Max Resolution

Hello, I am interested in using your inference code directly for image to image translation. I would like to use your method to edit images with a resolution of 1920 x 1080. I am using an NVIDIA GeForce RTX 4090, which I believe should handle that resolution. However, I am encountering a memory allocation error (torch.cuda.OutOfMemoryError: CUDA out of memory) whenever I set the resolution higher than 900. Additionally, I receive this warning: UserWarning: Failed to load the image Python extension: libtorch_cuda_cu.so: cannot open shared object file: No such file or directory, when running my code. Could this be related? I am curious if you can identify a problem based on the information provided. I tried simplifying the model (e.g., reducing the number of channels, attention resolutions, etc.), but the decrease in quality is too significant. Are there other strategies I should consider?