Dario Serez

Hi everyone, thank you for the amazing job on this repository. I'm trying to replicate @kobiso work on CUB200 dataset (#131) on a 2 GPU architecture. I'm not directly using...

Hi, thank you for the repo. I am trying to add the module to my code, but I noticed that the used GPU memory increases a lot, due to the...

Expectation Over Transformation Wrapper

2

Closes #563 and #191 My approach for implementing [EoT](https://arxiv.org/pdf/1707.07397.pdf). I tried to follow the suggestions of @zimmerrol. Basically, I created a new ExpectationOverTransformation Wrapper class that computes the forward pass...

Hi, thank you for the amazing repo. Would it be feasible to add a BigGan trained on ImageNet with the method proposed in: https://arxiv.org/abs/1907.02544 The GAN comes with an extra...

Large .beton files slow down or even freeze learning during loading [possible bug]

1

Hello, I have created two different versions of the imagenet .beton file. The code is the following: ``` from torch.utils.data import Dataset from PIL import Image from ffcv.fields import RGBImageField...

Hi, I am trying to reproduce your work on the Imagenet1000 case. Following your pipeline, I can reproduce the results on the Real case (training SimClr on Imagenet). However, when...

Hi, Thank you for the repo, it is very useful! For my current work, I would like to get also the pre-trained Discriminators from the GAN models that I load....

DLR Loss and Zero Gradient

1

Hi, I am trying to understand how DLR loss works, and in particular, I am running some tests with Batch Size = 1. The DLR loss defined in https://github.com/fra31/auto-attack/blob/master/autoattack/autopgd_base.py is...