a
PSic( @ sP d dl Z d dlmZ d dlmZmZmZmZ d dlZd dlm Z d dlm
Z
ddl m Z ddl
mZ dd lmZ d
d lmZmZ d
d lmZ d
d
lmZmZmZ g dZG dd de ZeZ eZ G dd de
j Z G dd de
j Z ddedZ!G dd deZ"ede"j#fddd d ee" e$ee d d d Z%d
d!lm&Z& e&d e"j#j'iZ(dS )" N)partial)CallableAnyOptionalList)Tensor)nn )Conv2dNormActivation)ImageClassification)_log_api_usage_once ) WeightsEnumWeights)_IMAGENET_CATEGORIES)handle_legacy_interface_ovewrite_named_param_make_divisible) MobileNetV2MobileNet_V2_Weights mobilenet_v2c s e Zd Z fddZ ZS )_DeprecatedConvBNActc sV t dt |dd d u r&tj|d< |dd d u r@tj|d< t j|i | d S )NzThe ConvBNReLU/ConvBNActivation classes are deprecated since 0.12 and will be removed in 0.14. Use torchvision.ops.misc.Conv2dNormActivation instead.
norm_layeractivation_layer) warningswarn
FutureWarninggetr BatchNorm2dReLU6super__init__)selfargskwargs __class__ Z/var/www/html/django/DPS/env/lib/python3.9/site-packages/torchvision/models/mobilenetv2.pyr! s
z _DeprecatedConvBNAct.__init__)__name__
__module__ __qualname__r!
__classcell__r' r' r% r( r s r c sL e Zd Zd eeeeeedejf dd fdd
Ze e dddZ
Z S )
InvertedResidualN.)inpoupstride expand_ratior returnc s t || _|dvr&td| |d u r4tj}tt|| }| jdkoT||k| _g }|dkr~| t
||d|tj d | t
|||||tj dtj
||ddddd||g tj| | _|| _|dk| _d S ) N)r
r z"stride should be 1 or 2 insted of r
kernel_sizer r )r0 groupsr r r F)bias)r r! r0
ValueErrorr r introunduse_res_connectappendr
r extendConv2d
Sequentialconv out_channelsZ_is_cn)r" r. r/ r0 r1 r
hidden_dimlayersr% r' r( r! ( s:
zInvertedResidual.__init__xr2 c C s" | j r|| | S | |S d S N)r: r? r" rD r' r' r( forwardP s zInvertedResidual.forward)N) r) r* r+ r8 r r r Moduler! r rG r, r' r' r% r( r- ' s (r- c
sz e Zd Zdeeeeee eeedej f eedej f edd fdd
Z
e e d
d d Z e e d
d
dZ
ZS )r ?N 皙?.) num_classes
width_multinverted_residual_setting
round_nearestblockr dropoutr2 c s" t t| |du r t}|du r,tj}d}d} |du rjg dg dg dg dg dg d g d
g}t|d kst|d d krtd
| t|| |}t| t d| || _
t d|d|tj dg}
|D ]V\} } }
}t| | |}t
|
D ]2}|d kr|nd}|
||||| |d |}qq|
t || j
d|tj d tj|
| _ttj|dt| j
|| _| D ]}t|tjrtjj|jdd |jdurtj|j n`t|tjtj frtj |j tj|j n.t|tjrztj |jd d tj|j qzdS )aw
MobileNet V2 main class
Args:
num_classes (int): Number of classes
width_mult (float): Width multiplier - adjusts number of channels in each layer by this amount
inverted_residual_setting: Network structure
round_nearest (int): Round the number of channels in each layer to be a multiple of this number
Set to 1 to turn off rounding
block: Module specifying inverted residual building block for mobilenet
norm_layer: Module specifying the normalization layer to use
dropout (float): The droupout probability
N i )r
r
r
) r r )rU rS r )rU @ r )rU ` rW r
)rU rW r )rU i@ r
r
r rY zGinverted_residual_setting should be non-empty or a 4-element list, got rJ rW r )r0 r r r
)r1 r r3 )pfan_out)modeg{Gz?) r r! r r- r r lenr7 r max last_channelr
r ranger; r> featuresDropoutLinear
classifiermodules
isinstancer= initkaiming_normal_weightr6 zeros_ GroupNormones_normal_)r" rM rN rO rP rQ r rR
input_channelra rc tcnsoutput_channelir0 mr% r' r( r! X sh
zMobileNetV2.__init__rC c C s2 | |}tj|d}t|d}| |}|S )Nr
r
r
)rc r
functionaladaptive_avg_pool2dtorchflattenrf rF r' r' r(
_forward_impl s
zMobileNetV2._forward_implc C s
| |S rE )r} rF r' r' r( rG s zMobileNetV2.forward)rI rJ NrK NNrL )r) r* r+ r8 floatr r r r rH r! r r} rG r, r' r' r% r( r W s& _
r iz5 rx )
num_paramsmin_size
categoriesc @ sr e Zd Zedeeddi edddddid d
d Zed eedd
di edddddidd
d ZeZ dS )r z=https://download.pytorch.org/models/mobilenet_v2-b0353104.pth ) crop_sizezQhttps://github.com/pytorch/vision/tree/main/references/classification#mobilenetv2z ImageNet-1Kgx&1Q@gMV@)zacc@1zacc@5zXThese weights reproduce closely the results of the paper using a simple training recipe.)recipe_metrics_docs)url
transformsmetaz=https://download.pytorch.org/models/mobilenet_v2-7ebf99e0.pth )r resize_sizezHhttps://github.com/pytorch/vision/issues/3995#new-recipe-with-reg-tuningg`" R@gS㥛V@a$
These weights improve upon the results of the original paper by using a modified version of TorchVision's
`new training recipe
`_.
N)
r) r* r+ r r r _COMMON_META
IMAGENET1K_V1
IMAGENET1K_V2DEFAULTr' r' r' r( r s>
r
pretrained)weightsT)r progress)r r r$ r2 c K sT t | } | dur(t|dt| jd tf i |}| durP|| j|d |S )a MobileNetV2 architecture from the `MobileNetV2: Inverted Residuals and Linear
Bottlenecks `_ paper.
Args:
weights (:class:`~torchvision.models.MobileNet_V2_Weights`, optional): The
pretrained weights to use. See
:class:`~torchvision.models.MobileNet_V2_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.mobilenetv2.MobileNetV2``
base class. Please refer to the `source code
`_
for more details about this class.
.. autoclass:: torchvision.models.MobileNet_V2_Weights
:members:
NrM r )r )r verifyr r_ r r load_state_dictget_state_dict)r r r$ modelr' r' r( r s
r )
_ModelURLs))r functoolsr typingr r r r r{ r r Zops.miscr
Ztransforms._presetsr utilsr _apir r _metar _utilsr r r __all__r Z
ConvBNReLUZConvBNActivationrH r- r r r r boolr r r
model_urlsr' r' r' r( sB 0o&
$