Yu, Guangye
|
35e44067c4
|
Add unified memory APIs for torch.accelerator (#152932)
# Motivation
The following API will be put under torch.accelerator
- empty_cache
- max_memory_allocated
- max_memory_reserved
- memory_allocated
- memory_reserved
- memory_stats
- reset_accumulated_memory_stats
- reset_peak_memory_stats
Pull Request resolved: https://github.com/pytorch/pytorch/pull/152932
Approved by: https://github.com/albanD
ghstack dependencies: #138222
|
2025-06-24 07:57:48 +00:00 |
|