A 7B model trained on a new 322K indoor spatial QA dataset achieves the best average VSI-Bench score, but not on all eight tasks as claimed.
Did I miss any? Did I double-count? The identified items seem distinct and fit the definition of cabinets/storage furniture
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Visuospatial Cognitive Assistant
A 7B model trained on a new 322K indoor spatial QA dataset achieves the best average VSI-Bench score, but not on all eight tasks as claimed.