LaZSL aligns local image patches with LLM-generated attribute descriptions using optimal transport, achieving slightly higher average zero-shot accuracy than prior interpretable CLIP baselines on nine benchmarks.
Robust calibration of large vision- language adapters
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
-
Interpretable Zero-Shot Learning with Locally-Aligned Vision-Language Model
LaZSL aligns local image patches with LLM-generated attribute descriptions using optimal transport, achieving slightly higher average zero-shot accuracy than prior interpretable CLIP baselines on nine benchmarks.