ReAlign corrects the modality gap in unpaired data to let MLLMs learn visual distributions from text alone before instruction tuning, reducing dependence on expensive paired corpora.
Decap: Decoding clip latents for zero-shot captioning via text-only training
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Modality Gap-Driven Subspace Alignment Training Paradigm For Multimodal Large Language Models
ReAlign corrects the modality gap in unpaired data to let MLLMs learn visual distributions from text alone before instruction tuning, reducing dependence on expensive paired corpora.