ValueError: Target size (torch.Size([128, 1])) must be the same as input size (torch.Size([256, 1]))
fly2tortoise opened this issue · 13 comments
Hello, distinguished researcher
Your research is very interesting and contributes greatly to evolutionary algorithms. But I have some problems when I reproduce your code, can you help me?
Personally, I guess I forgot to modify a parameter. It is ok for me to reproduce EGAN before.
Looking forward to your reply and best wishes
Thank you for interesting in our work.
Just as you guesting, you may not correctly modify the related hyperparameter.
Please carefully check your codes.
Best!
Shiming
Distinguished researcher, your code may still have some problems, and I'm still not working properly. but
I have the same problem, have you solved it?
I have the same problem, have you solved it?
It's a better solution, I forget the specifics, it seems to require deleting part of the Tensor list, I'll dig through the code sometime.
It's a better solution, I forget the specifics, it seems to require deleting part of the Tensor list, I'll dig through the code sometime.
wait your answer
It's a better solution, I forget the specifics, it seems to require deleting part of the Tensor list, I'll dig through the code sometime.
wait your answer
I changed one line of code and it works fine, it's in models/cdegan_model.py,about 150 lines or so
It's a better solution, I forget the specifics, it seems to require deleting part of the Tensor list, I'll dig through the code sometime.
wait your answer
I changed one line of code and it works fine, it's in models/cdegan_model.py,about 150 lines or so
in options/train_options.py,add bs ,it work!
parser.add_argument('--bs', type=int, default=2048, help='dataset fragment')
Thank you for interesting in our work. Just as you guesting, you may not correctly modify the related hyperparameter. Please carefully check your codes.
Best! Shiming
I would like to ask you one more question. Could you repeat the experiment? Don't know where the code is
Thank you for interesting in our work. Just as you guesting, you may not correctly modify the related hyperparameter. Please carefully check your codes.
Best! Shiming
I would like to ask you one more question. Can you calculate the fid of TTUR?
Yes, CDE-GAN is a very good paper. Please configure your environment strictly according to the settings given by the author. Use tensorflow 1.15 for computation.
qq272574497 , please install and configure according to the following environment, if you can't run TTUR, the probability is that it's a problem with the TensorFlow version as well as CUDA, thanks.
PyTorch 1.3.0
TensorFlow 1.15.0
cuda 10.0
qq272574497 , please install and configure according to the following environment, if you can't run TTUR, the probability is that it's a problem with the TensorFlow version as well as CUDA, thanks. PyTorch 1.3.0 TensorFlow 1.15.0 cuda 10.0
cuda 11.7 TensorFlow 2.X PyTorch 1.7 it work
Thank you for interesting in our work. Just as you guesting, you may not correctly modify the related hyperparameter. Please carefully check your codes.
Best! Shiming
I would like to ask you one more question. Could you repeat the experiment? Don't know where the code is