search

NVlabs / CF-3DGS

Other
394 stars 43 forks source link

OOM error always occurred. #33

Open Yuhuoo opened 3 months ago

Yuhuoo commented 3 months ago

Why does the CF-3DGS always report the cuda out of memory error under co3d scene? I have tried several scenes. In addition, I can't find the co3d scene presented in your paper at the link https://ucsdcloud-my.sharepoint.com/:u:/g/personal/yafu_ucsd_edu/EftJV9Xpn0hNjmOiGKZuzyIBW5j6hAVEGhewc8aUcFShEA?e=x1aXVx you provided.

  File "/data1/GaoAo/software/anaconda3/envs/cf3dgs/lib/python3.10/site-packages/diff_gaussian_rasterization/__init__.py", line 92, in forward
    num_rendered, color, depth, alpha, radii, geomBuffer, binningBuffer, imgBuffer = _C.rasterize_gaussians(*args)
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 17.15 GiB (GPU 0; 47.53 GiB total capacity; 17.37 GiB already allocated; 6.79 GiB free; 40.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
Training progress:  10%|██████▎                                                        | 100/1000 [00:23<03:30,  4.28it/s, PSNR=15.41, Number points=5805203]