Where I can load the mode and then I can generate several images without issue. Then by after like 20 or so images, and I try to make another image with the same settings as all the ones that worked before, the program will crash.
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1024.00 MiB (GPU 0; 11.00 GiB total capacity; 8.21 GiB already allocated; 593.00 MiB free; 9.23 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
And then I need to restart my program / reload the model.
Is this expected behavior? why do the VRAM requirements go up over time? Shouldn't it be able to check how much is available before crashing the program?
I have a very repeatable issue.
Where I can load the mode and then I can generate several images without issue. Then by after like 20 or so images, and I try to make another image with the same settings as all the ones that worked before, the program will crash.
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1024.00 MiB (GPU 0; 11.00 GiB total capacity; 8.21 GiB already allocated; 593.00 MiB free; 9.23 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
And then I need to restart my program / reload the model.
Is this expected behavior? why do the VRAM requirements go up over time? Shouldn't it be able to check how much is available before crashing the program?