kleinyoni comments

Repositories
Issues
Comments

Results 5 comments of


                                            kleinyoni

I met come errors in trainer.py

add module like this G.module.grow_network This worked for me

RuntimeError: CUDA error: out of memory (Is there any problem with pytorch)

How much memory do you have in your GPU? I think 12 should be OK.

RuntimeError: CUDA error: out of memory (Is there any problem with pytorch)

@AlexanderKozhevin alI think a single card. Since the primary data loading is done to one GPU(i think. not sure). Try to use --ngf 32 instead. It should use less memory....

about train code

Hi, great paper. When do you think the training code will be released?

Different backbone

Thanks for the fast reply. Regarding 2, I don't need to change the Anchor size even if the output resolution of the backbone is different? The Detection models are taking...