+
jR 4 ^ RI t ^ RIt^ RIHtHt ^ RIHt ^ RIHt ^ RIH t H
t
H t H t ^ RI
t^ RIt^ RIHt ^RIHtHt ^RIHt ^R IHtHt ^R
IHt ] ]P" ]P" ] ]P: , ] ]P: , 3, t ] ]P" ]P" ] ]P: , 3, t Rt ! R R]]4 t! ! R R ]!4 t" ! R R ]!4 t# ! R R]!4 t$ ! R R
]!4 t% ! R R]!4 t&R R lt'R R lt(R# ) N)ABCabstractmethod)globPath)AnyCallableOptionalUnion)Image)
decode_png read_file)default_loader) _read_pfmverify_str_arg)
VisionDataset KittiFlowSintelFlyingThings3D FlyingChairsHD1Kc a a ] tR t^ t oRtR]3V3R lV 3R llltV3R lR lt]V3R lR l4 t V3R lR
lt
V3R lR lt V3R
lR lt Rt
VtV ;t# ) FlowDatasetFNc n < V ^8 d QhRS[ S[S[3, RS[S[, RS[S[.S[3, RR/# ) root
transformsloaderreturnNr strr r
r r )format
__classdict__s "c/home/luca/screen-checking/.venv/lib/python3.14/site-packages/torchvision/datasets/_optical_flow.py __annotate__FlowDataset.__annotate__$ sL CI X& #$
c \ < \ SV ` VR 7 W n . V n . V n W0n R# ))r N)super__init__r
_flow_list _image_list_loader)selfr r r __class__s &&&&r$ r* FlowDataset.__init__$ s. d#$%',. r' c d < V ^8 d QhRS[ RS[S[P S[P 3, /# r file_namer )r! r r torchTensor)r" r# s "r$ r% r& 2 s- ' '3 '5ell1J+K 'r' c $ V P V4 # N)r- r. r3 s &&r$ _read_imgFlowDataset._read_img2 s ||I&&r' c < V ^8 d QhRS[ /# )r r3 )r! )r" r# s "r$ r% r& 6 s
C
r' c R # r7 r8 s &&r$
_read_flowFlowDataset._read_flow5 s
r' c < < V ^8 d QhRS[ RS[S[S[3, /# r indexr intr T1T2)r" r# s "r$ r% r& : s# $ $ $r2v $r' c V P V P V, ^ , 4 pV P V P V, ^, 4 pV P ' d= V P V P V, 4 pV P ' d Vw rEMRpMR;rEV P
e V P W#WE4 w r#rEV P ' g Ve W#WE3# W#V3# )r N)r9 r, r+ r> _has_builtin_flow_maskr )r. rB img1img2flowvalid_flow_masks && r$ __getitem__FlowDataset.__getitem__: s ~~d . .u 5a 89~~d . .u 5a 89 ?????4??5#9:D***(,%o"&%) )D ?? &04D0b -D & & &/*Et4 4t# #r' c < V ^8 d QhRS[ /# )r r )rD )r" r# s "r$ r% r& Q s % % %r' c , \ V P 4 # r7 )lenr, )r. s &r$ __len__FlowDataset.__len__Q s 4##$$r' c b < V ^8 d QhRS[ RS[P P P /# )r vr )rD r4 utilsdata
ConcatDataset)r" r# s "r$ r% r& T s+ : :# :%++"2"2"@"@ :r' c d \ P P P V .V, 4 # r7 )r4 rV rW rX )r. rU s &&r$ __rmul__FlowDataset.__rmul__T s# {{ --tfqj99r' )r+ r, r- r )__name__
__module__ __qualname____firstlineno__rH r r* r9 r r> rM rR rZ __static_attributes____classdictcell__
__classcell__r/ r# s @@r$ r r s` #
*.'5 ' '
$ $.% %: : :r' r c x a a ] tR t^Xt oRtRRR]3V3R lV 3R llltV3R lV 3R lltV3R lR
ltR t Vt
V ;t # ) r a: `Sintel `_ Dataset for optical flow.
The dataset is expected to have the following structure: ::
root
Sintel
testing
clean
scene_1
scene_2
...
final
scene_1
scene_2
...
training
clean
scene_1
scene_2
...
final
scene_1
scene_2
...
flow
scene_1
scene_2
...
Args:
root (str or ``pathlib.Path``): Root directory of the Sintel Dataset.
split (string, optional): The dataset split, either "train" (default) or "test"
pass_name (string, optional): The pass to use, either "clean" (default), "final", or "both". See link above for
details on the different passes.
transforms (callable, optional): A function/transform that takes in
``img1, img2, flow, valid_flow_mask`` and returns a transformed version.
``valid_flow_mask`` is expected for consistency with other datasets which
return a built-in valid mask, such as :class:`~torchvision.datasets.KittiFlow`.
loader (callable, optional): A function to load an image given its path.
By default, it uses PIL as its image loader, but users could also pass in
``torchvision.io.decode_image`` for decoding image data into tensors directly.
traincleanNc z < V ^8 d QhRS[ S[S[3, RS[RS[RS[S[, RS[S[.S[3, RR/# )r r split pass_namer r r Nr )r" r# s "r$ r% Sintel.__annotate__ sh V VCI V V V
X& V #$
V
Vr' c < \ S
V ` WVR 7 \ VRRR7 \ VRRR7 VR8X d RR.MV.p\ V4 R, pVR , R
, pV EF pVR8X d R MTpW, V, p \ P
! V 4 F p
\
\ \ W, R , 4 4 4 p \ \ V 4 ^,
4 F2 p V ;P W, W^, , ..,
un K4 VR8X g K V ;P \
\ \ Wz, R , 4 4 4 ,
un K EK R
# )r r r rh re valid_valuesri rf finalbothr trainingrK *.png*.floNre testrf ro rp )
r) r* r r oslistdirsortedr r! rangerQ r, r+ )r. r rh ri r r passes flow_root split_dir
image_rootscene
image_listir/ s &&&&&& r$ r* Sintel.__init__ s d&Iug4EFy+ Sintel._read_flow
##r' r= r\ r] r^ r_ __doc__r r* rM r> r` ra rb rc s @@r$ r r X s> )\ )-'5
V V8
*
* $ $ $r' c z a a ] tR t^t oRtRtRR]3V3R lV 3R llltV3R lV 3R lltV3R lR
lt R t
Vt V ;t # ) r aw `KITTI `__ dataset for optical flow (2015).
The dataset is expected to have the following structure: ::
root
KittiFlow
testing
image_2
training
image_2
flow_occ
Args:
root (str or ``pathlib.Path``): Root directory of the KittiFlow Dataset.
split (string, optional): The dataset split, either "train" (default) or "test"
transforms (callable, optional): A function/transform that takes in
``img1, img2, flow, valid_flow_mask`` and returns a transformed version.
loader (callable, optional): A function to load an image given its path.
By default, it uses PIL as its image loader, but users could also pass in
``torchvision.io.decode_image`` for decoding image data into tensors directly.
Tre Nc
t < V ^8 d QhRS[ S[S[3, RS[RS[S[, RS[S[.S[3, RR/# r r rh r r r Nr )r" r# s "r$ r% KittiFlow.__annotate__ s] P PCI P P X& P
#$ P
Pr' c 8 < \ S V ` WVR 7 \ VRR R7 \ V4 R, VR, , p\ \ \
VR, R, 4 4 4 p\ \ \
VR, R, 4 4 4 pV' d V' g \ R 4 h\ WV4 F w rxV ;P Wx..,
un K" VR8X d3 \ \ \
VR
, R, 4 4 4 V n
R # R # )
rl rh re rm r ingimage_2z*_10.pngz*_11.pngzZCould not find the Kitti flow images. Please make sure the directory structure is correct.flow_occNrt ) r) r* r r ry r r! FileNotFoundErrorzipr, r+ )
r. r rh r r images1images2rI rJ r/ s
&&&&& r$ r* KittiFlow.__init__ s d&Iug4EFDzK'55=9c$"2Z"? @ABc$"2Z"? @ABg#l
g/JD $ . 0 G $T#dZ.?*.L*M%N ODO r' c < < V ^8 d QhRS[ RS[S[S[3, /# rA rC )r" r# s "r$ r% r r r' c " < \ SV ` V4 # )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"``.
r r s &&r$ rM KittiFlow.__getitem__ r r' c d < V ^8 d QhRS[ RS[S[P S[P 3, /# r2 r! tupler r )r" r# s "r$ r% r 1 D DC DE"**bjj2H,I Dr' c \ V4 # r7 )_read_16bits_png_with_flow_and_valid_maskr8 s &&r$ r> KittiFlow._read_flow
8CCr' )r+ )
r\ r] r^ r_ r rH r r* rM r> r` ra rb rc s @@r$ r r sE , "
)-'5 P P4
*
* D D Dr' c p a a ] tR t^t oRtR V3R lV 3R llltV3R lV 3R lltV3R lR ltRtVt V ;t
# )
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`.
c V < V ^8 d QhRS[ S[S[3, RS[RS[S[, RR/# )r r rh r r N)r r! r r
r )r" r# s "r$ r% FlyingChairs.__annotate__ s@ I IU39 - Ic IQYZbQc Ios Ir' c 2 < \ S
V ` WR 7 \ VRR
R7 \ V4 R, p\ \ \
VR, R, 4 4 4 p\ \ \
VR, R, 4 4 4 pR p\ P P W, 4 ' g \ R
4 h\ P ! \
W, 4 \ P R 7 p\ \ V4 4 F pWx, p VR8X d V ^8X g VR8X g K" V ^8X g K+ V ;P WX, .,
un V ;P" V^V, , V^V, ^, , ..,
un K R # ))r r rh re valrm r rW z*.ppmrs zFlyingChairs_train_val.txtzmThe FlyingChairs_train_val.txt file was not found - please download it from the dataset page (see docstring).)dtypeN)re r )r) r* r r ry r r! rw pathexistsr r loadtxtint32rz rQ r+ r, ) r. r rh r imagesflowssplit_file_name
split_listr 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' c < < V ^8 d QhRS[ RS[S[S[3, /# rA rC )r" r# s "r$ r% r * r r' c " < \ SV ` V4 # )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.
r r s &&r$ rM FlyingChairs.__getitem__* r r' c : < V ^8 d QhRS[ RS[P /# r2 r )r" r# s "r$ r% r 9 r r' c \ V4 # r7 r r8 s &&r$ r> FlyingChairs._read_flow9 r r' r= )re N) r\ r] r^ r_ r r* rM r> r` ra rb rc s @@r$ r r s/ 2I I.
*
* $ $ $r' c z a a ] tR tRt oRtRRRR]3V3R lV 3R llltV3R lV 3R
lltV3R lR ltR
t Vt
V ;t # )r i= a `FlyingThings3D `_ dataset for optical flow.
The dataset is expected to have the following structure: ::
root
FlyingThings3D
frames_cleanpass
TEST
TRAIN
frames_finalpass
TEST
TRAIN
optical_flow
TEST
TRAIN
Args:
root (str or ``pathlib.Path``): Root directory of the intel FlyingThings3D Dataset.
split (string, optional): The dataset split, either "train" (default) or "test"
pass_name (string, optional): The pass to use, either "clean" (default) or "final" or "both". See link above for
details on the different passes.
camera (string, optional): Which camera to return images from. Can be either "left" (default) or "right" or "both".
transforms (callable, optional): A function/transform that takes in
``img1, img2, flow, valid_flow_mask`` and returns a transformed version.
``valid_flow_mask`` is expected for consistency with other datasets which
return a built-in valid mask, such as :class:`~torchvision.datasets.KittiFlow`.
loader (callable, optional): A function to load an image given its path.
By default, it uses PIL as its image loader, but users could also pass in
``torchvision.io.decode_image`` for decoding image data into tensors directly.
re rf leftNc < V ^8 d QhRS[ S[S[3, RS[RS[RS[RS[S[, RS[S[.S[3, RR/# ) r r rh ri camerar r r Nr )r" r# s "r$ r% FlyingThings3D.__annotate__] sj 1: 1:CI 1: 1: 1:
1: X&
1: #$1:
1:r' c
<aa \ SV ` WVR 7 \ VRRR7 VP 4 p\ VRRR7 RR.RR.RRR./V, p\ SR RR7 SR8X d R
R .MS.p\ V4 R , pRp \
P ! WxV 4 EF w poo\ \ \ W, V, R, 4 4 4 p
\ V3R lV
4 4 p
\ \ \ VR, V, R, 4 4 4 p \ VV3R lV 4 4 p V
' d V ' g \ R4 h\ W4 EF- w r\ \ \ V R, 4 4 4 p\ \ \ V
R, 4 4 4 p\ \ V4 ^,
4 F pSR
8X dU V ;P VV, VV^, , ..,
un V ;P VV, .,
un K^ SR8X g Kg V ;P VV^, , VV, ..,
un V ;P VV^, , .,
un K EK0 EK R# )rl rh rm ri rf ro rp frames_cleanpassframes_finalpassr r rightr into_future into_pastz*/*c 3 H <" T F p\ V4 S, x K R # 5ir7 r ).0 image_dirr s & r$ *FlyingThings3D.__init__..z s U*YY& 8 8*s " optical_flowc 3 V <" T F p\ V4 S, S, x K R # 5ir7 r )r flow_dirr directions & r$ r r } s$ ]S\xtH~ 9F B BS\s &)zcCould not find the FlyingThings3D flow images. Please make sure the directory structure is correct.rr z*.pfmNrt rv )r r rp )r r )r) r* r upperr itertoolsproductry r r! r r rz rQ r, r+ )r. r rh ri r r r r{ cameras
directions
image_dirs flow_dirsr r r r r r r/ s &&&&f&& @r$ r* FlyingThings3D.__init__] s# d&Iug4EF
y+? tC7(:$; <= s5zA~.A M1 ((fQiA -G,HH( E!H:5"k1 ((fQUmVAY-G,HH( E!a%L>9
/ (B -[r' c < < V ^8 d QhRS[ RS[S[S[3, /# rA rC )r" r# s "r$ r% r r r' c " < \ SV ` V4 # r r r s &&r$ rM FlyingThings3D.__getitem__ r r' c : < V ^8 d QhRS[ RS[P /# r2 r )r" r# s "r$ r% r r r' c \ V4 # r7 )r r8 s &&r$ r> FlyingThings3D._read_flow r r' r= r rc s @@r$ r r = s@ D )-'51: 1:f
*
* $ $ $r' c z a a ] tR tRt oRtRtRR]3V3R lV 3R llltV3R lR ltV3R
lV 3R llt R t
Vt V ;t # )
r i ak `HD1K `__ dataset for optical flow.
The dataset is expected to have the following structure: ::
root
hd1k
hd1k_challenge
image_2
hd1k_flow_gt
flow_occ
hd1k_input
image_2
Args:
root (str or ``pathlib.Path``): Root directory of the HD1K Dataset.
split (string, optional): The dataset split, either "train" (default) or "test"
transforms (callable, optional): A function/transform that takes in
``img1, img2, flow, valid_flow_mask`` and returns a transformed version.
loader (callable, optional): A function to load an image given its path.
By default, it uses PIL as its image loader, but users could also pass in
``torchvision.io.decode_image`` for decoding image data into tensors directly.
Tre Nc
t < V ^8 d QhRS[ S[S[3, RS[RS[S[, RS[S[.S[3, RR/# r r )r" r# s "r$ r% HD1K.__annotate__ sV CI X&
#$
r' c
< \ S
V ` WVR 7 \ VRRR7 \ V4 R, pVR8X d \ ^$4 F p\ \
\ VR, R, VR
R2, 4 4 4 p\ \
\ VR , R
, VR
R2, 4 4 4 p\ \ V4 ^,
4 FR pV ;P Wh, .,
un V ;P Wx, Wx^, , ..,
un
KT K M\ \
\ VR , R
, R , 4 4 4 p \ \
\ VR , R
, R
, 4 4 4 p
\ W4 F w rV ;P W..,
un
K" V P ' g \ R4 hR# )rl rh re rm hd1k hd1k_flow_gtr 06dz_*.png
hd1k_inputr hd1k_challengez*10.pngz*11.pngzTCould not find the HD1K images. Please make sure the directory structure is correct.Nrt )
r) r* r r rz ry r r! rQ r+ r, r r )r. r rh r r seq_idxr r r r r image1image2r/ s &&&&& r$ r*
HD1K.__init__ s d&Iug4EFDzF" G 9 tC~(=
(JPWX[}\bMc(c$d ef S ) HD1K._read_flow r r' c < < V ^8 d QhRS[ RS[S[S[3, /# rA rC )r" r# s "r$ r% r r r' c " < \ SV ` V4 # )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"``.
r r s &&r$ rM HD1K.__getitem__ r r' r= )
r\ r] r^ r_ r rH r r* r> rM r` ra rb rc s @@r$ r r sB . "
)-'5 >D D
*
*
*r' c D V ^8 d QhR\ R\ P /# r2 r )r" s "r$ r% r% s 8 8 8 8r' c
\ V R4 ;_uu_ 4 p\ P ! VR^R7 P 4 pVR8w d \ R4 h\ P ! VR^R7 P 4 p\ P ! VR^R7 P 4 p\ P ! VR^V, V, R7 pVP
WC^4 P ^^ ^4 uuRRR4 # + ' g i R# ; i) 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"1r'