search

tangbinh / question-answering

30 stars 3 forks source link

CUDA out of memory #2

Open mady143 opened 4 years ago

mady143 commented 4 years ago

Hi @tangbinh , while i am training the BIDAF i.e., python train.py --arch bidaf --embed-path wordvec/glove/glove.840B.300d.txt --checkpoint-dir checkpoints/bidaf --log-file logs --lr 0.01 getting an error like RuntimeError: CUDA out of memory. Tried to allocate 4.27 GiB (GPU 0; 10.92 GiB total capacity; 5.10 GiB already allocated; 1.00 GiB free; 9.09 GiB reserved in total by PyTorch)

but i am inruuning in a gpu system having python 3.6 11.gb graphic card 32 gb ram and Nvidia support also i was unable to understand why this error is getting can you explain me clearly

Thank you , Manikantha Sekhar

tangbinh commented 4 years ago

@mady143 The RuntimeError indicates you're out of GPU memory. You can try to reduce the batch size.