Traditional approaches to adapting multi-modal large language models (MLLMs) to new tasks have relied heavily on fine-tuning. This paper introduces Efficient Multi-Modal Long Context Learning (EMLoC), a novel training-free alternative that embeds demonstration examples directly into the model input. EMLoC offers a more efficient, flexible, and scalable solution for task adaptation. Because extremely lengthy inputs introduce prohibitive computational and memory overhead, EMLoC contributes a chunk-wise compression mechanism combined with layer-wise adaptive pruning. It condenses long-context multimodal inputs into compact, task-specific memory representations. By adaptively pruning tokens at each layer under a Jensen-Shannon divergence constraint, our method achieves a dramatic reduction in inference complexity without sacrificing performance. This approach is the first to seamlessly integrate compression and pruning techniques for multi-modal long-context learning, offering a scalable and efficient solution for real-world applications. Extensive experiments on diverse vision-language benchmarks demonstrate that EMLoC achieves performance on par with or superior to naive long-context approaches. Our results highlight the potential of EMLoC as a groundbreaking framework for efficient and flexible adaptation of multi-modal models in resource-constrained environments. Codes are publicly available at https://github.com/Zehong-Ma/EMLoC.

Existing acceleration techniques for video diffusion models often rely on uniform heuristics or time-embedding variants to skip timesteps and reuse cached features. These approaches typically require extensive calibration with curated prompts and risk inconsistent outputs due to prompt-specific overfitting. In this paper, we introduce a novel and robust discovery: a unified magnitude law observed across different models and prompts. Specifically, the magnitude ratio of successive residual outputs decreases monotonically, steadily in most timesteps while rapidly in the last several steps. Leveraging this insight, we introduce a Magnitude-aware Cache (MagCache) that adaptively skips unimportant timesteps using an error modeling mechanism and adaptive caching strategy. Unlike existing methods requiring dozens of curated samples for calibration, MagCache only requires a single sample for calibration. Experimental results show that MagCache achieves 2.10×-2.68× speedups on Open-Sora, CogVideoX, Wan 2.1, and HunyuanVideo, while preserving superior visual fidelity. It significantly outperforms existing methods in LPIPS, SSIM, and PSNR, under similar computational budgets.

Fine-grained text-to-image retrieval aims to retrieve a fine-grained target image with a given text query. Existing methods typically assume that each training image is accurately depicted by its textual descriptions. However, textual descriptions can be ambiguous and fail to depict discriminative visual details in images, leading to inaccurate representation learning. To alleviate the effects of text ambiguity, we propose a Multi-Modal Reference learning framework to learn robust representations. We first propose a multi-modal reference construction module to aggregate all visual and textual details of the same object into a comprehensive multi-modal reference. The multi-modal reference hence facilitates the subsequent representation learning and retrieval similarity computation. Specifically, a reference-guided representation learning module is proposed to use multi-modal references to learn more accurate visual and textual representations. Additionally, we introduce a reference-based refinement method that employs the object references to compute a reference-based similarity that refines the initial retrieval results. Extensive experiments are conducted on five fine-grained text-to-image retrieval datasets for different text-to-image retrieval tasks. The proposed method has achieved superior performance over state-of-the-art methods. For instance, on the text-to-person image retrieval dataset RSTPReid, our method achieves the Rank1 accuracy of 56.2%, surpassing the recent CFine by 5.6%.

The challenge of open-vocabulary recognition lies in the model has no clue of new categories it is applied to. Existing works have proposed different methods to embed category cues into the model, e.g., through few-shot fine-tuning, providing category names or textual descriptions to Vision-Language Models. Fine-tuning is time-consuming and degrades the generalization capability. Textual descriptions could be ambiguous and fail to depict visual details. This paper tackles open-vocabulary recognition from a different perspective by referring to multi-modal clues composed of textual descriptions and exemplar images. Our method, named OVMR, adopts two innovative components to pursue a more robust category cues embedding. A multi-modal classifier is first generated by dynamically complementing textual descriptions with image exemplars. A preference-based refinement module is hence applied to fuse uni-modal and multi-modal classifiers, with the aim to alleviate issues of low-quality exemplar images or textual descriptions. The proposed OVMR is a plug-and-play module, and works well with exemplar images randomly crawled from the Internet. Extensive experiments have demonstrated the promising performance of OVMR, e.g., it outperforms existing methods across various scenarios and setups. Codes are publicly available at https://github.com/Zehong-Ma/OVMR.

Text-based person search is a task to retrieve the corresponding person in a large-scale image database given a textual description, which has important value in various fields like video surveillance. In the inferring phase, language descriptions, serving as queries, guide to search the corresponding person images. Most existing methods apply cross-modal signals to guide feature refinement. However, they employ visual features from the gallery to refine textual features, which may cause high similarity between unmatched pairs. Besides, the similarity-based cross-modal attention could disturb the choice of interested areas for descriptions. In this paper, we analyze the deficiency of previous methods and carefully design a Text-guided Visual Feature Refinement network (TVFR), which utilizes text as reference to refine visual representations. Firstly, we divide each visual feature into several horizontal stripes for fine-grained refinement. After that, we employ a text-based filter generation module to generate description-customized filters, which are used to indicate the corresponding stripes mentioned in the textual input. Thereafter, we employ a text-guided visual feature refinement module to fuse part-level visual features adaptively for each description. In experiments, we validate our TVFR through extensive experiments on CUHK-PEDES, which is the only available dataset for text-based person search. To the best of our knowledge, the TVFR outperforms other state-of-the-art methods.