Nereus: Adaptive Parallelism for LLM Post-Training
Nereus adapts parallelism during LLM reinforcement-learning post-training, cutting step latency 27.7% versus a fixed layout.
1 stories, newest first · one card per story
Nereus adapts parallelism during LLM reinforcement-learning post-training, cutting step latency 27.7% versus a fixed layout.
Today 1.7K readers · 3.7K views · 2.8K reads · all time 161K views