warp-ctc icon indicating copy to clipboard operation
warp-ctc copied to clipboard

Published 20 hours ago • verified publisher icon SeanNaren

RuntimeError: CUDA out of memory. Tried to allocate 17179869113.53 GiB

Open vinayakarannil opened this issue 4 years ago • 1 comments

Getting out of memory error

File "/usr/local/lib/python2.7/dist-packages/torch/nn/modules/module.py", line 493, in call result = self.forward(*input, **kwargs) File "build/bdist.linux-x86_64/egg/warpctc_pytorch/init.py", line 82, in forward self.length_average, self.blank) File "build/bdist.linux-x86_64/egg/warpctc_pytorch/init.py", line 32, in forward blank) RuntimeError: CUDA out of memory. Tried to allocate 17179869113.53 GiB (GPU 0; 10.92 GiB total capacity; 36.04 MiB already allocated; 9.58 GiB free; 29.96 MiB cached) (malloc at /pytorch/c10/cuda/CUDACachingAllocator.cpp:267)

pytorch version-1.1 cuda version -10.1

vinayakarannil avatar Oct 03 '19 11:10 vinayakarannil

Encountered this myself. It looks similar to issue 118 see Sean's comment

dzubke avatar Jan 22 '20 18:01 dzubke