# When Do We Not Need Larger Vision Models?

arXiv: https://arxiv.org/abs/2403.13043 · HTML: https://mlsanity.com/paper/2403.13043/

## Key facts

- Cited by 67 papers in the map (data through 14 Oct 2026).
- Builds on 67 of its 87 references that are in the map.
- First posted on arXiv 19 Mar 2024.

## Builds on (most cited first, 30 of 67)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Hierarchical Text-Conditional Image Generation with CLIP Latents](https://mlsanity.com/paper/2204.06125/) (2022)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [Feature Pyramid Networks for Object Detection](https://mlsanity.com/paper/1612.03144/) (2016)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering](https://mlsanity.com/paper/1612.00837/) (2016)
- [Masked-attention Mask Transformer for Universal Image Segmentation](https://mlsanity.com/paper/2112.01527/) (2021)
- [MMBench: Is Your Multi-modal Model an All-around Player?](https://mlsanity.com/paper/2307.06281/) (2023)
- [MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI](https://mlsanity.com/paper/2311.16502/) (2023)
- [Unsupervised Feature Learning via Non-Parametric Instance-level Discrimination](https://mlsanity.com/paper/1805.01978/) (2018)
- [Towards VQA Models That Can Read](https://mlsanity.com/paper/1904.08920/) (2019)
- [MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts](https://mlsanity.com/paper/2310.02255/) (2023)
- [Deep High-Resolution Representation Learning for Visual Recognition](https://mlsanity.com/paper/1908.07919/) (2019)
- [Scaling Autoregressive Models for Content-Rich Text-to-Image Generation](https://mlsanity.com/paper/2206.10789/) (2022)
- [Unified Perceptual Parsing for Scene Understanding](https://mlsanity.com/paper/1807.10221/) (2018)
- [MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilities](https://mlsanity.com/paper/2308.02490/) (2023)
- [Reproducible scaling laws for contrastive language-image learning](https://mlsanity.com/paper/2212.07143/) (2022)

## Built on by (most cited first, 30 of 67)

- [NVILA: Efficient Frontier Visual Language Models](https://mlsanity.com/paper/2412.04468/) (2024)
- [Eagle: Exploring The Design Space for Multimodal LLMs with Mixture of Encoders](https://mlsanity.com/paper/2408.15998/) (2024)
- [GenAI-Bench: Evaluating and Improving Compositional Text-to-Visual Generation](https://mlsanity.com/paper/2406.13743/) (2024)
- [Multimodal Autoregressive Pre-training of Large Vision Encoders](https://mlsanity.com/paper/2411.14402/) (2024)
- [Why are Visually-Grounded Language Models Bad at Image Classification?](https://mlsanity.com/paper/2405.18415/) (2024)
- [Efficient Multimodal Large Language Models: A Survey](https://mlsanity.com/paper/2405.10739/) (2024)
- [Unveiling Encoder-Free Vision-Language Models](https://mlsanity.com/paper/2406.11832/) (2024)
- [MedTrinity-25M: A Large-scale Multimodal Dataset with Multigranular Annotations for Medicine](https://mlsanity.com/paper/2408.02900/) (2024)
- [FastVLM: Efficient Vision Encoding for Vision Language Models](https://mlsanity.com/paper/2412.13303/) (2024)
- [DenseFusion-1M: Merging Vision Experts for Comprehensive Multimodal Perception](https://mlsanity.com/paper/2407.08303/) (2024)
- [OmniVinci: Enhancing Architecture and Data for Omni-Modal Understanding LLM](https://mlsanity.com/paper/2510.15870/) (2025)
- [CuMo: Scaling Multimodal LLM with Co-Upcycled Mixture-of-Experts](https://mlsanity.com/paper/2405.05949/) (2024)
- [Visual Agents as Fast and Slow Thinkers](https://mlsanity.com/paper/2408.08862/) (2024)
- [TopViewRS: Vision-Language Models as Top-View Spatial Reasoners](https://mlsanity.com/paper/2406.02537/) (2024)
- [Med-MoE: Mixture of Domain-Specific Experts for Lightweight Medical Vision-Language Models](https://mlsanity.com/paper/2404.10237/) (2024)
- [The Larger the Better? Improved LLM Code-Generation via Budget Reallocation](https://mlsanity.com/paper/2404.00725/) (2024)
- [LLaVA-MORE: A Comparative Study of LLMs and Visual Backbones for Enhanced Visual Instruction Tuning](https://mlsanity.com/paper/2503.15621/) (2025)
- [Scaling Vision Pre-Training to 4K Resolution](https://mlsanity.com/paper/2503.19903/) (2025)
- [Mini-Monkey: Alleviating the Semantic Sawtooth Effect for Lightweight MLLMs via Complementary Image Pyramid](https://mlsanity.com/paper/2408.02034/) (2024)
- [MoVE-KD: Knowledge Distillation for VLMs with Mixture of Visual Encoders](https://mlsanity.com/paper/2501.01709/) (2025)
- [VILA2: VILA Augmented VILA](https://mlsanity.com/paper/2407.17453/) (2024)
- [Why Vision Language Models Struggle with Visual Arithmetic? Towards Enhanced Chart and Geometry Understanding](https://mlsanity.com/paper/2502.11492/) (2025)
- [Triad: Vision Foundation Model for 3D Magnetic Resonance Imaging](https://mlsanity.com/paper/2502.14064/) (2025)
- [FALCON: Resolving Visual Redundancy and Fragmentation in High-resolution Multimodal Large Language Models via Visual Registers](https://mlsanity.com/paper/2501.16297/) (2025)
- [SEA: Supervised Embedding Alignment for Token-Level Visual-Textual Integration in MLLMs](https://mlsanity.com/paper/2408.11813/) (2024)
- [Fewer Tokens, Greater Scaling: Self-Adaptive Visual Bases for Efficient and Expansive Representation Learning](https://mlsanity.com/paper/2511.19515/) (2025)
- [What Makes for Good Image Captions?](https://mlsanity.com/paper/2405.00485/) (2024)
- [VideoGameBunny: Towards vision assistants for video games](https://mlsanity.com/paper/2407.15295/) (2024)
- [FlashSloth: Lightning Multimodal Large Language Models via Embedded Visual Compression](https://mlsanity.com/paper/2412.04317/) (2024)
- [Native Visual Understanding: Resolving Resolution Dilemmas in Vision-Language Models](https://mlsanity.com/paper/2506.12776/) (2025)

## Abstract (arXiv)

Scaling up the size of vision models has been the de facto standard to obtain more powerful visual representations. In this work, we discuss the point beyond which larger vision models are not necessary. First, we demonstrate the power of Scaling on Scales (S2), whereby a pre-trained and frozen smaller vision model (e.g., ViT-B or ViT-L), run over multiple image scales, can outperform larger models (e.g., ViT-H or ViT-G) on classification, segmentation, depth estimation, Multimodal LLM (MLLM) benchmarks, and robotic manipulation. Notably, S2 achieves state-of-the-art performance in detailed understanding of MLLM on the V* benchmark, surpassing models such as GPT-4V. We examine the conditions under which S2 is a preferred scaling approach compared to scaling on model size. While larger models have the advantage of better generalization on hard examples, we show that features of larger vision models can be well approximated by those of multi-scale smaller models. This suggests most, if not all, of the representations learned by current large pre-trained models can also be obtained from multi-scale smaller models. Our results show that a multi-scale smaller model has comparable learning capacity to a larger model, and pre-training smaller models with S2 can match or even exceed the advantage of larger models. We release a Python package that can apply S2 on any vision model with one line of code: https://github.com/bfshi/scalingonscales.