Trending Papers

Z-Image, a 6B-parameter Scalable Single-Stream Diffusion Transformer (S3-DiT) model, achieves high-performance image generation with reduced computational cost, offering sub-second inference and compatibility with consumer hardware.

The study reveals that in text-to-image generation, CFG Augmentation is the primary driver of few-step distillation in Distribution Matching Distillation (DMD), while the distribution matching term acts as a regularizer.

WizardCoder, a Code LLM fine-tuned with complex instructions using Evol-Instruct, outperforms other open-source and closed LLMs on several code generation benchmarks.

LightRAG improves Retrieval-Augmented Generation by integrating graph structures for enhanced contextual awareness and efficient information retrieval, achieving better accuracy and response times.

Segment Anything Model 3 achieves state-of-the-art performance in promptable concept segmentation and tracking by leveraging a unified model architecture with decoupled recognition and localization.

PaddleOCR-VL, a vision-language model combining NaViT-style dynamic resolution and ERNIE, achieves state-of-the-art performance in document parsing and element recognition with high efficiency.

A multi-agent framework using large language models for stock trading simulates real-world trading firms, improving performance metrics like cumulative returns and Sharpe ratio.

Agent Lightning is a flexible RL framework for training LLMs in various agents, using a hierarchical RL algorithm and decoupling execution from training to handle complex interactions.

Flawed-Aware Policy Optimization (FAPO) enhances reinforcement learning with verifiable rewards by penalizing flawed-positive rollouts, improving reasoning capability and training stability in large language models.

STARFlow-V, a normalizing flow-based video generator, offers end-to-end learning, robust causal prediction, and high-quality video generation with practical sampling efficiency.

SAM 3D is a generative model that reconstructs 3D objects from single images using a multi-stage training framework that includes synthetic pretraining and real-world alignment, achieving high performance in human preference tests.

HunyuanOCR, a lightweight Vision-Language Model, achieves state-of-the-art performance in OCR tasks through a unified end-to-end architecture combining Vision Transformer and lightweight LLM, supported by data-driven and RL strategies.

MinerU2.5, a 1.2B-parameter document parsing vision-language model, achieves state-of-the-art recognition accuracy with computational efficiency through a coarse-to-fine parsing strategy.

OneThinker, an all-in-one multimodal reasoning model, unifies image and video understanding across various tasks using RL and demonstrates strong performance and knowledge transfer.

GigaWorld-0 is a unified world model framework that integrates video generation and 3D modeling to produce high-quality, diverse, and physically plausible VLA data, enabling strong real-world performance in embodied AI without real-world training.

MinerU is an open-source tool that enhances document content extraction using fine-tuned models and pre/postprocessing rules across diverse document types.

VLASH is an asynchronous inference framework for Vision-Language-Action models that achieves high speed and low latency without sacrificing accuracy, enabling precise robotic tasks like ping-pong and whack-a-mole.

CLaRa enhances retrieval-augmented generation by introducing unified embedding-based compression and joint optimization, achieving state-of-the-art performance in QA benchmarks.

LatentMAS enables efficient, lossless collaboration among LLM agents in latent space, improving performance and reducing computational costs compared to text-based methods.

Zep, a memory layer service, outperforms MemGPT in the DMR benchmark and LongMemEval by excelling in dynamic knowledge integration and temporal reasoning, critical for enterprise use cases.

LlamaFactory is a unified framework enabling efficient fine-tuning of large language models across various tasks using a web-based user interface.

MobiAgent, a comprehensive mobile agent system, achieves state-of-the-art performance in real-world mobile scenarios through its MobiMind-series models, AgentRR framework, and MobiFlow benchmarking suite, while also reducing data annotation costs.

ViSAudio, an end-to-end framework using conditional flow matching, generates high-quality binaural audio from silent video, providing spatial immersion and consistency across various acoustic conditions.

PPTAgent, a two-stage approach, improves presentation generation by analyzing reference presentations and ensuring structural and content consistency, outperforming traditional methods across content, design, and coherence.

GAM, a novel framework that employs JIT compilation principles, improves memory efficiency and task completion by leveraging a lightweight memorizer and researcher in conjunction with reinforcement learning.

A novel GPT-based model, OmniFlatten, enables real-time natural full-duplex spoken dialogue through a multi-stage post-training technique that integrates speech and text without altering the original model's architecture.

Depth Anything 3 (DA3) uses a plain transformer for geometry prediction from visual inputs, achieving state-of-the-art results in camera pose estimation, any-view geometry, visual rendering, and monocular depth estimation.

DeepAnalyze-8B, an agentic LLM, autonomously completes the data science pipeline from raw data to research reports using curriculum-based training and data-grounded trajectory synthesis.

A two-stage deterministic framework, Lotus-2, leverages diffusion models' world priors for high-quality geometric inference, achieving state-of-the-art results in monocular depth estimation and competitive surface normal prediction with limited training data.

DualCamCtrl is a diffusion model for camera-controlled video generation that uses a dual-branch framework and Semantic Guided Mutual Alignment to improve consistency and disentangle appearance and geometry modeling.

MemOS, a memory operating system for Large Language Models, addresses memory management challenges by unifying plaintext, activation-based, and parameter-level memories, enabling efficient storage, retrieval, and continual learning.

IndexTTS, an enhanced text-to-speech system combining XTTS and Tortoise models, offers improved naturalness, enhanced voice cloning, and controllable usage through hybrid character-pinyin modeling and optimized vector quantization.

Vidi, a family of Large Multimodal Models, excels in temporal retrieval for video editing by processing long, multimodal video content and outperforming proprietary models on the VUE-TR benchmark.

Deep Research systems integrate LLMs with external tools to enhance problem-solving capabilities, involving query planning, information acquisition, memory management, and answer generation.

Code2Video generates educational videos using a code-centric agent framework, improving coherence and interpretability compared to direct code generation.

Mem0, a memory-centric architecture with graph-based memory, enhances long-term conversational coherence in LLMs by efficiently extracting, consolidating, and retrieving information, outperforming existing memory systems in terms of accuracy and computational efficiency.

A small orchestrator using ToolOrchestra method coordinates various intelligent tools with reinforcement learning, achieving higher accuracy and efficiency in solving complex tasks like Humanity's Last Exam compared to larger models.

Echo-4o-Image, a synthetic dataset generated by GPT-4o, enhances image generation models by addressing rare scenarios and providing clean supervision, leading to improved performance and transferability.

Today's denoising diffusion models do not "denoise" in the classical sense, i.e., they do not directly predict clean images. Rather, the neural networks predict noise or a noised quantity. In this paper, we suggest that predicting clean data and predicting noised quantities are fundamentally different. According to the manifold assumption, natural data should lie on a low-dimensional manifold, whereas noised quantities do not. With this assumption, we advocate for models that directly predict clean data, which allows apparently under-capacity networks to operate effectively in very high-dimensional spaces. We show that simple, large-patch Transformers on pixels can be strong generative models: using no tokenizer, no pre-training, and no extra loss. Our approach is conceptually nothing more than "Just image Transformers", or JiT, as we call it. We report competitive results using JiT with large patch sizes of 16 and 32 on ImageNet at resolutions of 256 and 512, where predicting high-dimensional noised quantities can fail catastrophically. With our networks mapping back to the basics of the manifold, our research goes back to basics and pursues a self-contained paradigm for Transformer-based diffusion on raw natural data.

Agent0-VL, a self-evolving vision-language agent, incorporates tool usage into both reasoning and self-evaluation, enabling continual improvement through evidence-grounded analysis and reinforcement learning.

A simple head-specific sigmoid gate applied after Scaled Dot-Product Attention improves performance, stability, and scaling in large models, mitigating 'attention sink' and enhancing long-context extrapolation.

MIRIX, a modular multi-agent memory system, enhances language models' memory capabilities by integrating diverse memory types and a dynamic framework, achieving superior performance in multimodal and long-form conversation benchmarks.

InternVL3 is a multimodal pre-trained language model that jointly learns from both multimodal data and text, improving performance and scalability through advanced techniques and setting a new state-of-the-art in multimodal tasks.

Agent0, a self-evolving framework utilizing multi-step co-evolution and tool integration, enhances LLM reasoning capabilities without human-curated data.

The proposed AnyTalker framework generates high-quality multi-person talking videos by extending Diffusion Transformer with identity-aware attention, leveraging single-person videos for training, and using a specialized dataset for evaluation.

GigaBrain-0, a VLA foundation model, uses world model-generated data to enhance cross-task generalization and policy robustness, improving real-world performance on complex manipulation tasks.

EvoVLA, a self-supervised VLA framework, enhances long-horizon robotic manipulation by addressing stage hallucination through triplet contrastive learning, pose-based exploration, and long-horizon memory, achieving improved success rates and sample efficiency on both simulated and real-world tasks.

SimWorld, a new Unreal Engine 5-based simulator, enables the development and evaluation of LLM/VLM agents in realistic, real-world-like settings with diverse physical and social reasoning scenarios.

The PyTorch distributed data parallel module optimizes large-scale model training using techniques like gradient bucketing, computation-communication overlap, and selective synchronization to achieve near-linear scalability.

A large-scale dataset and extensive comparative analysis are provided for multilingual speech translation in the medical domain, enhancing communication and healthcare efficiency.