RuntimeError: CUDA out of memory. Tried to allocate 2.00 MiB (GPU 0; 14.76 GiB total capacity; 13.59 GiB already allocated; 3.75 MiB free; 13.59 GiB reserved in total by PyTorch

enghock1 / Real-Time-2D-and-3D-Hand-Pose-Estimation

Real-Time 2D and 3D Hand Pose Estimation from RGB Image

GNU General Public License v2.0

48 stars 15 forks source link

RuntimeError: CUDA out of memory. Tried to allocate 2.00 MiB (GPU 0; 14.76 GiB total capacity; 13.59 GiB already allocated; 3.75 MiB free; 13.59 GiB reserved in total by PyTorch #10

Closed KhaLee2307 closed 2 years ago

KhaLee2307 commented 2 years ago

Hi, thank you for your interesting project. I have a question for you. I ran the hand tracking on both google colab (15GiB GPU) and my local computer (4GiB GPU), but in both cases ran out of memory. I think this is because your code used so much GPU but don't free it. So do you have any solutions to solve this problem? I would appreciate it if you reply me immediately.

KhaLee2307 commented 2 years ago

I just found the solution. Thanks!