LLaVA-MORE: A Comparative Study of LLMs and Visual Backbones for Enhanced Visual Instruction Tuning

• URL: http://arxiv.org/abs/2503.15621v1
• Date: Wed, 19 Mar 2025 18:10:12 GMT
• Title: LLaVA-MORE: A Comparative Study of LLMs and Visual Backbones for Enhanced Visual Instruction Tuning
• Authors: Federico Cocchi, Nicholas Moratelli, Davide Caffagni, Sara Sarto, Lorenzo Baraldi, Marcella Cornia, Rita Cucchiara,
• Abstract summary: Trade-offs between model size, architecture, and performance remain underexplored.<n>In this paper, we introduce LLaVA-MORE, a new family of MLLMs that integrates recent language models with diverse visual backbones.<n>To ensure fair comparisons, we employ a unified training protocol applied consistently across all architectures.
• Score: 39.54891426369773
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent progress in Multimodal Large Language Models (MLLMs) has highlighted the critical roles of both the visual backbone and the underlying language model. While prior work has primarily focused on scaling these components to billions of parameters, the trade-offs between model size, architecture, and performance remain underexplored. Additionally, inconsistencies in training data and evaluation protocols have hindered direct comparisons, making it difficult to derive optimal design choices. In this paper, we introduce LLaVA-MORE, a new family of MLLMs that integrates recent language models with diverse visual backbones. To ensure fair comparisons, we employ a unified training protocol applied consistently across all architectures. Our analysis systematically explores both small- and medium-scale LLMs -- including Phi-4, LLaMA-3.1, and Gemma-2 -- to evaluate multimodal reasoning, generation, and instruction following, while examining the relationship between model size and performance. Beyond evaluating the LLM impact on final results, we conduct a comprehensive study of various visual encoders, ranging from CLIP-based architectures to alternatives such as DINOv2, SigLIP, and SigLIP2. Additional experiments investigate the effects of increased image resolution and variations in pre-training datasets. Overall, our results provide insights into the design of more effective MLLMs, offering a reproducible evaluation framework that facilitates direct comparisons and can guide future model development. Our source code and trained models are publicly available at: https://github.com/aimagelab/LLaVA-MORE.

Related papers

• CoMMIT: Coordinated Instruction Tuning for Multimodal Large Language Models [68.64605538559312]
  In this paper, we analyze the MLLM instruction tuning from both theoretical and empirical perspectives. Inspired by our findings, we propose a measurement to quantitatively evaluate the learning balance. In addition, we introduce an auxiliary loss regularization method to promote updating of the generation distribution of MLLMs.
  arXiv  Detail & Related papers  (2024-07-29T23:18:55Z)
• LLAVADI: What Matters For Multimodal Large Language Models Distillation [77.73964744238519]
  In this work, we do not propose a new efficient model structure or train small-scale MLLMs from scratch. Our studies involve training strategies, model choices, and distillation algorithms in the knowledge distillation process. By evaluating different benchmarks and proper strategy, even a 2.7B small-scale model can perform on par with larger models with 7B or 13B parameters.
  arXiv  Detail & Related papers  (2024-07-28T06:10:47Z)
• Evaluating Linguistic Capabilities of Multimodal LLMs in the Lens of Few-Shot Learning [15.919493497867567]
  This study aims to evaluate the performance of Multimodal Large Language Models (MLLMs) on the VALSE benchmark. We conducted a comprehensive assessment of state-of-the-art MLLMs, varying in model size and pretraining datasets.
  arXiv  Detail & Related papers  (2024-07-17T11:26:47Z)
• Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs [61.143381152739046]
  We introduce Cambrian-1, a family of multimodal LLMs (MLLMs) designed with a vision-centric approach.<n>Our study uses LLMs and visual instruction tuning as an interface to evaluate various visual representations.<n>We provide model weights, code, supporting tools, datasets, and detailed instruction-tuning and evaluation recipes.
  arXiv  Detail & Related papers  (2024-06-24T17:59:42Z)
• Explore In-Context Segmentation via Latent Diffusion Models [132.26274147026854]
  In-context segmentation aims to segment objects using given reference images.<n>Most existing approaches adopt metric learning or masked image modeling to build the correlation between visual prompts and input image queries.<n>This work approaches the problem from a fresh perspective - unlocking the capability of the latent diffusion model for in-context segmentation.
  arXiv  Detail & Related papers  (2024-03-14T17:52:31Z)
• The Revolution of Multimodal Large Language Models: A Survey [46.84953515670248]
  Multimodal Large Language Models (MLLMs) can seamlessly integrate visual and textual modalities. This paper provides a review of recent visual-based MLLMs, analyzing their architectural choices, multimodal alignment strategies, and training techniques.
  arXiv  Detail & Related papers  (2024-02-19T19:01:01Z)
• From Training-Free to Adaptive: Empirical Insights into MLLMs' Understanding of Detection Information [32.57246173437492]
  Vision detection models excel at recognizing fine-grained image details.<n>One effective strategy is to infuse detection information in text format, which has proven simple and effective.<n>This paper addresses the question: How does training impact MLLMs' understanding of infused textual detection information?
  arXiv  Detail & Related papers  (2024-01-31T16:38:32Z)
• MoE-LLaVA: Mixture of Experts for Large Vision-Language Models [49.32669226551026]
  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.
  arXiv  Detail & Related papers  (2024-01-29T08:13:40Z)
• An Empirical Study of Scaling Instruct-Tuned Large Multimodal Models [116.50367506746713]
  We present an empirical study of scaling LLaVA up to 33B and 65B/70B. We find that scaling LMM consistently enhances model performance and improves language capabilities. We hope that this study makes state-of-the-art LMM research at a larger scale more accessible.
  arXiv  Detail & Related papers  (2023-09-18T17:30:46Z)

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.