cst_captioning icon indicating copy to clipboard operation
cst_captioning copied to clipboard
verified publisher icon mynlp

Problem about multi processing

Open SilencerChen opened this issue 7 years ago • 1 comments

Thanks for your great work at first. I run you code and find out the usage rate of my GPU is always 0% when calculate the scores of the val captions. I train to use multiprocessing and need to add "num_works" in torch.utils.data.DataLoader but I find you write the class all by yourself. So is there a way to use the mutiprocessing?

SilencerChen avatar Mar 24 '18 13:03 SilencerChen

  • I think the COCO Caption Evaluation code does not support running on GPU.
  • You can implement a DataLoader class to support parallel loading; however, I don't think it improves the speed significantly since data loading is quite fast.
  • You may need to run on multiple GPUs; however, I haven't tried to implement this. Any contribution is welcome.

plsang avatar Mar 26 '18 03:03 plsang