Cannot run dorado correct without triggering a memory issue - using A100 GPUs with 80Gb memory. Error message is:
RuntimeError: CUDA out of memory. Tried to allocate 18.56 GiB (GPU 0; 79.15 GiB total capacity; 56.85 GiB already allocated; 16.23 GiB free; 62.28 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
Steps to reproduce the issue:
Try running dorado correct on a 90Gb+ ultra long run.
Issue Report
Please describe the issue:
Cannot run dorado correct without triggering a memory issue - using A100 GPUs with 80Gb memory. Error message is:
RuntimeError: CUDA out of memory. Tried to allocate 18.56 GiB (GPU 0; 79.15 GiB total capacity; 56.85 GiB already allocated; 16.23 GiB free; 62.28 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
Steps to reproduce the issue:
Try running dorado correct on a 90Gb+ ultra long run.
Run environment:
Logs
Full error message is: