I have this error, i have RTX 4090. ubuntu Ubuntu 22.04.1
RuntimeError: CUDA out of memory. Tried to allocate 3.20 GiB (GPU 0; 22.20 GiB total capacity; 18.38 GiB already allocated; 2.83 GiB free; 18.46 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
Ran into this as well. My issue was that --sample_batch_size defaults to 4, and when the script tries to generate class images, I run out of CUDA memory. Adding that param and setting it to 1-2 worked for me.
Describe the bug
I have this error, i have RTX 4090. ubuntu Ubuntu 22.04.1
RuntimeError: CUDA out of memory. Tried to allocate 3.20 GiB (GPU 0; 22.20 GiB total capacity; 18.38 GiB already allocated; 2.83 GiB free; 18.46 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
Command:
Reproduction
No response
Logs
No response
System Info
Copy-and-paste the text below in your GitHub issue and FILL OUT the two last points.
diffusers
version: 0.9.0