Beyond Model Size: Probing the Gaps in Visual in-Context Learning by Training a Tiny Model

• URL: http://arxiv.org/abs/2606.10905v1
• Date: Tue, 09 Jun 2026 14:13:58 GMT
• Title: Beyond Model Size: Probing the Gaps in Visual in-Context Learning by Training a Tiny Model
• Abstract summary: We aim at making progress towards adaptive vision models, that can adapt to a new task at test-time.<n>We train a tiny visual in-context model with merely $1$ million parameters and a modest amount of $70,000$ images.<n>We compare the results of this severely capacity capped tiny model to $7,000times$ larger VICL models in different adaptive settings.
• Score: 11.54554255426278
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Visual in-Context Learning (VICL) aims at making progress towards adaptive vision models, that can -- based on a few examples -- adapt to a new task at test-time. With the history of in-context learning in natural language processing research, where large, parameter-heavy models are in use, one pathway that current VICL methods take is model- and data-scaling as key ingredients. Yet, it is not clear, whether these ingredients are the key for in-context learning to take shape in vision models. To stress-test such large models, we challenge them with an extreme counterexample: we train a tiny visual in-context model with merely $1$ million parameters and a modest amount of $70,000$ images. We compare the results of this severely capacity capped tiny model to $7,000\times$ larger VICL models in different adaptive settings, (1) on image data with small distribution shifts, (2) on unseen task encodings and (3) on a completely new task, i.e., the setting VICL envisions. With the chasm of training resources between the tiny- and large models, our experiments showcase a lack in how adaptive capabilities are measured, with respect to how tasks are encoded, which tasks were used in pre-training and the choice of metrics. These gaps in current VICL benchmarking underscore a need for innovation in evaluation of adaptive capabilities.

Related papers

• Quo Vadis, Visual In-Context Learning? A Unified Benchmark Across Domains and Tasks [47.20669655453283]
  We build a broad Visual In-Context BEnchmark (VIBE) with a focus on diverse imaging domains and a wide range of tasks.<n>We stress test six models on $14$ datasets and $12$ tasks and compare them under a unified, reproducible evaluation protocol.<n>Our evaluation uncovers key insights on the state of visual in-context learning, including limitations, systematic failure modes and promising directions.
  arXiv  Detail & Related papers  (2026-06-09T15:08:43Z)
• Virchow2: Scaling Self-Supervised Mixed Magnification Models in Pathology [2.6761982943661438]
  We introduce three new models: Virchow2, a 632 million parameter vision transformer, Virchow2G, a 1.9 billion parameter vision transformer, and Virchow2G Mini, a 22 million parameter distillation of Virchow2G. We achieve state of the art performance on 12 tile-level tasks, as compared to the top performing competing models.
  arXiv  Detail & Related papers  (2024-08-01T17:35:58Z)
• An exactly solvable model for emergence and scaling laws in the multitask sparse parity problem [2.598133279943607]
  We present a framework where each new ability (a skill) is represented as a basis function. We find analytic expressions for the emergence of new skills, as well as for scaling laws of the loss with training time, data size, model size, and optimal compute. Our simple model captures, using a single fit parameter, the sigmoidal emergence of multiple new skills as training time, data size or model size increases in the neural network.
  arXiv  Detail & Related papers  (2024-04-26T17:45:32Z)
• Data-efficient Large Vision Models through Sequential Autoregression [58.26179273091461]
  We develop an efficient, autoregression-based vision model on a limited dataset. We demonstrate how this model achieves proficiency in a spectrum of visual tasks spanning both high-level and low-level semantic understanding. Our empirical evaluations underscore the model's agility in adapting to various tasks, heralding a significant reduction in the parameter footprint.
  arXiv  Detail & Related papers  (2024-02-07T13:41:53Z)
• GistScore: Learning Better Representations for In-Context Example Selection with Gist Bottlenecks [3.9638110494107095]
  In-context Learning (ICL) is the ability of Large Language Models (LLMs) to perform new tasks when conditioned on prompts. We propose Example Gisting, a novel approach for training example encoders through supervised fine-tuning. We show that our fine-tuned models get state-of-the-art ICL performance with over 20% absolute gain over off-the-shelf retrievers.
  arXiv  Detail & Related papers  (2023-11-16T06:28:05Z)
• An Efficient General-Purpose Modular Vision Model via Multi-Task Heterogeneous Training [79.78201886156513]
  We present a model that can perform multiple vision tasks and can be adapted to other downstream tasks efficiently. Our approach achieves comparable results to single-task state-of-the-art models and demonstrates strong generalization on downstream tasks.
  arXiv  Detail & Related papers  (2023-06-29T17:59:57Z)
• Towards Foundation Models for Scientific Machine Learning: Characterizing Scaling and Transfer Behavior [32.74388989649232]
  We study how pre-training could be used for scientific machine learning (SciML) applications. We find that fine-tuning these models yields more performance gains as model size increases.
  arXiv  Detail & Related papers  (2023-06-01T00:32:59Z)
• Delving Deeper into Data Scaling in Masked Image Modeling [145.36501330782357]
  We conduct an empirical study on the scaling capability of masked image modeling (MIM) methods for visual recognition. Specifically, we utilize the web-collected Coyo-700M dataset. Our goal is to investigate how the performance changes on downstream tasks when scaling with different sizes of data and models.
  arXiv  Detail & Related papers  (2023-05-24T15:33:46Z)
• 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)
• FAME-ViL: Multi-Tasking Vision-Language Model for Heterogeneous Fashion Tasks [129.49630356651454]
  We propose a novel FAshion-focused Multi-task Efficient learning method for Vision-and-Language tasks (FAME-ViL) Our FAME-ViL can save 61.5% of parameters over alternatives, while significantly outperforming the conventional independently trained single-task models.
  arXiv  Detail & Related papers  (2023-03-04T19:07:48Z)
• SVL-Adapter: Self-Supervised Adapter for Vision-Language Pretrained Models [9.017387427570538]
  Vision-language models such as CLIP are pretrained on large volumes of internet sourced image and text pairs. Due to their size, fine-tuning these models on new datasets can be prohibitively expensive, both in terms of the supervision and compute required. We present a new approach called SVL-Adapter that combines the complementary strengths of both vision-language pretraining and self-supervised representation learning.
  arXiv  Detail & Related papers  (2022-10-07T19:35:08Z)
• Advancing Plain Vision Transformer Towards Remote Sensing Foundation Model [97.9548609175831]
  We resort to plain vision transformers with about 100 million parameters and make the first attempt to propose large vision models customized for remote sensing tasks. Specifically, to handle the large image size and objects of various orientations in RS images, we propose a new rotated varied-size window attention. Experiments on detection tasks demonstrate the superiority of our model over all state-of-the-art models, achieving 81.16% mAP on the DOTA-V1.0 dataset.
  arXiv  Detail & Related papers  (2022-08-08T09:08:40Z)

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.