hN< & S SK Jr S SKJrJrJr S SKrS SKJr S SKJ r SSK
J r SSK J
r
SS KJrJrJr SS
KJr SS KJrJr / S QrS
\ S\S\ 4S jr " S S\R4 5 r " S S\R4 5 r S\\ S\ S\S\S\ 4
S jr S\SS .r " S S \5 r " S S \5 r! " S! S"\5 r" " S# S$\5 r#\" 5 \" S%\ RH 4S&9SS'S(.S\\ S\ S\S\ 4S) jj5 5 r%\" 5 \" S%\!RH 4S&9SS'S(.S\\! S\ S\S\ 4S* jj5 5 r&\" 5 \" S%\"RH 4S&9SS'S(.S\\" S\ S\S\ 4S+ jj5 5 r'\" 5 \" S%\#RH 4S&9SS'S(.S\\# S\ S\S\ 4S, jj5 5 r(g)- )partial)AnyCallableOptionalN)Tensor )ImageClassification)_log_api_usage_once )register_modelWeights WeightsEnum)_IMAGENET_CATEGORIES)_ovewrite_named_paramhandle_legacy_interface) ShuffleNetV2ShuffleNet_V2_X0_5_WeightsShuffleNet_V2_X1_0_WeightsShuffleNet_V2_X1_5_WeightsShuffleNet_V2_X2_0_Weightsshufflenet_v2_x0_5shufflenet_v2_x1_0shufflenet_v2_x1_5shufflenet_v2_x2_0xgroupsreturnc U R 5 u p#pEX1- nU R X!XdU5 n [ R " U SS5 R 5 n U R X#XE5 n U $ )Nr r )sizeviewtorch transpose
contiguous)r r batchsize num_channelsheightwidthchannels_per_groups W/var/www/auris/envauris/lib/python3.13/site-packages/torchvision/models/shufflenetv2.pychannel_shuffler* sb -.VVX*IV%/
y"4eDA
1a ++-A
y6A H c ^ \ rS rSrS\S\S\SS4U 4S jjr\ SS\S \S
\S\S \S \S\R 4S
jj5 r
S\ S\ 4S jr Sr
U =r$ )InvertedResidual+ inpoupstrider Nc > [ TU ] 5 SUs=:: a S:: d
O [ S5 eX0l US- nU R S:X a XS- :w a [ SU SU SU S35 eU R S: a [ R
" U R
XSU R SS 9[ R " U5 [ R " XSSS
S S 9[ R " U5 [ R " S
S95 U l
O[ R
" 5 U l
[ R
" [ R " U R S: a UOUUSSS
S S 9[ R " U5 [ R " S
S9U R
XDSU R SS 9[ R " U5 [ R " XDSSS
S S 9[ R " U5 [ R " S
S95 U l g )Nr zillegal stride valuer z Invalid combination of stride z, inp z and oup zB values. If stride == 1 then inp should be equal to oup // 2 << 1. kernel_sizer1 paddingr F)r5 r1 r6 biasTinplace) super__init__
ValueErrorr1 nn
Sequentialdepthwise_conv BatchNorm2dConv2dReLUbranch1branch2)selfr/ r0 r1 branch_features __class__s r) r; InvertedResidual.__init__, s
V q 34 4 ( KK1 3Q*>#> 0se9SE RT U
;;? ==##C!DKKYZ#[s# #AaQRY^_/% DL ==?DL}} II ao
NN? + GGD ! aX\XcXcmn o NN? + IIoAaYZaf g NN? + GGD !
r+ ior5 r6 r7 c
0 [ R " XX#XEU S9$ )N)r7 r )r= rA )rI rJ r5 r1 r6 r7 s r) r? InvertedResidual.depthwise_convV s yy{GqQQr+ r c U R S:X a8 UR SSS9u p#[ R " X R U5 4SS9nO5[ R " U R U5 U R U5 4SS9n[
US5 nU$ )Nr r )dim)r1 chunkr! catrD rC r* )rE r x1x2outs r) forwardInvertedResidual.forward\ st ;;! WWQAW&FB))R b!1 2:C))T\\!_dll1o >AFC c1%
r+ )rC rD r1 )r r F)__name__
__module__ __qualname____firstlineno__intr; staticmethodboolr= rA r? r rT __static_attributes__
__classcell__rG s @r) r- r- + s (
C (
c (
3 (
4 (
T Z_R RR%(R25RDGRSWR R R
F r+ r- c ^ \ rS rSrS\4S\\ S\\ S\S\S\R 4 SS 4
U 4S
jjjr
S \ S\ 4S jr S \ S\ 4S
jr
SrU =r$ )r h i stages_repeatsstages_out_channels num_classesinverted_residual.r Nc > [ T U ] 5 [ U 5 [ U5 S:w a [ S5 e[ U5 S:w a [ S5 eX l SnU R
S n[ R " [ R " XVSSSSS 9[ R " U5 [ R " S
S 95 U l Un[ R " SSSS 9U l
U U U S
Vs/ s H nSU 3PM
nn[ XU R
SS 5 H\ u pnU" XVS5 /n [ U
S-
5 H nU R! U" XfS5 5 M [# X [ R " U 6 5 UnM^ U R
S n[ R " [ R " XVSSSSS 9[ R " U5 [ R " S
S 95 U l [ R& " Xc5 U l g s snf )Nr3 z2expected stages_repeats as list of 3 positive ints z7expected stages_out_channels as list of 5 positive intsr r r F)r7 Tr8 r4 )r r3 stage)r: r; r
lenr< _stage_out_channelsr= r> rA r@ rB conv1 MaxPool2dmaxpoolziprangeappendsetattrconv5Linearfc)
rE rb rc rd re input_channelsoutput_channelsrI stage_namesnamerepeatsseqrG s
r) r; ShuffleNetV2.__init__i s D! ~ ! # QR R " #q ( VW W#6 2215]] IInq!QU K NN? + GGD !
)||!QG
,56Iqqc{I 6.1+tOgOghihjOk.l *D?$^aHIC7Q;'
,_qQR ( D
s 3 4 ,N /m 2226]] IInq!QU K NN? + GGD !
))O9 7s &G4r c U R U5 nU R U5 nU R U5 nU R U5 nU R U5 nU R U5 nUR
SS/5 nU R U5 nU$ )Nr r3 )rm ro stage2stage3stage4rt meanrv rE r s r)
_forward_implShuffleNetV2._forward_impl ss JJqM LLO KKN KKN KKN JJqM
FFAq6N GGAJr+ c $ U R U5 $ )N)r r s r) rT ShuffleNetV2.forward s !!!$$r+ )rl rm rt rv ro )rV rW rX rY r- listrZ r r= Moduler; r r rT r] r^ r_ s @r) r r h s
6F .: S .: "#Y.: .:
$CN3 .:
.: .:`
v
&
% %F % %r+ r weightsprogressargskwargsc U b# [ US[ U R S 5 5 [ U0 UD6nU b UR U R USS95 U$ )Nrd
categoriesT)r
check_hash)r rk metar load_state_dictget_state_dict)r r r r models r)
_shufflenetv2r s] fmS l9S5TU $ )& )E
g 4 4hSW 4 XY Lr+ )r r z2https://github.com/ericsun99/Shufflenet-v2-Pytorch)min_sizer recipec P \ rS rSr\" S\" \SS90 \ESSSSS .0S
S S S
.ES9r\r Sr
g)r zDhttps://download.pytorch.org/models/shufflenetv2_x0.5-f707e7126e.pth crop_sizei ImageNet-1Kg-FN@g9voT@zacc@1zacc@5g{Gz?gT㥛 @VThese weights were trained from scratch to reproduce closely the results of the paper.
num_params_metrics_ops
_file_size_docsurl
transformsr N rV rW rX rY r
r r _COMMON_META
IMAGENET1K_V1DEFAULTr] r r+ r) r r sT R.#>
! # # q
M$ Gr+ r c P \ rS rSr\" S\" \SS90 \ESSSSS .0S
S S S
.ES9r\r Sr
g)r zBhttps://download.pytorch.org/models/shufflenetv2_x1-5666bf0f80.pthr r i" r gI +WQ@gNbX9V@r g(\?gE!@r r r r Nr r r+ r) r r sT P.#>
! # # q
M$ Gr+ r c T \ rS rSr\" S\" \SSS90 \ESSSS S
S .0S S
SS.ES9r\r Sr
g)r zBhttps://download.pytorch.org/models/shufflenetv2_x1_5-3c479a10.pthr r resize_size+https://github.com/pytorch/vision/pull/5906iv5 r g9v?R@g/$V@r gl?gw/ +@
These weights were trained from scratch by using TorchVision's `new training recipe
`_.
r r r r r r r r Nr r r+ r) r r [ P.#3O
C! # #
M* Gr+ r c T \ rS rSr\" S\" \SSS90 \ESSSS S
S .0S S
SS.ES9r\r Sr
g)r i zBhttps://download.pytorch.org/models/shufflenetv2_x2_0-8be3c8ee.pthr r r r ip r g QS@gMb@W@r g-?g+n<@r r r r Nr r r+ r) r r r r+ r
pretrained)r T)r r c P [ R U 5 n [ X/ SQ/ SQ40 UD6$ )a
Constructs a ShuffleNetV2 architecture with 0.5x output channels, as described in
`ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
`__.
Args:
weights (:class:`~torchvision.models.ShuffleNet_V2_X0_5_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ShuffleNet_V2_X0_5_Weights` below for
more details, and possible values. By default, no pre-trained
weights are used.
progress (bool, optional): If True, displays a progress bar of the
download to stderr. Default is True.
**kwargs: parameters passed to the ``torchvision.models.shufflenetv2.ShuffleNetV2``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ShuffleNet_V2_X0_5_Weights
:members:
rh rh ) 0 ` )r verifyr r r r s r) r r s* 4 )//8G I7N YRX YYr+ c P [ R U 5 n [ X/ SQ/ SQ40 UD6$ )a
Constructs a ShuffleNetV2 architecture with 1.0x output channels, as described in
`ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
`__.
Args:
weights (:class:`~torchvision.models.ShuffleNet_V2_X1_0_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ShuffleNet_V2_X1_0_Weights` below for
more details, and possible values. By default, no pre-trained
weights are used.
progress (bool, optional): If True, displays a progress bar of the
download to stderr. Default is True.
**kwargs: parameters passed to the ``torchvision.models.shufflenetv2.ShuffleNetV2``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ShuffleNet_V2_X1_0_Weights
:members:
r )r t r i r )r r r r s r) r r > * 4 )//8G I7P [TZ [[r+ c P [ R U 5 n [ X/ SQ/ SQ40 UD6$ )a
Constructs a ShuffleNetV2 architecture with 1.5x output channels, as described in
`ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
`__.
Args:
weights (:class:`~torchvision.models.ShuffleNet_V2_X1_5_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ShuffleNet_V2_X1_5_Weights` below for
more details, and possible values. By default, no pre-trained
weights are used.
progress (bool, optional): If True, displays a progress bar of the
download to stderr. Default is True.
**kwargs: parameters passed to the ``torchvision.models.shufflenetv2.ShuffleNetV2``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ShuffleNet_V2_X1_5_Weights
:members:
r )r i` i r )r r r r s r) r r ] r r+ c P [ R U 5 n [ X/ SQ/ SQ40 UD6$ )a
Constructs a ShuffleNetV2 architecture with 2.0x output channels, as described in
`ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
`__.
Args:
weights (:class:`~torchvision.models.ShuffleNet_V2_X2_0_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ShuffleNet_V2_X2_0_Weights` below for
more details, and possible values. By default, no pre-trained
weights are used.
progress (bool, optional): If True, displays a progress bar of the
download to stderr. Default is True.
**kwargs: parameters passed to the ``torchvision.models.shufflenetv2.ShuffleNetV2``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ShuffleNet_V2_X2_0_Weights
:members:
r )r i i i )r r r r s r) r r | r r+ )) functoolsr typingr r r r! torch.nnr= r transforms._presetsr utilsr
_apir r
r _metar _utilsr r __all__rZ r* r r- r r\ r r r r r r r r r r r r r+ r) r sn * * 5 ' 6 6 ' B
v
s
v
:ryy :z>%299 >%B
k
"
$ &B , , 2 2 ,0J0X0X!YZ7;dZ34ZGKZ^aZZ [ Z: ,0J0X0X!YZ7;d\34\GK\^a\\ [ \: ,0J0X0X!YZ7;d\34\GK\^a\\ [ \: ,0J0X0X!YZ7;d\34\GK\^a\\ [ \r+