Concatenating visual tokens from multiple MLLMs in the same family and linearly merging their language model deltas yields a training-free fusion that improves multimodal benchmark scores over each source model.
An image is worth 1/2 tokens after layer 2: Plug-and-play inference acceleration for large vision-language models, 2024
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Enhancing Perception Capabilities of Multimodal LLMs with Training-Free Fusion
Concatenating visual tokens from multiple MLLMs in the same family and linearly merging their language model deltas yields a training-free fusion that improves multimodal benchmark scores over each source model.