CUDA out of memory

[BobH62]

I tried to run the vanillaGS Renderer, but if I execute the client.py, the main.py gives:

(gaussian_splatting) root@0343a59430d5:/workspace/gaussian-splatting# python /src/main.py
Gaussian Splatting renderer ready for requests...
R
[[ 1. 0. 0.]
[ 0. -1. 0.]
[ 0. 0. -1.]]
T [0. 0. 1.] | |T| = 1.0
num_gaussians = 500036
camera 1280 720
FoVx (deg) = 59.99999999999999
Error during rendering: CUDA out of memory. Tried to allocate 130814.35 GiB. GPU 0 has a total capacity of 15.45 GiB of which 13.76 GiB is free. Including non-PyTorch memory, this process has 646.00 MiB memory in use. Of the allocated memory 299.44 MiB is allocated by PyTorch, and 10.56 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)

Could you have any ideas what's wrong?

[j3soon]

Is this error from the poster example in the README? Or are you using a custom 3DGS model?

I think this error may be caused by a large/incorrect 3DGS model that cannot fit inside your 16 GB VRAM.

Could you also provide a screenshot of the error?