DriveStack-VLA: Render-Teacher Alignment for BEV-Based DeepStack Vision-Language-Action Model
Open paper
Benchmark
Model
Rank
Results
bench2drive-on-bench2drive
DriveStack-VLA
–
Driving Score: 79.49
Success Rate: 56.36
navsim-on-navsim-v2-navtest
DriveStack-VLA
–
EPDMS: 91.0
navsim-on-openscene
DriveStack-VLA
–
PDMS: 91.6
Rank counts only results with a code link.