NP Merge learns a separate interpolation weight for each parameter when merging two aligned networks, improving on several baselines but not on ensembles.
On power laws in deep ensembles,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Non-Uniform Parameter-Wise Model Merging
NP Merge learns a separate interpolation weight for each parameter when merging two aligned networks, improving on several baselines but not on ensembles.