VLAQuantBench: Closed-Loop Evaluation of Post-Training Quantization for Vision-Language-Action Models

T1Covered by 1 source · first reported by arXiv at 24 Sept, 04:00 UTC

arXiv:2609.25376v1 Announce Type: cross Abstract: Post-training quantization reduces the memory requirements of vision-language-action (VLA) models, but precision selection must account for the interaction between layer scope, numerical format, and calibration. We introduce \textbf{VLAQuantBench}…

science

Sources

• T1VLAQuantBench: Closed-Loop Evaluation of Post-Training Quantization for Vision-Language-Action ModelsarXiv24 Sept, 04:00