erogol / resnet.torch Goto Github PK

Thanks for the work so far, but I was curious if you could add an example on how to perform regression using resnet.

I don't see how you can specify a target (as the original resnet is folder based for the classes).

Is it possible to finetune an imagenet-trained network using regression?

May you also add AlexNet for model comparison?

the original nnlr can't support the Adam optimizer but only SGD, see https://github.com/gpleiss/nnlr/issues/8, your modified nnlr package whether supports the Adam?
In addition, I want to fine-tune the BatchNormalization layers in the residual unit and freeze others, how to set learning rates to each layer in the residual unit?

When I try to convert the final nn.linear(nFeatures, 1000) layer be convolutional (in order to help me handle multi-dimensional inputs later on), I get a bunch of errors. Is it possible to change it only in the model file (which would mean that I'm messing up my input sizes), or are there additional changes that I'd need to make to the code to make it work?

For reference, I am replacing it with model:add(Convolution(nFeatures,1000, 3,3,1,1)) or would I need to do