Tao

icon indicating a website link https://www.cnblogs.com/txu2k8/ icon indicating an email [email protected]

icon location china icon location Software QA engineer - Python/Go, DevQA, Automation, Storage, Stress, Performance.

Results 2 comments of Tao

Assigning different accelerator count per host does not work

We meet the same error when running checkpoting 70B model with 64 progresses. We have limited hardware hosts and each host has only 256GB memory(required total 912GB total). it's very...

Guidance on Implementing CPU Binding for UNet3D (PyTorch) with MPI

Hi @zhenghh04 ,the checkpointing datasize subcommand has no "args.mpi_params",Bug it used in function update_args: cli.py ` ## Line 330: def add_checkpointing_arguments(checkpointing_parsers): .... if _parser == run_benchmark: _parser.add_argument('--exec-type', '-et', type=EXEC_TYPE, choices=list(EXEC_TYPE),...