ڧg
d dl mZ d dlmZmZmZmZmZmZ d dl Z d dl
m Z d dl m Z ddl
mZ ddlmZ dd lmZmZmZ dd
lmZ dd lmZmZ g d ZdEd
ededededede j f dZ dFd
ededede j fdZ G d de j Z G d de j Z! G d de j Z"deee e!f d ee d ee d e#d ede"f d Z$d!ed"Z% G d# d$e Z& G d% d&e Z' G d' d(e Z( G d) d*e Z) G d+ d,e Z* G d- d.e Z+ G d/ d0e Z, G d1 d2e Z- G d3 d4e Z. G d5 d6e Z/ e ed7e&j0 f8 dd9d:d ee& d e#d ede"fd; Z1 e ed7e'j0 f8 dd9d:d ee' d e#d ede"fd< Z2 e ed7e(j0 f8 dd9d:d ee( d e#d ede"fd= Z3 e ed7e)j0 f8 dd9d:d ee) d e#d ede"fd> Z4 e ed7e*j0 f8 dd9d:d ee* d e#d ede"fd? Z5 e ed7e+j0 f8 dd9d:d ee+ d e#d ede"fd@ Z6 e ed7e,j0 f8 dd9d:d ee, d e#d ede"fdA Z7 e ed7e-j0 f8 dd9d:d ee- d e#d ede"fdB Z8 e ed7e.j0 f8 dd9d:d ee. d e#d ede"fdC Z9 e ed7e/j0 f8 dd9d:d ee/ d e#d ede"fdD Z:dS )G )partial)AnyCallableListOptionalTypeUnionN)Tensor )ImageClassification)_log_api_usage_once )register_modelWeights WeightsEnum)_IMAGENET_CATEGORIES)_ovewrite_named_paramhandle_legacy_interface)ResNetResNet18_WeightsResNet34_WeightsResNet50_WeightsResNet101_WeightsResNet152_WeightsResNeXt50_32X4D_WeightsResNeXt101_32X8D_WeightsResNeXt101_64X4D_WeightsWide_ResNet50_2_WeightsWide_ResNet101_2_Weightsresnet18resnet34resnet50 resnet101 resnet152resnext50_32x4dresnext101_32x8dresnext101_64x4dwide_resnet50_2wide_resnet101_2 in_planes
out_planesstridegroupsdilationreturnc
: t j | |d|||d| S )z 3x3 convolution with padding F) kernel_sizer, paddingr- biasr. nnConv2d)r* r+ r, r- r. s U/var/www/html/ai-engine/env/lib/python3.11/site-packages/torchvision/models/resnet.pyconv3x3r9 ( s3
9
c 4 t j | |d|d S )z1x1 convolutionr F)r2 r, r4 r5 )r* r+ r, s r8 conv1x1r< 6 s
9Y
&u U U UUr: c e Zd ZU dZeed< ddedededeej d ed
ed ed ee d
ej f ddf fd
Z
de de fdZ xZ
S )
BasicBlockr expansionN@ inplanesplanesr,
downsampler-
base_widthr.
norm_layer.r/ c t | t j }|dk s|dk rt d |dk rt d t
||| | _ || | _ t j d | _
t
|| | _ || | _ || _
|| _ d S )Nr r@ z3BasicBlock only supports groups=1 and base_width=64z(Dilation > 1 not supported in BasicBlockTinplace)super__init__r6 BatchNorm2d
ValueErrorNotImplementedErrorr9 conv1bn1ReLUreluconv2bn2rC r, )
selfrA rB r, rC r- rD r. rE __class__s
r8 rJ zBasicBlock.__init__> s J Q;;*** RSS S a<<%&PQQ Q Xvv66
:f%%GD))) VV,,
:f%%$ r: xc H |}| | }| | }| | }| | }| | }| j | | }||z
}| | }|S N)rN rO rQ rR rS rC rT rV identityouts r8 forwardzBasicBlock.forwardY s jjmmhhsmmiinnjjoohhsmm ? &q))H xiinn
r: r Nr r@ r N__name__
__module__ __qualname__r? int__annotations__r r6 Moduler rJ r
r\
__classcell__rU s @r8 r> r> ; s Is *. 9=
RY'
Xc29n 56
6 F r: r> c e Zd ZU dZeed< ddededed eej d
ed ed ed
ee dej f ddf fd
Z
de de fdZ xZ
S )
Bottleneck r? r Nr@ rA rB r, rC r- rD r. rE .r/ c t | t j }t ||dz z |z } t || | _ || | _ t | | ||| | _ || | _
t | || j z | _ ||| j z | _
t j d | _ || _ || _ d S )Ng P@TrG )rI rJ r6 rK rb r< rN rO r9 rR rS r? conv3bn3rP rQ rC r, ) rT rA rB r, rC r- rD r. rE widthrU s r8 rJ zBottleneck.__init__u s JFj4 /011F: Xu--
:e$$ UE668DD
:e$$ UFT^$;<<
:ft~ 566GD))) $ r: rV c |}| | }| | }| | }| | }| | }| | }| | }| | }| j | | }||z
}| | }|S rX )rN rO rQ rR rS rk rl rC rY s r8 r\ zBottleneck.forward s jjmmhhsmmiinnjjoohhsmmiinnjjoohhsmm ? &q))H xiinn
r: r] r^ rf s @r8 rh rh l s Is *. 9=
RY'
Xc29n 56
4 F r: rh c $ e Zd Z ddeeeef dee ded e d
ed ed e
ee d
e
e de j
f ddf fd
Z ddeeeef dededede de j f dZdedefdZdedefdZ xZS )r Fr r@ Nblocklayers num_classeszero_init_residualr- width_per_group replace_stride_with_dilationrE .r/ c t t | | t j }|| _ d| _ d| _ |g d}t | dk rt d| || _
|| _ t j d| j dddd | _
|| j | _ t j d
| _ t j ddd | _ | |d|d
| _ | |d|d d|d
| _ | |d|d d|d | _ | |d|d d|d | _ t j d | _ t j d|j z | | _ | D ]} t= | t j r(t j | j! dd Dt= | t j t j" f rJt j # | j! d t j # | j$ d
|r| D ]} t= | tJ r7| j& j! +t j # | j& j! d
Nt= | tN r6| j( j! *t j # | j( j! d
d S d S )Nr@ r )FFFr1 zFreplace_stride_with_dilation should be None or a 3-element tuple, got r F)r2 r, r3 r4 TrG )r2 r, r3 r )r, dilate i r r fan_outrQ )mode nonlinearity))rI rJ r
r6 rK _norm_layerrA r. lenrL r- rD r7 rN rO rP rQ MaxPool2dmaxpool _make_layerlayer1layer2layer3layer4AdaptiveAvgPool2davgpoolLinearr? fcmodules
isinstanceinitkaiming_normal_weight GroupNorm constant_r4 rh rl r> rS ) rT rq rr rs rt r- ru rv rE mrU s r8 rJ zResNet.__init__ s D!!! J%
' / ,A+@+@ ( + , , 1 1 L-IL L
)Yq$-QqRSZ_```
:dm,,GD))) |!QGGG &&ub&)<< &&uc6!9QOklmOn&oo &&uc6!9QOklmOn&oo &&uc6!9QOklmOn&oo +F33 )C%/ 1;?? - -A!RY''
-''yv'VVVVA =>>
-!!!(A...!!!&!,,,
7\\^^
7
7 a,, 7 1IG%%aelA6666 :.. 715<3KG%%aelA666 7 7
7
7r: rB blocksr, rz c < | j }d }| j }|r| xj |z c_ d}|dk s| j ||j z k rBt j t
| j ||j z | |||j z }g } | || j |||| j | j || ||j z | _ t d| D ]:}
| || j || j | j | j | ;t j | S )Nr )r- rD r. rE ) r r. rA r? r6
Sequentialr< appendr- rD range) rT rq rB r r, rz rE rC previous_dilationrr _s r8 r zResNet._make_layer sU %
M MMV #MMF Q;;$-6EO+CCC
v'?HH
6EO344 J
E
vvz4;Yjlv
0
q&!!
A MMM ; # !] )
}f%%r: rV c | | }| | }| | }| | }| | }| | }| | }| | }| | }t j
|d }| | }|S )Nr ) rN rO rQ r r r r r r torchflattenr rT rV s r8
_forward_implzResNet._forward_impl
s JJqMM HHQKK IIaLL LLOO KKNN KKNN KKNN KKNN LLOO M!Q GGAJJr: c , | | S rX )r r s r8 r\ zResNet.forward s !!!$$$r: )rp Fr r@ NN)r F)r_ r` ra r r r> rh r rb boolr r r6 rd rJ r r r
r r\ re rf s @r8 r r s
#( !=A9=87 87E*j01287 S 87 87
! 87
87 87 '/tDz&:87 Xc29n 5687
87 87 87 87 87 87~
'& '&E*j012'& '& '&
'&
'&
'& '& '& '&Rv & $% %F % % % % % % % %r: r rq rr weightsprogresskwargsc |)t |dt |j d t | |fi |}|*| | |d |S )Nrs
categoriesT)r
check_hash)r r metar load_state_dictget_state_dict)rq rr r r r models r8 _resnetr su fmS l9S5T5TUUU 5& + +F + +E
g 4 4hSW 4 X XYYY Lr: r| )min_sizer c d e Zd Z ed eed i edddddd id
d d d
ZeZdS )r z9https://download.pytorch.org/models/resnet18-f37072fd.pth crop_sizei(^ Lhttps://github.com/pytorch/vision/tree/main/references/classification#resnet ImageNet-1KgnpQ@gEDV@zacc@1zacc@5g/$?gS㥛TF@XThese weights reproduce closely the results of the paper using a simple training recipe.
num_paramsrecipe_metrics_ops
_file_size_docsurl
transformsr N r_ r` ra r r r _COMMON_META
IMAGENET1K_V1DEFAULT r: r8 r r 8 G G7.#>>>
"d # # s
M$ GGGr: r c d e Zd Z ed eed i edddddd id
d d d
ZeZdS )r z9https://download.pytorch.org/models/resnet34-b627a593.pthr r i(Lr r gjtTR@g{GV@r g Zd;O
@gT@r r r Nr r r: r8 r r N r r: r c e Zd Z ed eed i edddddd id
d d d
Z ed eedd i edddddd id
ddd
ZeZ dS )r z9https://download.pytorch.org/models/resnet50-0676ba61.pthr r i(r r g QS@gI +7W@r gB`"[@gDlqX@r r r z9https://download.pytorch.org/models/resnet50-11ad3fa6.pth r resize_sizezEhttps://github.com/pytorch/vision/issues/3995#issuecomment-1013906621gx6T@gW@g(\rX@
These weights improve upon the results of the original paper by using TorchVision's `new training recipe
`_.
N
r_ r` ra r r r r r
IMAGENET1K_V2r r r: r8 r r d s G G7.#>>>
"d # # s
M$ G G7.#3OOO
"] # #
M* GGGr: r c e Zd Z ed eed i edddddd id
d d d
Z ed eedd i edddddd id
ddd
ZeZ dS )r z:https://download.pytorch.org/models/resnet101-63fe2227.pthr r i(ħr r g-WS@gmbW@r gNbX94 @g1 ZPe@r r r z:https://download.pytorch.org/models/resnet101-cd907fc2.pthr r 8https://github.com/pytorch/vision/issues/3995#new-recipegbX9xT@gR W@g)\Pe@r Nr r r: r8 r r s G H7.#>>>
"d # # !s
M$ G H7.#3OOO
"P # #
M* GGGr: r c e Zd Z ed eed i edddddd id
d d d
Z ed eedd i edddddd id
ddd
ZeZ dS )r z:https://download.pytorch.org/models/resnet152-394f9c45.pthr r i(xr r gS@gmW@r gI +'@gSl@r r r z:https://download.pytorch.org/models/resnet152-f82ba261.pthr r r gV-T@g㥛 X@gI +l@r Nr r r: r8 r r s G H7.#>>>
"d # # !s
M$ G H7.#3OOO
"P # # !
M* GGGr: r c e Zd Z ed eed i edddddd id
d d d
Z ed eedd i edddddd id
ddd
ZeZ dS )r z@https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pthr r i(}Mhttps://github.com/pytorch/vision/tree/main/references/classification#resnextr gd;OgS@g&1lW@r gQ @g"~W@r r r z@https://download.pytorch.org/models/resnext50_32x4d-1a0047aa.pthr r r g&1LT@g(\W@gZd;OW@r Nr r r: r8 r r s G N7.#>>>
"e # # s
M$ G N7.#3OOO
"P # #
M* GGGr: r c e Zd Z ed eed i edddddd id
d d d
Z ed eedd i edddddd id
ddd
ZeZ dS )r zAhttps://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pthr r i(Jr r gS@glW@r gDli0@gL7A`9u@r r r zAhttps://download.pytorch.org/models/resnext101_32x8d-110c445d.pthr r Dhttps://github.com/pytorch/vision/issues/3995#new-recipe-with-fixresgL7A`T@g;OX@gT㥛:u@r Nr r r: r8 r r s G O7.#>>>
"e # # !s
M$ G O7.#3OOO
"\ # # !
M* GGGr: r c f e Zd Z ed eedd i eddddd d
id d d
d ZeZdS )r zAhttps://download.pytorch.org/models/resnext101_64x4d-173b62eb.pthr r r i(mz+https://github.com/pytorch/vision/pull/5935r g9vT@gV X@r gQ .@g +s@z
These weights were trained from scratch by using TorchVision's `new training recipe
`_.
r r Nr r r: r8 r r ; s G O7.#3OOO
"C # # !
M* GGGr: r c e Zd Z ed eed i edddddd id
d d d
Z ed eedd i edddddd id
ddd
ZeZ dS )r z@https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pthr r i(:https://github.com/pytorch/vision/pull/912#issue-445437439r gˡES@g/$W@r g&@g
ףp=z`@r r r z@https://download.pytorch.org/models/wide_resnet50_2-9ba9bcbe.pthr r r gJ +fT@gnW@gDlqp@r Nr r r: r8 r r T s G N7.#>>>
"R # # s
M$ G N7.#3OOO
"\ # # !
M* GGGr: r c e Zd Z ed eed i edddddd id
d d d
Z ed eedd i edddddd id
ddd
ZeZ dS )r zAhttps://download.pytorch.org/models/wide_resnet101_2-32ee1156.pthr r i(#r r gʡES@gV-W@r gT㥛6@g&1\n@r r r zAhttps://download.pytorch.org/models/wide_resnet101_2-d733dc28.pthr r r gq=
ףT@gzGX@gˡEK~@r Nr r r: r8 r r s G O7.#>>>
#R # # !s
M$ G O7.#3OOO
#P # # !
M* GGGr: r
pretrained)r T)r r c d t | } t t g d| |fi |S )ap ResNet-18 from `Deep Residual Learning for Image Recognition `__.
Args:
weights (:class:`~torchvision.models.ResNet18_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ResNet18_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ResNet18_Weights
:members:
)r r r r )r verifyr r> r r r s r8 r r 7 * %%g..G :|||Wh I I& I IIr: c d t | } t t g d| |fi |S )ap ResNet-34 from `Deep Residual Learning for Image Recognition `__.
Args:
weights (:class:`~torchvision.models.ResNet34_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ResNet34_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ResNet34_Weights
:members:
r1 ri r1 )r r r r> r s r8 r! r! r r: c d t | } t t g d| |fi |S )a ResNet-50 from `Deep Residual Learning for Image Recognition `__.
.. note::
The bottleneck of TorchVision places the stride for downsampling to the second 3x3
convolution while the original paper places it to the first 1x1 convolution.
This variant improves the accuracy and is known as `ResNet V1.5
`_.
Args:
weights (:class:`~torchvision.models.ResNet50_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ResNet50_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ResNet50_Weights
:members:
r )r r r rh r s r8 r" r" s7 6 %%g..G :|||Wh I I& I IIr: c d t | } t t g d| |fi |S )a ResNet-101 from `Deep Residual Learning for Image Recognition `__.
.. note::
The bottleneck of TorchVision places the stride for downsampling to the second 3x3
convolution while the original paper places it to the first 1x1 convolution.
This variant improves the accuracy and is known as `ResNet V1.5
`_.
Args:
weights (:class:`~torchvision.models.ResNet101_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ResNet101_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ResNet101_Weights
:members:
r1 ri r1 )r r r rh r s r8 r# r# 7 6 &&w//G :}}}gx J J6 J JJr: c d t | } t t g d| |fi |S )a ResNet-152 from `Deep Residual Learning for Image Recognition `__.
.. note::
The bottleneck of TorchVision places the stride for downsampling to the second 3x3
convolution while the original paper places it to the first 1x1 convolution.
This variant improves the accuracy and is known as `ResNet V1.5
`_.
Args:
weights (:class:`~torchvision.models.ResNet152_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ResNet152_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ResNet152_Weights
:members:
)r1 $ r1 )r r r rh r s r8 r$ r$ r r: c t | } t |dd t |dd t t g d| |fi |S )a ResNeXt-50 32x4d model from
`Aggregated Residual Transformation for Deep Neural Networks `_.
Args:
weights (:class:`~torchvision.models.ResNeXt50_32X4D_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ResNext50_32X4D_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ResNeXt50_32X4D_Weights
:members:
r- ru ri r )r r r r rh r s r8 r% r% > s\ . &,,W55G&(B///&"3Q777 :|||Wh I I& I IIr: c t | } t |dd t |dd t t g d| |fi |S )a ResNeXt-101 32x8d model from
`Aggregated Residual Transformation for Deep Neural Networks `_.
Args:
weights (:class:`~torchvision.models.ResNeXt101_32X8D_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ResNeXt101_32X8D_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ResNeXt101_32X8D_Weights
:members:
r- r ru r r )r r r r rh r s r8 r& r& \ \ . '--g66G&(B///&"3Q777 :}}}gx J J6 J JJr: c t | } t |dd t |dd t t g d| |fi |S )a ResNeXt-101 64x4d model from
`Aggregated Residual Transformation for Deep Neural Networks `_.
Args:
weights (:class:`~torchvision.models.ResNeXt101_64X4D_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.ResNeXt101_64X4D_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.ResNeXt101_64X4D_Weights
:members:
r- r@ ru ri r )r r r r rh r s r8 r' r' z r r: c t | } t |dd t t g d| |fi |S )a Wide ResNet-50-2 model from
`Wide Residual Networks `_.
The model is the same as ResNet except for the bottleneck number of channels
which is twice larger in every block. The number of channels in outer 1x1
convolutions is the same, e.g. last block in ResNet-50 has 2048-512-2048
channels, and in Wide ResNet-50-2 has 2048-1024-2048.
Args:
weights (:class:`~torchvision.models.Wide_ResNet50_2_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.Wide_ResNet50_2_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.Wide_ResNet50_2_Weights
:members:
ru ry r )r r r r rh r s r8 r( r( sJ 8 &,,W55G&"3V<<< :|||Wh I I& I IIr: c t | } t |dd t t g d| |fi |S )a Wide ResNet-101-2 model from
`Wide Residual Networks `_.
The model is the same as ResNet except for the bottleneck number of channels
which is twice larger in every block. The number of channels in outer 1x1
convolutions is the same, e.g. last block in ResNet-101 has 2048-512-2048
channels, and in Wide ResNet-101-2 has 2048-1024-2048.
Args:
weights (:class:`~torchvision.models.Wide_ResNet101_2_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.Wide_ResNet101_2_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.resnet.ResNet``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.Wide_ResNet101_2_Weights
:members:
ru ry r )r r r r rh r s r8 r) r) sJ 8 '--g66G&"3V<<< :}}}gx J J6 J JJr: )r r r )r ); functoolsr typingr r r r r r r torch.nnr6 r
transforms._presetsr utilsr
_apir r r _metar _utilsr r __all__rb r7 r9 r< rd r> rh r r r r r r r r r r r r r r r r r! r" r# r$ r% r& r' r( r) r r: r8 r s = = = = = = = = = = = = = = = = 5 5 5 5 5 5 ' ' ' ' ' ' 6 6 6 6 6 6 6 6 6 6 ' ' ' ' ' ' B B B B B B B B 2 s S c Y\ egen V Vs V VS V V V V V
. . . . . . . .b7 7 7 7 7 7 7 7tw% w% w% w% w%RY w% w% w%t j*,- . I k
"
& & { , { ,( ( ( ( ( { ( ( ( V( ( ( ( ( ( ( ( V( ( ( ( ( ( ( ( V( ( ( ( ( k ( ( ( V( ( ( ( ( { ( ( ( V { 2( ( ( ( ( k ( ( ( V( ( ( ( ( { ( ( ( V ,0@0N!OPPP6:T J J J"23 Jd J]` Jek J J J QP J0 ,0@0N!OPPP6:T J J J"23 Jd J]` Jek J J J QP J0 ,0@0N!OPPP6:T J J J"23 Jd J]` Jek J J J QP J< ,0A0O!PQQQ8