VIVAS: Vitalizing Visual Perception in VLM Pre-training via Vision-language Unified Autoregressive Supervision
Open paper
Benchmark
Model
Rank
Results
computer-use-agents-on-screenspot-pro
VIVAS
–
Accuracy: 57.1
multimodal-reasoning-on-visulogic
VIVAS
–
Accuracy: 23.1
vision-language-understanding-on-mmstar
VIVAS
–
Accuracy: 70.4
Rank counts only results with a code link.