SAIL-Embedding Technical Report: Omni-modal Embedding Foundation Model

Multimodal embedding models aim to yield informative unified representations that empower diverse cross-modal tasks. Despite promising developments in the evolution from CLIP-based dual-tower architectures to large vision-language models, prior works still face unavoidable challenges in real-world applications and business scenarios, such as the limited modality support, unstable training mechanisms, and industrial domain gaps. In this work, we introduce SAIL-Embedding, an omni-modal embedding foundation model that addresses these issues through tailored training strategies and architectural design. In the optimization procedure, we propose a multi-stage training scheme to boost the multifaceted effectiveness of representation learning. Specifically, the content-aware progressive training aims to enhance the model's adaptability to diverse downstream tasks and master enriched cross-modal proficiency. The collaboration-aware recommendation enhancement training further adapts multimodal representations for recommendation scenarios by distilling knowledge from sequence-to-item and ID-to-item embeddings while mining user historical interests. Concurrently, we develop the stochastic specialization and dataset-driven pattern matching to strengthen model training flexibility and generalizability. Experimental results show that SAIL-Embedding achieves SOTA performance compared to other methods in different retrieval tasks. In online experiments across various real-world scenarios integrated with our model, we observe a significant increase in Lifetime (LT), which is a crucial indicator for the recommendation experience. For instance, the model delivers the 7-day LT gain of +0.158% and the 14-day LT gain of +0.144% in the Douyin-Selected scenario. For the Douyin feed rank model, the match features produced by SAIL-Embedding yield a +0.08% AUC gain.

MedQ-Bench: Evaluating and Exploring Medical Image Quality Assessment Abilities in MLLMs

Medical Image Quality Assessment (IQA) serves as the first-mile safety gate for clinical AI, yet existing approaches remain constrained by scalar, score-based metrics and fail to reflect the descriptive, human-like reasoning process central to expert evaluation. To address this gap, we introduce MedQ-Bench, a comprehensive benchmark that establishes a perception-reasoning paradigm for language-based evaluation of medical image quality with Multi-modal Large Language Models (MLLMs). MedQ-Bench defines two complementary tasks: (1) MedQ-Perception, which probes low-level perceptual capability via human-curated questions on fundamental visual attributes; and (2) MedQ-Reasoning, encompassing both no-reference and comparison reasoning tasks, aligning model evaluation with human-like reasoning on image quality. The benchmark spans five imaging modalities and over forty quality attributes, totaling 2,600 perceptual queries and 708 reasoning assessments, covering diverse image sources including authentic clinical acquisitions, images with simulated degradations via physics-based reconstructions, and AI-generated images. To evaluate reasoning ability, we propose a multi-dimensional judging protocol that assesses model outputs along four complementary axes. We further conduct rigorous human-AI alignment validation by comparing LLM-based judgement with radiologists. Our evaluation of 14 state-of-the-art MLLMs demonstrates that models exhibit preliminary but unstable perceptual and reasoning skills, with insufficient accuracy for reliable clinical use. These findings highlight the need for targeted optimization of MLLMs in medical IQA. We hope that MedQ-Bench will catalyze further exploration and unlock the untapped potential of MLLMs for medical image quality evaluation.

SAIL-VL2 Technical Report

We introduce SAIL-VL2, an open-suite vision-language foundation model (LVM) for comprehensive multimodal understanding and reasoning. As the successor to SAIL-VL, SAIL-VL2 achieves state-of-the-art performance at the 2B and 8B parameter scales across diverse image and video benchmarks, demonstrating strong capabilities from fine-grained perception to complex reasoning. Its effectiveness is driven by three core innovations. First, a large-scale data curation pipeline with scoring and filtering strategies enhances both quality and distribution across captioning, OCR, QA, and video data, improving training efficiency. Second, a progressive training framework begins with a powerful pre-trained vision encoder (SAIL-ViT), advances through multimodal pre-training, and culminates in a thinking-fusion SFT-RL hybrid paradigm that systematically strengthens model capabilities. Third, architectural advances extend beyond dense LLMs to efficient sparse Mixture-of-Experts (MoE) designs. With these contributions, SAIL-VL2 demonstrates competitive performance across 106 datasets and achieves state-of-the-art results on challenging reasoning benchmarks such as MMMU and MathVista. Furthermore, on the OpenCompass leaderboard, SAIL-VL2-2B ranks first among officially released open-source models under the 4B parameter scale, while serving as an efficient and extensible foundation for the open-source multimodal community.

PersonaAnimator: Personalized Motion Transfer from Unconstrained Videos

Recent advances in motion generation show remarkable progress. However, several limitations remain: (1) Existing pose-guided character motion transfer methods merely replicate motion without learning its style characteristics, resulting in inexpressive characters. (2) Motion style transfer methods rely heavily on motion capture data, which is difficult to obtain. (3) Generated motions sometimes violate physical laws. To address these challenges, this paper pioneers a new task: Video-to-Video Motion Personalization. We propose a novel framework, PersonaAnimator, which learns personalized motion patterns directly from unconstrained videos. This enables personalized motion transfer. To support this task, we introduce PersonaVid, the first video-based personalized motion dataset. It contains 20 motion content categories and 120 motion style categories. We further propose a Physics-aware Motion Style Regularization mechanism to enforce physical plausibility in the generated motions. Extensive experiments show that PersonaAnimator outperforms state-of-the-art motion transfer methods and sets a new benchmark for the Video-to-Video Motion Personalization task.

eMotions: A Large-Scale Dataset and Audio-Visual Fusion Network for Emotion Analysis in Short-form Videos

Short-form videos (SVs) have become a vital part of our online routine for acquiring and sharing information. Their multimodal complexity poses new challenges for video analysis, highlighting the need for video emotion analysis (VEA) within the community. Given the limited availability of SVs emotion data, we introduce eMotions, a large-scale dataset consisting of 27,996 videos with full-scale annotations. To ensure quality and reduce subjective bias, we emphasize better personnel allocation and propose a multi-stage annotation procedure. Additionally, we provide the category-balanced and test-oriented variants through targeted sampling to meet diverse needs. While there have been significant studies on videos with clear emotional cues (e.g., facial expressions), analyzing emotions in SVs remains a challenging task. The challenge arises from the broader content diversity, which introduces more distinct semantic gaps and complicates the representations learning of emotion-related features. Furthermore, the prevalence of audio-visual co-expressions in SVs leads to the local biases and collective information gaps caused by the inconsistencies in emotional expressions. To tackle this, we propose AV-CANet, an end-to-end audio-visual fusion network that leverages video transformer to capture semantically relevant representations. We further introduce the Local-Global Fusion Module designed to progressively capture the correlations of audio-visual features. Besides, EP-CE Loss is constructed to globally steer optimizations with tripolar penalties. Extensive experiments across three eMotions-related datasets and four public VEA datasets demonstrate the effectiveness of our proposed AV-CANet, while providing broad insights for future research. Moreover, we conduct ablation studies to examine the critical components of our method. Dataset and code will be made available at Github.

SAILViT: Towards Robust and Generalizable Visual Backbones for MLLMs via Gradual Feature Refinement

Vision Transformers (ViTs) are essential as foundation backbones in establishing the visual comprehension capabilities of Multimodal Large Language Models (MLLMs). Although most ViTs achieve impressive performance through image-text pair-based contrastive learning or self-supervised mechanisms, they struggle to engage in connector-based co-training directly with LLMs due to potential parameter initialization conflicts and modality semantic gaps. To address the above challenges, this paper proposes SAILViT, a gradual feature learning-enhanced ViT for facilitating MLLMs to break through performance bottlenecks in complex multimodal interactions. SAILViT achieves coarse-to-fine-grained feature alignment and world knowledge infusion with gradual feature refinement, which better serves target training demands. We perform thorough empirical analyses to confirm the powerful robustness and generalizability of SAILViT across different dimensions, including parameter sizes, model architectures, training strategies, and data scales. Equipped with SAILViT, existing MLLMs show significant and consistent performance improvements on the OpenCompass benchmark across extensive downstream tasks. SAILViT series models are released at https://huggingface.co/BytedanceDouyinContent.

LingoLoop Attack: Trapping MLLMs via Linguistic Context and State Entrapment into Endless Loops

Multimodal Large Language Models (MLLMs) have shown great promise but require substantial computational resources during inference. Attackers can exploit this by inducing excessive output, leading to resource exhaustion and service degradation. Prior energy-latency attacks aim to increase generation time by broadly shifting the output token distribution away from the EOS token, but they neglect the influence of token-level Part-of-Speech (POS) characteristics on EOS and sentence-level structural patterns on output counts, limiting their efficacy. To address this, we propose LingoLoop, an attack designed to induce MLLMs to generate excessively verbose and repetitive sequences. First, we find that the POS tag of a token strongly affects the likelihood of generating an EOS token. Based on this insight, we propose a POS-Aware Delay Mechanism to postpone EOS token generation by adjusting attention weights guided by POS information. Second, we identify that constraining output diversity to induce repetitive loops is effective for sustained generation. We introduce a Generative Path Pruning Mechanism that limits the magnitude of hidden states, encouraging the model to produce persistent loops. Extensive experiments demonstrate LingoLoop can increase generated tokens by up to 30 times and energy consumption by a comparable factor on models like Qwen2.5-VL-3B, consistently driving MLLMs towards their maximum generation limits. These findings expose significant MLLMs' vulnerabilities, posing challenges for their reliable deployment. The code will be released publicly following the paper's acceptance.

AdaLRS: Loss-Guided Adaptive Learning Rate Search for Efficient Foundation Model Pretraining

Learning rate is widely regarded as crucial for effective foundation model pretraining. Recent research explores and demonstrates the transferability of learning rate configurations across varying model and dataset sizes, etc. Nevertheless, these approaches are constrained to specific training scenarios and typically necessitate extensive hyperparameter tuning on proxy models. In this work, we propose \textbf{AdaLRS}, a plug-in-and-play adaptive learning rate search algorithm that conducts online optimal learning rate search via optimizing loss descent velocities. We provide experiment results to show that the optimization of training loss and loss descent velocity in foundation model pretraining are both convex and share the same optimal learning rate. Relying solely on training loss dynamics, AdaLRS involves few extra computations to guide the search process, and its convergence is guaranteed via theoretical analysis. Experiments on both LLM and VLM pretraining show that AdaLRS adjusts suboptimal learning rates to the neighborhood of optimum with marked efficiency and effectiveness, with model performance improved accordingly. We also show the robust generalizability of AdaLRS across varying training scenarios, such as different model sizes, training paradigms, and base learning rate scheduler choices.

DanmakuTPPBench: A Multi-modal Benchmark for Temporal Point Process Modeling and Understanding

We introduce DanmakuTPPBench, a comprehensive benchmark designed to advance multi-modal Temporal Point Process (TPP) modeling in the era of Large Language Models (LLMs). While TPPs have been widely studied for modeling temporal event sequences, existing datasets are predominantly unimodal, hindering progress in models that require joint reasoning over temporal, textual, and visual information. To address this gap, DanmakuTPPBench comprises two complementary components: (1) DanmakuTPP-Events, a novel dataset derived from the Bilibili video platform, where user-generated bullet comments (Danmaku) naturally form multi-modal events annotated with precise timestamps, rich textual content, and corresponding video frames; (2) DanmakuTPP-QA, a challenging question-answering dataset constructed via a novel multi-agent pipeline powered by state-of-the-art LLMs and multi-modal LLMs (MLLMs), targeting complex temporal-textual-visual reasoning. We conduct extensive evaluations using both classical TPP models and recent MLLMs, revealing significant performance gaps and limitations in current methods' ability to model multi-modal event dynamics. Our benchmark establishes strong baselines and calls for further integration of TPP modeling into the multi-modal language modeling landscape. The code and dataset have been released at https://github.com/FRENKIE-CHIANG/DanmakuTPPBench

MCCD: Multi-Agent Collaboration-based Compositional Diffusion for Complex Text-to-Image Generation

Diffusion models have shown excellent performance in text-to-image generation. Nevertheless, existing methods often suffer from performance bottlenecks when handling complex prompts that involve multiple objects, characteristics, and relations. Therefore, we propose a Multi-agent Collaboration-based Compositional Diffusion (MCCD) for text-to-image generation for complex scenes. Specifically, we design a multi-agent collaboration-based scene parsing module that generates an agent system comprising multiple agents with distinct tasks, utilizing MLLMs to extract various scene elements effectively. In addition, Hierarchical Compositional diffusion utilizes a Gaussian mask and filtering to refine bounding box regions and enhance objects through region enhancement, resulting in the accurate and high-fidelity generation of complex scenes. Comprehensive experiments demonstrate that our MCCD significantly improves the performance of the baseline models in a training-free manner, providing a substantial advantage in complex scene generation.