Error Encountered When Training Neural Network

AI Discussions
1

Hello

I’ve been facing an error while training a neural network for a computer vision project, and I’m asking for help to resolve the issue. During the training process, I consistently receive the following error message: ‘RuntimeError: CUDA out of memory’. I’ve tried to address this by reducing batch sizes and optimizing memory usage, but the error still exist .

Has anyone else faced a similar ‘CUDA out of memory’ error when training neural networks on GPUs? If yes , I would greatly appreciate recommendations on how to effectively manage GPU memory usage and resolve this issue.

Thank you in advance for your help.

Best regards,
stevediaz

2

Try to use Colab and their resources and lets see what happens with that error!