keremturgutlu / deeplearning Goto Github PK

@KeremTurgutlu,
I ran your notebook and ran into an error

RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.FloatTensor) should be the same

I thought the error was becuse I was on pytorcg 0.3.1, I have now upgraded to pytorch version 0.4.0.
The error still occured.

Any guess why?

Many thanks

Hi,
I noticed an undefined function, "get_contoured_mask" is called in the PART 2 CONTOUR LABELS.ipynb and Stage 1 train + Stage 1 test .ipynb notebooks but doesn't appear to exist in the repo

Hi-

Thanks for posting this code! I am trying to get it working but an getting an out of memory error:
RuntimeError: cuda runtime error (2) : out of memory at d:\pytorch\pytorch\torch\lib\thc\generic/THCStorage.cu:58

on this step:
model = MixedInputModel(emb_szs, n_cont=0, emb_drop=0, out_sz=2, szs=[500], drops=[0.5]).cuda()

I tried reducing the max embedding size, but I still get the error. My machine has an older GPU with only 2Gb. I suspect that the 2Gb is a significant limitation, but I was hoping I could still getting it working.

Any ideas?

Kerem,
Can you help me understand how to define ae_model in Porto Seguro.ipynb?
When I reach:

learn = StructuredLearner(data, ae_model)

I get an error...ae_model is not defined.

Thank you in advance