Fine-tunes LLaVA on up to 4k bridge images for damage description and priority scoring, with a two-stage SLM quality guard showing peak semantic similarity at 3k samples.
Few-shot1/aanomalies feed- back: Damage vision mining opportunity and embedding feature imbalance.arXiv preprint arXiv:2307.12676, 2023
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Fine-Tuning Vision-Language Models for Understanding Current Damage and Scoring Priority with Quality Guard Agent
Fine-tunes LLaVA on up to 4k bridge images for damage description and priority scoring, with a two-stage SLM quality guard showing peak semantic similarity at 3k samples.