search

MozarTuring / badgan

1 stars 0 forks source link

I have a question about exceeding the graphics memory #1

Open XiaoHao-Chen opened 5 years ago

XiaoHao-Chen commented 5 years ago

First of all, thank you very much for your help. When I ran the mnist_trainer.py you provided, there was a problem beyond the memory of the video, at which point my GPU was 8G. The problem is still on the pixelcnn.py , in the for loop inside the code, constantly calling some models defined by bad gan himself, such as down_shifted_conv2d,down_shifted_deconv2d, and so on, Then the amount of GPU did not drop after the for loop ended, so I wondered if this part of the memory had not been released. I tried to free up this part of the memory, but it failed. I would like to ask you what the Python and Pytorch versions were when you ran the program? Would you mind posting the mnist_trainer.py related code? Thank you again for your help!

MozarTuring commented 5 years ago

updated, please check

XiaoHao-Chen commented 5 years ago

Thank you

XiaoHao-Chen commented 5 years ago

Thank you very much for your help, the code has been successfully run! Thanks again!

XiaoHao-Chen commented 5 years ago

Hello, I have a question to ask. Do you have any good Suggestions for training pixelcnn in this program? Because I want to train my data set.