Closed Data-drone closed 5 years ago
You're running out of CUDA memory. Either you have other open notebooks taking up the GPU RAM, or you have too little to start with. Use nvidia-smi
to debug this.
If you still can't figure it out post in https://forums.fast.ai/c/part1-v3, including the output of:
python -c 'import fastai; fastai.show_install(1)'
I think it was actually a cuda 9.1 issue. I moved to cuda 9.2 and now it seems fine
I have the same question with you. My cuda version is 9.0. When I run the pytorch example code of MNIST, everything is ok. But when I run the pytorch example code of imagenet. this problem appears. Can you tell me how to solve this problem. Thanks.
@yang-yk, please post at forums, each lesson has its own thread where you can ask for help, describing specifically what doesn't work, showing your code if it's custom one or which notebook you have a difficulty with if it's unmodified. And including your setup information (3 comments up).
When trying to run Lesson 7 superres
in the train section gives:
this is with
1.0.37
has anyone else gotten this error?