Omni-SMoLA: Boosting Generalist Multimodal Models with Soft Mixture of Low-rank Experts

arXiv:2312.00968v2 Announce Type: replace
Abstract: Large multi-modal models (LMMs) exhibit remarkable performance across numerous tasks. However, generalist LMMs often suffer from performance degradation when tuned over a large collection of tasks. Recent research suggests that Mixture of Experts (MoE) architectures are useful for instruction tuning, but for LMMs of parameter size around O(50-100B), the prohibitive cost of replicating and storing the expert models severely limits the number of experts we can use. We propose Omni-SMoLA, an architecture that uses the …

abstract architectures arxiv boosting collection cs.cl cs.cv experts however lmms low mixture of experts modal moe multi-modal multimodal multimodal models performance research tasks type