Running your code on the server will appear: Segmentation fault (core dumped)

ShangtongZhang / DeepRL

Modularized Implementation of Deep RL Algorithms in PyTorch

MIT License

3.21k stars 684 forks source link

Running your code on the server will appear: Segmentation fault (core dumped) #60

Closed Kchu closed 5 years ago

Kchu commented 5 years ago

 Sorry to bother you, I am getting an error when submitting your code to the server: Segmentation fault (core dumped). There is no problem running on my stand-alone machine. I don't know why, please explain it.

ShangtongZhang commented 5 years ago

Did you use Docker?

Kchu commented 5 years ago

Sorry, I don't use docker, but our lab's server assigns everyone a virtual machine-like account.

ShangtongZhang commented 5 years ago

You should try docker

Kchu commented 5 years ago

I've found the problem. The Initialization function nn.init.orthogonal_ caused this question. If we use another function like nn.init.xaviernormal then there won't be this question and code works fine!