A run error occurred after modifying the network layer！

[15732031137]

Hello! First of all, congratulations on your paper getting the best ICCV paper, thank you for your contribution to GAN network research, and share your code. Recently I was fortunate to study your paper and program. I did not have any problems when running your original code, but I added a layer of attention mechanism after the fourth layer in the generator and discriminator in the models.py file. I ran the picture twice and it was 500 and 250 respectively. Both times I got the following error. The 250 size picture is your own picture. Is this just because the GPU is not big enough? Thank you for reading in your busy schedule! I really want your advice! I wish you a happy life and youth forever!

File "main_train.py", line 29, in
train(opt, Gs, Zs, reals, NoiseAmp)
File "/root/Z-SinGAN-master/SinGAN/training.py", line 39, in train
z_curr,in_s,G_curr = train_single_scale(D_curr,G_curr,reals,Gs,Zs,in_s,NoiseAmp,opt)
File "/root/Z-SinGAN-master/SinGAN/training.py", line 162, in train_single_scale
gradient_penalty.backward()
File "/root/anaconda3/lib/python3.7/site-packages/torch/tensor.py", line 107, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph)
File "/root/anaconda3/lib/python3.7/site-packages/torch/autograd/init.py", line 93, in backward
allow_unreachable=True) # allow_unreachable flag
RuntimeError: CUDA out of memory. Tried to allocate 772.00 MiB (GPU 0; 11.17 GiB total capacity; 9.83 GiB already allocated; 410.81 MiB free; 648.70 MiB cached)

[markstrefford]

Yes it seems that the GPU doesn’t have enough memory for the new network architecture you are using.

…

On 28 Dec 2019, at 06:52, 15732031137 ***@***.***> wrote:  Hello! First of all, congratulations on your paper getting the best ICCV paper, thank you for your contribution to GAN network research, and share your code. Recently I was fortunate to study your paper and program. I did not have any problems when running your original code, but I added a layer of attention mechanism after the fourth layer in the generator and discriminator in the models.py file. I ran the picture twice and it was 500 and 250 respectively. Both times I got the following error. The 250 size picture is your own picture. Is this just because the GPU is not big enough? Thank you for reading in your busy schedule! I really want your advice! I wish you a happy life and youth forever! File "main_train.py", line 29, in train(opt, Gs, Zs, reals, NoiseAmp) File "/root/Z-SinGAN-master/SinGAN/training.py", line 39, in train z_curr,in_s,G_curr = train_single_scale(D_curr,G_curr,reals,Gs,Zs,in_s,NoiseAmp,opt) File "/root/Z-SinGAN-master/SinGAN/training.py", line 162, in train_single_scale gradient_penalty.backward() File "/root/anaconda3/lib/python3.7/site-packages/torch/tensor.py", line 107, in backward torch.autograd.backward(self, gradient, retain_graph, create_graph) File "/root/anaconda3/lib/python3.7/site-packages/torch/autograd/init.py", line 93, in backward allow_unreachable=True) # allow_unreachable flag RuntimeError: CUDA out of memory. Tried to allocate 772.00 MiB (GPU 0; 11.17 GiB total capacity; 9.83 GiB already allocated; 410.81 MiB free; 648.70 MiB cached) — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[15732031137]

Thank you for your reply and wish you a happy life!