pulsatrix 1.1.0
Loading...
Searching...
No Matches
pulsatrix::VGGConfig Struct Reference

A torchvision VGG without BatchNorm (VGG11 to VGG19; VGG16 is the one tested on the published weights). More...

#include <vision_models.hpp>

Public Attributes

std::vector< int64_t > features = {64, 64, 0, 128, 128, 0, 256, 256, 256, 0, 512, 512, 512, 0, 512, 512, 512, 0}
 The convolution widths in order, 0 for a 2x2 max pool: torchvision's config "D" for VGG16.
 
int64_t pool_size = 7
 The adaptive average pool's output size (7 in torchvision).
 
int64_t hidden = 4096
 The two hidden classifier widths (4096 in torchvision).
 
int64_t num_classes = 1000
 
int64_t in_channels = 3
 

Detailed Description

A torchvision VGG without BatchNorm (VGG11 to VGG19; VGG16 is the one tested on the published weights).

Member Data Documentation

◆ features

std::vector<int64_t> pulsatrix::VGGConfig::features = {64, 64, 0, 128, 128, 0, 256, 256, 256, 0, 512, 512, 512, 0, 512, 512, 512, 0}

The convolution widths in order, 0 for a 2x2 max pool: torchvision's config "D" for VGG16.

◆ hidden

int64_t pulsatrix::VGGConfig::hidden = 4096

The two hidden classifier widths (4096 in torchvision).

◆ in_channels

int64_t pulsatrix::VGGConfig::in_channels = 3

◆ num_classes

int64_t pulsatrix::VGGConfig::num_classes = 1000

◆ pool_size

int64_t pulsatrix::VGGConfig::pool_size = 7

The adaptive average pool's output size (7 in torchvision).


The documentation for this struct was generated from the following file: