InfoSAM improves SAM fine-tuning by compressing and distilling encoder-decoder relations using a Rényi entropy based mutual information loss.
Pkd: General distillation framework for object detectors via pearson correlation coefficient
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
-
InfoSAM: Fine-Tuning the Segment Anything Model from An Information-Theoretic Perspective
InfoSAM improves SAM fine-tuning by compressing and distilling encoder-decoder relations using a Rényi entropy based mutual information loss.