w a [ S5 e[ R " USSS9R 5 n[ R " USSS9R 5 n[ R " US S
U- U- S9nUR
XCS
5 R S
S S5 sS S S 5 $ ! , ( d f g = f)
z#Read .flo file in Middlebury formatrbc )counts PIEHz)Magic number incorrect. Invalid .flo filezr s # 1 1 , " , !
5;; Xbjj%98BJJ;O OP
5;; Xbjj%9 9: 7:#} 7:tX$[ X$vCD CDLA$; A$Hc$[ c$LI*; I*X8 8 8"1 1rzzSUS]S]G]A^ 1r+