A significant memory leak has been identified in the current workflow. System RAM (or VRAM) is not being released after an inference task is completed. Repeated usage leads to memory exhaustion (OOM), suggesting that objects or tensors remain cached in memory instead of being deallocated.
A significant memory leak has been identified in the current workflow. System RAM (or VRAM) is not being released after an inference task is completed. Repeated usage leads to memory exhaustion (OOM), suggesting that objects or tensors remain cached in memory instead of being deallocated.