3 papers
cs.LG2025
Merge to Mix: Mixing Datasets via Model Merging
Zhixu Silvia Tao, Kasper Vinken, Hao-Wei Yeh +2
Mixing datasets for fine-tuning large models (LMs) has become critical for maximizing performance on downstream tasks. However, composing effective dataset mixtures typically relie…
cs.LG2024
Task Arithmetic Through The Lens Of One-Shot Federated Learning
Zhixu Silvia Tao, Ian Mason, Sanjeev Kulkarni +1
Task Arithmetic is a model merging technique that enables the combination of multiple models' capabilities into a single model through simple arithmetic in the weight space, withou…
cs.LG2024
Rethinking VLMs and LLMs for Image Classification
Avi Cooper, Keizo Kato, Chia-Hsien Shih +8
Visual Language Models (VLMs) are now increasingly being merged with Large Language Models (LLMs) to enable new capabilities, particularly in terms of improved interactivity and op…