MoE-LLaVA: Mixture of Experts for Large Vision-Language Models
- URL: http://arxiv.org/abs/2401.15947v4
- Date: Sat, 6 Jul 2024 13:46:02 GMT
- Title: MoE-LLaVA: Mixture of Experts for Large Vision-Language Models
- Authors: Bin Lin, Zhenyu Tang, Yang Ye, Jiaxi Cui, Bin Zhu, Peng Jin, Jinfa Huang, Junwu Zhang, Yatian Pang, Munan Ning, Li Yuan,
- Abstract summary: We propose a simple yet effective training strategy MoE-Tuning for LVLMs.
MoE-LLaVA, a MoE-based sparse LVLM architecture, uniquely activates only the top-k experts through routers.
Experiments show the significant performance of MoE-LLaVA in a variety of visual understanding and object hallucination benchmarks.
- Score: 27.930351465266515
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Recent advances demonstrate that scaling Large Vision-Language Models (LVLMs) effectively improves downstream task performances. However, existing scaling methods enable all model parameters to be active for each token in the calculation, which brings massive training and inferring costs. In this work, we propose a simple yet effective training strategy MoE-Tuning for LVLMs. This strategy innovatively addresses the common issue of performance degradation in multi-modal sparsity learning, consequently constructing a sparse model with an outrageous number of parameters but a constant computational cost. Furthermore, we present the MoE-LLaVA, a MoE-based sparse LVLM architecture, which uniquely activates only the top-k experts through routers during deployment, keeping the remaining experts inactive. Extensive experiments show the significant performance of MoE-LLaVA in a variety of visual understanding and object hallucination benchmarks. Remarkably, with only approximately 3B sparsely activated parameters, MoE-LLaVA demonstrates performance comparable to the LLaVA-1.5-7B on various visual understanding datasets and even surpasses the LLaVA-1.5-13B in object hallucination benchmark. Through MoE-LLaVA, we aim to establish a baseline for sparse LVLMs and provide valuable insights for future research in developing more efficient and effective multi-modal learning systems. Code is released at https://github.com/PKU-YuanGroup/MoE-LLaVA.
Related papers
- Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs [56.391404083287235]
We introduce Cambrian-1, a family of multimodal LLMs (MLLMs) designed with a vision-centric approach.
Our study uses LLMs and visual instruction tuning as an interface to evaluate various visual representations.
We provide model weights, code, supporting tools, datasets, and detailed instruction-tuning and evaluation recipes.
arXiv Detail & Related papers (2024-06-24T17:59:42Z) - Boosting Continual Learning of Vision-Language Models via Mixture-of-Experts Adapters [65.15700861265432]
We present a parameter-efficient continual learning framework to alleviate long-term forgetting in incremental learning with vision-language models.
Our approach involves the dynamic expansion of a pre-trained CLIP model, through the integration of Mixture-of-Experts (MoE) adapters.
To preserve the zero-shot recognition capability of vision-language models, we introduce a Distribution Discriminative Auto-Selector.
arXiv Detail & Related papers (2024-03-18T08:00:23Z) - ALLaVA: Harnessing GPT4V-Synthesized Data for Lite Vision-Language Models [45.040292339670096]
Large vision-language models (LVLMs) have shown premise in a broad range of vision-language tasks with their strong reasoning and generalization capabilities.
This study aims to bridge the performance gap between traditional-scale LVLMs and resource-friendly lite versions by adopting high-quality training data.
arXiv Detail & Related papers (2024-02-18T19:26:49Z) - Mitigating Object Hallucination in Large Vision-Language Models via
Classifier-Free Guidance [56.04768229686853]
Large Vision-Language Models (LVLMs) tend to hallucinate non-existing objects in the images.
We introduce a framework called Mitigating hallucinAtion via classifieR-Free guIdaNcE (MARINE)
MARINE is both training-free and API-free, and can effectively and efficiently reduce object hallucinations during the generation process.
arXiv Detail & Related papers (2024-02-13T18:59:05Z) - Learning Semantic Proxies from Visual Prompts for Parameter-Efficient Fine-Tuning in Deep Metric Learning [13.964106147449051]
Existing solutions concentrate on fine-tuning the pre-trained models on conventional image datasets.
We propose a novel and effective framework based on learning Visual Prompts (VPT) in the pre-trained Vision Transformers (ViT)
We demonstrate that our new approximations with semantic information are superior to representative capabilities.
arXiv Detail & Related papers (2024-02-04T04:42:05Z) - Enhancing Multimodal Large Language Models with Vision Detection Models: An Empirical Study [32.57246173437492]
This paper presents an empirical study on enhancing MLLMs with state-of-the-art (SOTA) object detection and Optical Character Recognition (OCR) models to improve fine-grained understanding and reduce hallucination in responses.
We conduct systematic and extensive experiments with representative models such as LLaVA-1.5, DINO, PaddleOCRv2, and Grounding DINO.
Notably, the enhanced LLaVA-1.5 outperforms its original 7B/13B models on all 10 benchmarks, achieving an improvement of up to 12.5% on the normalized average score.
arXiv Detail & Related papers (2024-01-31T16:38:32Z) - When Parameter-efficient Tuning Meets General-purpose Vision-language
Models [65.19127815275307]
PETAL revolutionizes the training process by requiring only 0.5% of the total parameters, achieved through a unique mode approximation technique.
Our experiments reveal that PETAL not only outperforms current state-of-the-art methods in most scenarios but also surpasses full fine-tuning models in effectiveness.
arXiv Detail & Related papers (2023-12-16T17:13:08Z) - eP-ALM: Efficient Perceptual Augmentation of Language Models [70.47962271121389]
We propose to direct effort to efficient adaptations of existing models, and propose to augment Language Models with perception.
Existing approaches for adapting pretrained models for vision-language tasks still rely on several key components that hinder their efficiency.
We show that by freezing more than 99% of total parameters, training only one linear projection layer, and prepending only one trainable token, our approach (dubbed eP-ALM) significantly outperforms other baselines on VQA and Captioning.
arXiv Detail & Related papers (2023-03-20T19:20:34Z) - Scaling Vision-Language Models with Sparse Mixture of Experts [128.0882767889029]
We show that mixture-of-experts (MoE) techniques can achieve state-of-the-art performance on a range of benchmarks over dense models of equivalent computational cost.
Our research offers valuable insights into stabilizing the training of MoE models, understanding the impact of MoE on model interpretability, and balancing the trade-offs between compute performance when scaling vision-language models.
arXiv Detail & Related papers (2023-03-13T16:00:31Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.