I am trying to use Detectron2 and almost done. Detection is going well now but training gets this Error:
RuntimeError: CUDA out of memory. Tried to allocate 12.00 MiB (GPU 0; 1.95 GiB total capacity; 1.30 GiB already allocated; 1.56 MiB free; 1.34 GiB reserved in total by PyTorch)
Conda ( python 3.7.6)
PyTorch 1.5.0a0+e70c288 (compiled from source)
Detectron2 (latest version)
Anyone can solve it? Please bid with one sentence showing your idea.
I want to choose one and share my pc for him using AnyDesk or TeamView.
4 freelancers are bidding on average $106 for this job
hi the problem can be tackled mainly by reducing the batch size or the size of images being used. Resizing images to same size or setting it in config may solve this.