Using more flexible Bayesian posterior approximations in model merging gives better previews of multitask finetuning weightings on vision and language transformers.
Neural networks for pattern recognition
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
-
Variational Model Merging for Pareto Front Estimation in Multitask Finetuning
Using more flexible Bayesian posterior approximations in model merging gives better previews of multitask finetuning weightings on vision and language transformers.