Understanding ResNeXt code

akashgshastri · October 13, 2019, 7:40pm

I found the resnext architecture code in the link below :

fastai/imagenet-fast/blob/master/cifar10/models/resnext.py

# https://raw.githubusercontent.com/miraclewkf/ResNeXt-PyTorch/master/resnext.py
'''
New for ResNeXt:
1. Wider bottleneck
2. Add group for conv2
'''

import torch.nn as nn
import math

__all__ = ['ResNeXt', 'resnext18', 'resnext34', 'resnext50', 'resnext101',
           'resnext152']

def conv3x3(in_planes, out_planes, stride=1):
    """3x3 convolution with padding"""
    return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
                     padding=1, bias=False)


class BasicBlock(nn.Module):

This file has been truncated. show original

can somebody give me a brief understanding of this code with reference to the paper terms and what they’re doing?

for example: is cardinality in the paper ‘inplanes’ in the code? and so on

thank you