aanoosheh/combogan

combogan's Issues

Dataset Error 404: not found

Hi, thanks for your work! I have tried to download datasets by download_dataset.sh scripts and wget but failed.
I have noticed issue3 and 2. But it still doesn't work.

The dataset is unavailable

Need help,your dataset is not available

Objective function for four domains

Hi, thanks for such a nice work.

In your paper, you have mentioned the objective of CycleGAN which is:

This is the objective for two domains only. As your work in ComboGAN can translate images from multiple domains, then if you have four domains, what will be the objective function for four domains?

Explanation for forward loss

Thanks for such a nice and clean code. Can you please explain what lambda_fwd is intended to do?

https://github.com/AAnoosheh/ComboGAN/blob/master/models/combogan_model.py#L155

Thanks.

--netG_n_shared flag

Hi, I was wondering what exactly this --netG_n_shared flag is for ?
What I understood is that these are the Resblocks shared by all encoders (they help bring the different images into the same shared latent space) right?

The problem is that when I change the default value (zero) of the flag I get a Cuda issue:

RuntimeError: chunk expects at least a 1-dimensional tensor (chunk at /pytorch/aten/src/ATen/native/TensorShape.cpp:186)
frame #0: std::function<std::string ()>::operator()() const + 0x11 (0x7f44f02bb441 in /usr/lib/python3.7/site-packages/torch/lib/libc10.so)
frame #1: c10::Error::Error(c10::SourceLocation, std::string const&) + 0x2a (0x7f44f02bad7a in /usr/lib/python3.7/site-packages/torch/lib/libc10.so)
frame #2: at::native::chunk(at::Tensor const&, long, long) + 0x1f3 (0x7f449c7dbf93 in /usr/lib/python3.7/site-packages/torch/lib/libcaffe2.so)
frame #3: at::TypeDefault::chunk(at::Tensor const&, long, long) const + 0x12 (0x7f449ca3a442 in /usr/lib/python3.7/site-packages/torch/lib/libcaffe2.so)
frame #4: torch::autograd::VariableType::chunk(at::Tensor const&, long, long) const + 0x282 (0x7f449ae270c2 in /usr/lib/python3.7/site-packages/torch/lib/libtorch.so.1)
frame #5: torch::cuda::scatter(at::Tensor const&, c10::ArrayRef, c10::optional<std::vector<long, std::allocator > > const&, long, c10::optional<std::vector<c10::optionalc10::cuda::CUDAStream, std::allocator<c10::optionalc10::cuda::CUDAStream > > > const&) + 0x5ad (0x7f449b38a1fd in /usr/lib/python3.7/site-packages/torch/lib/libtorch.so.1)
frame #6: + 0x5a41cf (0x7f44f0a781cf in /usr/lib/python3.7/site-packages/torch/lib/libtorch_python.so)
frame #7: + 0x130fac (0x7f44f0604fac in /usr/lib/python3.7/site-packages/torch/lib/libtorch_python.so)

frame #15: THPFunction_apply(_object*, _object*) + 0x6b1 (0x7f44f0888301 in /usr/lib/python3.7/site-packages/torch/lib/libtorch_python.so)

Test problem

There is something wrong when I try to test my own dataset, I've already trained my models and now I want to generate a new dataset so I write this command in my terminal :python test.py --phase test --name dimmer --dataroot ./datasets/stanfordcars/test0_day --n_domains 2 --which_epoch 195 --serial_test

GAN architecture

Hi,
Thank you for your work.
In GAn architecture, can you explain where the skip connections are for the residual blocks?
Thanks in advance.

your dataset is unavailable

Hi, thank you for your share. When i want to get the dataset using the ''download_dataset.sh', I found a '403 forbidden' problem.

datasets

Can you share your datasets？ thanks

'https://data.vision.ee.ethz.ch/cvl/combogan/$FILE.zip' is useless

dataset is unavailable

When i use .sh to download the dataset, the connection to the target url is always timed out. So how can i get the dataset?

Testing all Images from Test Folder

Hi,

I have 1000 images in each test folders. I need to generate images from all the test images. I have 3 domains. So, I need to generate 1000 * 3 = 3000 images from one test folders.
Please let me know how can I do that?

The number of testing images?

Thanks for you sharing your excellent codes. But a problem confused me.
The Alps Seasons dataset has 400 testing images, however, after doing testing, I obtained 1240 images in folder "image", why is not 400*5=2000 images generated rather 1240 images?

training hours

Hi, in your paper, you mentioned that "The fourteen painters dataset, for example, ran 1400 epochs in 220 hours on our nVidia Titan X GPU. Note that pairwise CycleGAN instead would have taken about 2860 hours, or four months."

How you calculated the training hours, i.e. 2860 hours for CycleGAN?