04:00
2026-08-17
machinebrief.com
machine-learning
Rollplex: Cross-Phase GPU Spatial Sharing for Vision Language Model Post-Training
Researchers introduced Rollplex, a runtime that enables cross-phase GPU spatial sharing for vision-language model post-training, achieving 1.23x–1.30x speedup over serial colocation and 1.57x–2.24x ov…