List your company

RLHF Training on Standard RLHF Workload (train)

99.84Mean Latency (s)OPPO
98.8192105.7096112.6119.4904Sep 30, 2025

Evaluation results

MethodLinks
OPPO2025.0999.84
AReaL2025.09109.92
VeRLParallelism Strategy=DP+SP2025.09120.47
VeRLParallelism Strategy=DP2025.09125.36