torch-residual-networks

torch-residual-networks copied to clipboard

Update residual-layers.lua to be consistent with other torch nn models

2

Hi, I updated residual-layers.lua so that it can be used with other available layers in torch community. I thought this might be useful for someone who might want to use...

lim0606

Thank you for your code firstly. I have tried to train resnet-20 on cifar10 dataset with pytorch and I used ShortcutA as you did. But I can't get as good...

I-Doctor

Why are 2 convolutions used in every res-block in a resnet?

1

This is more a general question about resnets. I am wondering why there are always 2 convolutional layers used in every res-block instad of just one? To me just one...

mbcel

Hello everyone, I met a problem about unexpected output behaviour of batch normalisation layer or Cadd table. I would like to investigate the output of these layers with layer.output attribute,...

pszyu

What is your Machine (Runtime Env.) Configuration ? like number of GPUs, and some other configuration info. like true batchsize, etc. And another question is, in part of "Alternate training...

whatbeg

BatchNorm after ReLU

8

Hi, I am performing somehow similar benchmark, but on caffenet128 (and moving to ResNets now) on ImageNet. One thing, that I have found - the best position of BN in...

ducha-aiki

training data also normalized?

2

Hi G, I am so new to Torch, just a quick concern about fetching training data. The training data is supposed to be normalized, too. However, I see no such...

hli2020

Hi Guys I'm trying to classify images with the pre-loaded models but something is strange.... $ th classify.lua resnet-34.t7 ImgnetCat.jpg Classes for ImgnetCat.jpg 0.0075856610201299 bucket, pail 0.0070607061497867 hook, claw 0.0061263972893357...

leonardoaraujosantos

Loss function explode under default settings

4

Hi all, I'm training on CIFAR-10 following the instructions but I found that in most cases the loss function explodes during the first few iterations. This is how it behaved...

yangky11

still on the same GPU when iter size is above 1 ?

1

Hi, when iter_size is over 1, say 3, it still computes on one GPU, right? some confusing [here](https://github.com/gcr/torch-residual-networks/blob/master/train-cifar.lua#L201) in the code. I guess the default GPU_id is 1, aka the...

hli2020