Kiran Scaria comments

Results 2 comments of


                                            Kiran Scaria

help ! RuntimeError: CUDA out of memory. Tried to allocate 1.50 GiB (GPU 0; 10.92 GiB total capacity; 8.62 GiB already allocated; 1.39 GiB free; 8.81 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

Faced the same issues. Things that worked for me. 1. Load the half-model as suggested by @xmvlad [here](https://github.com/CompVis/stable-diffusion/issues/86#issuecomment-1230309710). 2. Disabling safety checker and invisible watermarking 3. reducing number of samples...

help ! RuntimeError: CUDA out of memory. Tried to allocate 1.50 GiB (GPU 0; 10.92 GiB total capacity; 8.62 GiB already allocated; 1.39 GiB free; 8.81 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

@rjamesnw After using the half precision model, have the GPU consumption to peak to ~12-13GB. To lower the GPU consumption further you can refer Issue: #95 You can also look...