登入
洞見 - Dynamic Tuning Strategy for MLLMs
HyperLLaVA: Dynamic Visual and Language Expert Tuning for Multimodal Large Language Models
HyperLLaVA introduces adaptive tuning of the projector and LLM parameters, along with dynamic visual and language experts, to enhance performance on multimodal tasks.
HyperLLaVA: Dynamic Visual and Language Expert Tuning for Multimodal Large Language Models
HyperLLaVA introduces adaptive tuning of the projector and LLM parameters, along with dynamic visual and language experts, to enhance multimodal task performance.
