8000 Cuda Out of Memory! · Issue #13 · mlpc-ucsd/BLIVA · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
Cuda Out of Memory! #13
Open
Open
@Ethantequila

Description

@Ethantequila

Hi I try to run a demo(BLIVA_Vicuna 7B) in my local machine(V100, 16GB), It comes with a OOM error:
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 86.00 MiB (GPU 0; 15.77 GiB total capacity; 12.18 GiB already allocated; 54.88 MiB free; 12.38 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

So how much is the minimum amount of GPU memory required ? Or is there any way to reduce GPU usage?

Thanks a lot!

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions

      0