Closed CaiWang0503 closed 1 year ago
Hi, okay, that's weird Please try again with n_gpu = 1 (and device = "gpu") as argument in sjSDM_cv
it looks like a memory issue it said: "Your notebook tried to allocate more memory than is available" (it works well 2 month ago, so i am not sure if it's kaggle's problem)
but my dataset is very small (250 sites and 15 species), i have tried n_gpu = 1 and n_core=1, it still show the same issue. is that the new version's problem or just because kaggle not suit for it now. looks like i can't use it at the moment
does it run on the CPU?
it didn't use a lot of GPU like before, most of time i saw it only uses CPU, and GPU is used very occasionally...maybe is Kaggle's problem
Hi, I am running "SJSDM_CV" on Kaggle, but the monitor shows that it is not running with GPU, but "SJSDM" can run with GPU. I wonder if "SJSDM_CV" needs some time to load data and then enable GPU, or it's an error. (I already gave it the parameter device = "gpu"
)