basecls.models.mbnet#

MobileNet Series

MobileNetV1: “MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications”

MobileNetV2: “MobileNetV2: Inverted Residuals and Linear Bottlenecks”

MobileNetV3: “Searching for MobileNetV3”

引用

rwightman/pytorch-image-models rwightman/pytorch-image-models

class basecls.models.mbnet.MBConv(w_in, w_out, stride, kernel, exp_r, se_r, se_from_exp, se_act_name, se_approx, se_rd_fn, has_proj_act, has_skip, drop_path_prob, norm_name, act_name)[源代码]#

基类：Module

Mobile inverted bottleneck block with SE.

Version

Expansion

DWConv

SE

PWConv

OutAct

Skip

basic

[EXP, BN, AF]

[kxk_DW, BN, AF]

[SE]

[1x1_Conv, BN]

[ AF]

[Skip]

V1

[3x3_DW, BN, ReLU6]

[1x1_Conv, BN]

[ReLU6]

V2

[EXP, BN, ReLU6]

[3x3_DW, BN, ReLU6]

[1x1_Conv, BN]

[Skip]

V3

[EXP, BN, AF]

[kxk_DW, BN, AF]

[SE]

[1x1_Conv, BN]

[Skip]
参数

  • w_in (int) – input width.

  • w_out (int) – output width.

  • stride (int) – stride of depthwise conv.

  • kernel (int) – kernel of depthwise conv.

  • exp_r (float) – expansion ratio.

  • se_r (float) – SE ratio.

  • se_from_exp (bool) – calculate SE channels from expanded (mid) channels.

  • se_act_name (str) – activation function for SE.

  • se_approx (bool) – whether approximated sigmoid function (HSigmoid).

  • se_rd_fn (Callable) – SE round channel function.

  • has_proj_act (bool) – whether apply activation to output.

  • has_skip (bool) – whether apply skip connection.

  • drop_path_prob (float) – drop path probability.

  • norm_name (str) – normalization function.

  • act_name (str) – activation function.

forward(x)[源代码]#
class basecls.models.mbnet.MBStage(w_in, w_out, stride, depth, exp_r, drop_path_prob, **kwargs)[源代码]#

基类：Module

MBNet stage (sequence of blocks w/ the same output shape).

forward(x)[源代码]#
class basecls.models.mbnet.MBNet(stem_w, depths, widths, strides, kernels, exp_rs=1.0, se_rs=0.0, stage_act_names=None, has_proj_act=False, has_skip=True, drop_path_prob=0.0, width_mult=1.0, norm_name='BN', act_name='relu6', head=None)[源代码]#

基类：Module

MobileNet model.

参数

  • stem_w (int) – stem width.

  • depths (Sequence[int]) – depth for each stage (number of blocks in the stage).

  • widths (Sequence[int]) – width for each stage (width of each block in the stage).

  • strides (Sequence[int]) – strides for each stage (applies to the first block of each stage).

  • kernels (Sequence[int]) – kernel sizes for each stage.

  • exp_rs (Union[float, Sequence[Union[float, Sequence[float]]]]) – expansion ratios for MobileNet basic blocks in each stage. Default: 1.0

  • se_rs (Union[float, Sequence[Union[float, Sequence[float]]]]) – Squeeze-and-Excitation (SE) ratios. Default: 0.0

  • stage_act_names (Optional[Sequence[str]]) – activation function for stages. Default: None

  • has_proj_act (bool) – whether apply activation to output. Default: False

  • has_skip (bool) – whether apply skip connection. Default: True

  • drop_path_prob (float) – drop path probability. Default: 0.0

  • width_mult (float) – width multiplier. Default: 1.0

  • norm_name (str) – normalization function. Default: "BN"

  • act_name (str) – activation function. Default: "relu6"

  • head (Optional[Mapping[str, Any]]) – head args. Default: None

forward(x)[源代码]#