Update hpc_training_rules.adoc

mlcommons / training_policies

Issues related to MLPerf™ training policies, including rules and suggested changes

https://mlcommons.org/en/groups/training

Apache License 2.0

93 stars 66 forks source link

Update hpc_training_rules.adoc #478

Closed nvaprodromou closed 2 years ago

nvaprodromou commented 2 years ago

Added policy snippet that defines allowed max scale for resubmission of weakly-scaled results caused by HP borrowing. This was a point of discussion during MLPerf HPC v0.7 review period.

Rule can be removed when submitters are allowed to submit pruned log files in their submission to prove availability of claimed max scale.

github-actions[bot] commented 2 years ago

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

sparticlesteve commented 2 years ago

This PR was approved by the HPC WG in the group meeting on May 23, 2022.