"out of memory" cuda error and required GPU

[vutiqibyl]

Hello,

When launching train_all.sh and using a GTX 980Ti 6GB (on Ubuntu 17.10), I encounter a "Check failed: error == cudaSuccess (2 vs. 0) out of memory". How much GPU memory is required for the training ?

[pongsate1]

Around 10GB. When I was doing this, I need to use Titan X.

Thank you for your interest in my work.

On Wed, Mar 28, 2018, 16:56 vutiqibyl [email protected] wrote:

Hello,

When launching train_all.sh and using a GTX 980Ti 6GB (on Ubuntu 17.10), I encounter a "Check failed: error == cudaSuccess (2 vs. 0) out of memory". How much GPU memory is required for the training ?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/pongsate1/fashion-parsing/issues/7, or mute the thread https://github.com/notifications/unsubscribe-auth/AGrDTkD5tASlU8_HfHjmuMWJCrH0KCptks5ti0IogaJpZM4S-LTh .

[vutiqibyl]

Thank you for your quick answer.

[vutiqibyl]

Hello again,

Now using a Titan X 12GB, I encounter the error again.

Training the models fcn-32s, fcn-16s, and fcn-8s works as intended, but when training segc-8s-pre I get a "Check failed: error == cudaSuccess (2 vs. 0) out of memory" error. Would you know what might cause this ?

Here is the whole output : fp_error_output.txt