A memory-based editor with a multimodal scope classifier improves focused editing of one object in multi-entity images, supported by a new benchmark.
Lan- guage models are few-shot learners
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Visual-Oriented Fine-Grained Knowledge Editing for MultiModal Large Language Models
A memory-based editor with a multimodal scope classifier improves focused editing of one object in multi-entity images, supported by a new benchmark.