VGP uses low-rank virtual node, edge, and node prompts to adapt frozen Vision GNNs, matching or exceeding full fine-tuning on ten vision and nine graph classification benchmarks.
Our approach does not require the re-training of the entire model, which helps to mitigate the computational cost typically associated with full fine-tuning
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
-
Vision Graph Prompting via Semantic Low-Rank Decomposition
VGP uses low-rank virtual node, edge, and node prompts to adapt frozen Vision GNNs, matching or exceeding full fine-tuning on ten vision and nine graph classification benchmarks.