CycleGAN
CycleGAN copied to clipboard
junyanz
Software that can generate photos from paintings, turn horses into zebras, perform style transfer, and more.
you suggest to train our own datasets to create a csv file. i wanted to know the content of the csv file if we already specify in the dataset.py file...
I don't understand, is there any special role? thanks！
I am using my own dataset on Colab. I can train my dataset but when I need to test it. dataset [SingleDataset] was created initialize network with normal model [TestModel]...
Thanks for your great work. I want to reproduce the experiment of summer2winter and the dataset link you posted in README.md is invalid now. Could you please send me a...
At first let me thank you @junyanz for the massive work you have done on the project. I have a question, and I would like to apologize if that will...
Hello,I am the Beginner.I have a doubt about the initialization of the network for cyclegan. As we know,we create the network using the op ''InstanceNorm2d'',but why we initialize "BatchNorm2d" in...
Was going through exercises from https://pytorch.org/deep-learning-with-pytorch and found the following: > Search GitHub for projects that provide a hubconf.py file. > - How many repositories are returned? > - Find...
i have read your paper ,and can not figure out how to evaluted photo to label task on cityscapes dataset.what's your input into the FCN network?
Decent results require 3k-5k frames. My GPU session on colab gets disconnected due to usage while training. I am saving the checkpoints in the drive. Is there any way I...
I used the CycleGan to transfer the image of Yosemite in summer and winter, in 17093 iterations, the Fake image and reconstructed image were still blurred, should I wait for...
