+ jml^RIt^RIHt^RIHt^RIt^RIHt^RIHt H t .ROt R R lt !R R]4t !RR]PP4t!RR]PP4t!RR ]PP4t!RR ]PP4tR#)N)Enum)OptionalTensor) functionalInterpolationModeAutoAugmentPolicy AutoAugment RandAugmentTrivialAugmentWideAugMixc V^8dQhR\R\R\R\R\\ \,,/#)imgop_name magnitude interpolationfill)rstrfloatrrlist)formats"c/home/luca/screen-checking/.venv/lib/python3.14/site-packages/torchvision/transforms/autoaugment.py __annotate__r sGMM MM*/M@QMYabfglbmYnMc VR8XdO\P!VR^^.R\P!\P!V44R.VV^^.R7pV#VR8XdO\P!VR^^.RR\P!\P!V44.VV^^.R7pV#VR8Xd-\P!VR\ V4^.RVRR.VR7pV#VR8Xd-\P!VR^\ V4.RVRR.VR7pV#VR8Xd\P !WW4R 7pV#VR 8Xd!\P!VRV,4pV#VR 8Xd!\P!VRV,4pV#VR 8Xd!\P!VRV,4pV#VR 8Xd!\P!VRV,4pV#VR8Xd#\P!V\ V44pV#VR8Xd\P!W4pV#VR8Xd\P!V4pV#VR8Xd\P!V4pV#VR8Xd\P!V4pV#VR8XdV#\!RV R24h)ShearX?)angle translatescaleshearrrcenterShearY TranslateX)r r!r"rr#r TranslateYRotaterr BrightnessColorContrast Sharpness PosterizeSolarize AutoContrastEqualizeInvertIdentityzThe provided operator  is not recognized.)Faffinemathdegreesatanintrotateadjust_brightnessadjust_saturationadjust_contrastadjust_sharpness posterizesolarize autocontrastequalizeinvert ValueError)rrrrrs&&&&&r _apply_oprF s(hh !f<< ) 45s;'q6  F Js H hh !f TYYy%9:;'q6  l JY L hh 9~q)'* V JE L hh #i.)'* B J1 H hhs]N. J- L !!#sY7* J) G !!#sY7& J% J S9_5" J! K   cIo6 J K kk#s9~. J J jj( J N "nnS! J J jjo J H hhsm J J   J1':MNOOrc&]tRt^]tRtRtRtRtRtR#)r zgAutoAugment policies learned on different datasets. Available policies are IMAGENET, CIFAR10 and SVHN. imagenetcifar10svhnN) __name__ __module__ __qualname____firstlineno____doc__IMAGENETCIFAR10SVHN__static_attributes__rKrrr r ]sHG Drcaa]tRt^gtoRt]P ]PR3V3RlV3Rlllt V3RlRlt V3RlRlt ] V3R lR l4t V3R lR ltV3R lRltRtVtV;t#)r a AutoAugment data augmentation method based on `"AutoAugment: Learning Augmentation Strategies from Data" `_. If the image is torch Tensor, it should be of type torch.uint8, and it is expected to have [..., 1 or 3, H, W] shape, where ... means an arbitrary number of leading dimensions. If img is PIL Image, it is expected to be in mode "L" or "RGB". Args: policy (AutoAugmentPolicy): Desired policy enum defined by :class:`torchvision.transforms.autoaugment.AutoAugmentPolicy`. Default is ``AutoAugmentPolicy.IMAGENET``. interpolation (InterpolationMode): Desired interpolation enum defined by :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``. If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported. fill (sequence or number, optional): Pixel fill value for the area outside the transformed image. If given a number, the value is used for all bands respectively. NcP<V^8dQhRS[RS[RS[S[S[,,RR/#)rpolicyrrreturnN)r rrrr)r __classdict__s"rrAutoAugment.__annotate__xs< 3 3! 3) 3tE{# 3  3rct<\SV`4WnW nW0nVP V4VnR#N)super__init__rWrr _get_policiespolicies)selfrWrr __class__s&&&&rr^AutoAugment.__init__xs2  * **62 rc <V^8dQhRS[RS[S[S[S[S[S[S[,3,S[S[S[S[S[,3,3,,/#)rrWrX)r rtuplerrrr:)rrYs"rrrZsUXQXQ'XQ eE#uhsm34eCQT @r*rr+r,r-r.Fr/o@r0r1r2t?)torchlinspacearangeroundr:tensorrarrs&&&r_augmentation_spaceAutoAugment._augmentation_spacesz u~~c394@ u~~c394@ 5>>#}z!}/LhWY]^ 5>>#}z!}/LhWY]^ u~~c4:DA 5>>#sH=tD ennS#x8$? S(;TB %..c8$>+<+D+D&* 3 3XQXQt  & ' '8TTTrcaa]tRtRtoRt^^ ^]P R3V3RlV3RllltV3RlRltV3RlR lt V3R lR lt R t Vt V;t #) r iaBRandAugment data augmentation method based on `"RandAugment: Practical automated data augmentation with a reduced search space" `_. If the image is torch Tensor, it should be of type torch.uint8, and it is expected to have [..., 1 or 3, H, W] shape, where ... means an arbitrary number of leading dimensions. If img is PIL Image, it is expected to be in mode "L" or "RGB". Args: num_ops (int): Number of augmentation transformations to apply sequentially. magnitude (int): Magnitude for all the transformations. num_magnitude_bins (int): The number of different magnitude values. interpolation (InterpolationMode): Desired interpolation enum defined by :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``. If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported. fill (sequence or number, optional): Pixel fill value for the area outside the transformed image. If given a number, the value is used for all bands respectively. Nc \<V^8dQhRS[RS[RS[RS[RS[S[S[,,RR/#)rnum_opsrnum_magnitude_binsrrrXNr:rrrr)rrYs"rrRandAugment.__annotate__1sP      )  tE{#   rc`<\SV`4WnW nW0nW@nWPnR#r\)r]r^rrrrr)rarrrrrrbs&&&&&&rr^RandAugment.__init__1s+  ""4* rc n<V^8dQhRS[RS[S[S[3,RS[S[S[S[S[3,3,/#rr)rrYs"rrr@rrcvR\P!R4R3R\P!RRV4R3R\P!RRV4R3R\P!RRV^,,V4R3R\P!RRV^,,V4R3R \P!RR V4R3R \P!RR V4R3R \P!RR V4R3R\P!RR V4R3R\P!RR V4R3R^\P!V4V^, ^, , P 4P 4, R3R\P!RRV4R3R\P!R4R3R\P!R4R3/#)r3rFrrTr%r&r'r(rr*rr+r,r-r.r/rr0r1rrrrrrr:rs&&&rrRandAugment._augmentation_space@sz c*E2 u~~c394@ u~~c394@ 5>>#}z!}/LhWY]^ 5>>#}z!}/LhWY]^ u~~c4:DA 5>>#sH=tD ennS#x8$? S(;TB %..c8*84G!(!1 JDNOOVWDWj8==?@]`I%--400T! C)CUCU\`aC% 0s4F(c <V^8dQhRS[/#rr)rrYs"rrrns  # rc VPP RVP RVP RVP RVP RVP R2 pV#)z (num_ops=z , magnitude=z, num_magnitude_bins=, interpolation=rr)rbrLrrrrrrass& rrRandAugment.__repr__nsg~~&&'(||n4>>*#D$;$;#<t112dii[   r)rrrrrrLrMrNrOrPrrr^rrrrTrrrs@@rr r sN("$+<+D+D&*    &6   rcaa]tRtRtoRt^]P R3V3RlV3RllltV3RlRltV3RlR lt V3R lR lt R t Vt V;t #) r i{aDataset-independent data-augmentation with TrivialAugment Wide, as described in `"TrivialAugment: Tuning-free Yet State-of-the-Art Data Augmentation" `_. If the image is torch Tensor, it should be of type torch.uint8, and it is expected to have [..., 1 or 3, H, W] shape, where ... means an arbitrary number of leading dimensions. If img is PIL Image, it is expected to be in mode "L" or "RGB". Args: num_magnitude_bins (int): The number of different magnitude values. interpolation (InterpolationMode): Desired interpolation enum defined by :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``. If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported. fill (sequence or number, optional): Pixel fill value for the area outside the transformed image. If given a number, the value is used for all bands respectively. NcP<V^8dQhRS[RS[RS[S[S[,,RR/#)rrrrrXNr)rrYs"rrTrivialAugmentWide.__annotate__s<   ) tE{#   rcH<\SV`4WnW nW0nR#r\)r]r^rrr)rarrrrbs&&&&rr^TrivialAugmentWide.__init__s! "4* rcR<V^8dQhRS[RS[S[S[S[S[3,3,/#)rrrX)r:rrrerr)rrYs"rrrs.  C DeFDL>Q9Q4R rc>R\P!R4R3R\P!RRV4R3R\P!RRV4R3R\P!RRV4R3R \P!RRV4R3R \P!RR V4R3R \P!RRV4R3R \P!RRV4R3R\P!RRV4R3R\P!RRV4R3R^\P!V4V^, ^, , P 4P 4, R3R\P!RRV4R3R\P!R4R3R\P!R4R3/#)r3rFrgGz?Tr%r&g@@r'r(g`@r*r+r,r-r.r/rr0r1r)rars&&rr&TrivialAugmentWide._augmentation_spacesf c*E2 u~~c4:DA u~~c4:DA 5>>#tX>E 5>>#tX>E u~~c5(;TB 5>>#tX>E ennS$94@ T8`_. If the image is torch Tensor, it should be of type torch.uint8, and it is expected to have [..., 1 or 3, H, W] shape, where ... means an arbitrary number of leading dimensions. If img is PIL Image, it is expected to be in mode "L" or "RGB". Args: severity (int): The severity of base augmentation operators. Default is ``3``. mixture_width (int): The number of augmentation chains. Default is ``3``. chain_depth (int): The depth of augmentation chains. A negative value denotes stochastic depth sampled from the interval [1, 3]. Default is ``-1``. alpha (float): The hyperparameter for the probability distributions. Default is ``1.0``. all_ops (bool): Use all operations (including brightness, contrast, color and sharpness). Default is ``True``. interpolation (InterpolationMode): Desired interpolation enum defined by :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``. If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported. fill (sequence or number, optional): Pixel fill value for the area outside the transformed image. If given a number, the value is used for all bands respectively. rTNch<V^8dQhRS[RS[RS[RS[RS[RS[RS[S[S[,,RR /#) rseverity mixture_width chain_depthalphaall_opsrrrXN)r:rrrrr)rrYs"rrAugMix.__annotate__sd     )tE{# rc<\SV`4^ Vn^Tu;8:dVP8:gM\RVP RV R24hWnW nW0nW@nWPnW`n Wpn R#) z!The severity must be between [1, z]. Got z instead.N) r]r^_PARAMETER_MAXrErrrrr rr) rarrrrr rrrbs &&&&&&&&rr^AugMix.__init__sw  X4!4!44@ATAT@UU\]e\ffopq q *&  * rc n<V^8dQhRS[RS[S[S[3,RS[S[S[S[S[3,3,/#rr)rrYs"rrr sCCU38_QUVY[`agimam[nVnQorcR\P!RRV4R3R\P!RRV4R3R\P!RV^,R, V4R3R\P!RV^,R, V4R3R\P!RR V4R3R ^\P!V4V^, ^, , P4P 4, R 3R \P!R RV4R 3R\P !R4R 3R\P !R4R 3/ pVP 'dzVPR\P!RRV4R3R\P!RRV4R3R\P!RRV4R3R\P!RRV4R3/4V#)rrrTr%r&g@r'r(rr.Fr/rr0r1r*rr+r,r-)rrrrr:rr update)rarrrs&&& rrAugMix._augmentation_spaces u~~c394@ u~~c394@ 5>>#z!}s/BHMtT 5>>#z!}s/BHMtT u~~c4:DA !u||H5(Q,!9KLSSUYY[[]bc sH=uE U\\#.6 c*E2  <<< HH 5>>#sH#Et"LennS#x@$GS(!CT J%..c8"Dd!K  rc <V^8dQhRS[/#rr)rrYs"rrr s$$V$rc.\P!V4#r\)r5 pil_to_tensorrars&&r_pil_to_tensorAugMix._pil_to_tensorss##rc <V^8dQhRS[/#)rrr)rrYs"rrr s##&#rc.\P!V4#r\)r5 to_pil_imagers&&r_tensor_to_pilAugMix._tensor_to_pils~~c""rc&<V^8dQhRS[RS[/#)rparamsrXr)rrYs"rrr s///6/rc.\P!V4#r\)r_sample_dirichlet)rars&&rr!AugMix._sample_dirichlets&&v..rc&<V^8dQhRS[RS[/#)rorig_imgrXr)rrYs"rrr  s88868rc  VPp\P!V4wr4p\V\4'dRTp\V\ \ 34'd\ V4.V,pM0VeVUu.uFp\ V4NK ppMVPV4pVPVPWE34p\VP4p VP^.\^VP, ^4,V ,4p V P^4.^.V P^, ,,p VP!\"P$!VP&VP&.V P(R7P+V ^,R44p VP!\"P$!VP&.VP,,V P(R7P+V ^,R44V R,PV ^,R.4,p V R ,PV 4V ,p\/VP,4EFpT pVP0^8d VP0M/\ \"P2!^^R R7P544p\/V4EFp\ \"P2!\7V4R 4P544p\VP944V,pVV,wppVP^8dP\ V\"P2!VP:R \"P<R7,P544MRpV'd'\"P2!^R 4'd VR ,p\?VVVVP@VR7pEK VPCV RV3,PV 4V,4EK VPV 4PEVPFR7p\V\4'gVPIV4#V#uupi) r)deviceNNN)lowhighsizerrr))r'r)r'rrr)%rr5rrrr:rrrr rshapeviewmaxrr*r!rrrr&expandrrrrrrrrrrFradd_torr)rar$rrrrrrr orig_dimsbatch batch_dimsmcombined_weightsmixraugdepthrrrrrrs&& rrAugMix.forward sQyy"#"2"28"<% h ' 'C$e --d }x/!*./$Qa$/%%h/C**4+>+>PO !s1sxx<33i?@jjm_sejj1n'==   " " LL$**djj1%,, G N NzZ[}^` a   11 LL$**(:(::5<< P W WXbcdXegi j dGLL*Q-, -.gll:&.t))*AC(,(8(81(rGs  0 ]M`tT%((//tTnZ%((//ZzSSlUUXX__Ur