Trending Papers

TurboDiffusion accelerates video generation by 100-200x using attention acceleration, step distillation, and quantization, while maintaining video quality.

SHARP synthesizes photorealistic views from a single image using a 3D Gaussian representation, achieving state-of-the-art results with rapid processing.

Step-DeepResearch, an end-to-end agent enhanced with a data synthesis strategy and progressive training, achieves expert-level capabilities in deep research scenarios, outperforming established models.

A self-supervised framework optimizes prompts for both closed and open-ended tasks by evaluating LLM outputs without external references, reducing costs and required data.

Agentic coding tools receive goals written in natural language as input, break them down into specific tasks, and write or execute the actual code with minimal human intervention. Central to this process are agent context files ("READMEs for agents") that provide persistent, project-level instructions. In this paper, we conduct the first large-scale empirical study of 2,303 agent context files from 1,925 repositories to characterize their structure, maintenance, and content. We find that these files are not static documentation but complex, difficult-to-read artifacts that evolve like configuration code, maintained through frequent, small additions. Our content analysis of 16 instruction types shows that developers prioritize functional context, such as build and run commands (62.3%), implementation details (69.9%), and architecture (67.7%). We also identify a significant gap: non-functional requirements like security (14.5%) and performance (14.5%) are rarely specified. These findings indicate that while developers use context files to make agents functional, they provide few guardrails to ensure that agent-written code is secure or performant, highlighting the need for improved tooling and practices.

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

SmolDocling is a compact vision-language model that performs end-to-end document conversion with robust performance across various document types using 256M parameters and a new markup format.

A novel framework, Robust-R1, enhances multimodal large language models' robustness to visual degradations through explicit modeling, supervised fine-tuning, reward-driven alignment, and dynamic reasoning depth scaling, achieving state-of-the-art performance on real-world degradation benchmarks.

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.

PagedAttention algorithm and vLLM system enhance the throughput of large language models by efficiently managing memory and reducing waste in the key-value cache.

SAM Audio, a diffusion transformer-based foundation model, achieves superior performance in general audio separation using unified text, visual, and temporal span prompts across various audio types.

StoryMem enhances multi-shot video generation with cinematic quality and long-range consistency using a memory bank and pre-trained single-shot video diffusion models.

Unified Autoencoding combines semantic and pixel-level information through a frequency-band modulator, resulting in a latent space with state-of-the-art performance on image benchmarks.

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.

VibeVoice synthesizes long-form multi-speaker speech using next-token diffusion and a highly efficient continuous speech tokenizer, achieving superior performance and fidelity.

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

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.

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

DINOv3, a self-supervised learning model, achieves superior performance across various vision tasks by scaling datasets and models, addressing dense feature degradation, and enhancing flexibility with post-hoc strategies.

DeepCode, a fully autonomous framework, addresses the challenges of document-to-codebase synthesis by optimizing information flow through source compression, structured indexing, knowledge injection, and error correction, achieving state-of-the-art performance and surpassing human experts.

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.

FunAudioLLM enhances voice interactions by integrating SenseVoice for multilingual speech recognition, emotion detection, and audio event detection with CosyVoice for natural speech generation across languages, timbres, and styles.

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.

EgoX framework generates egocentric videos from exocentric inputs using video diffusion models with LoRA adaptation, unified conditioning, and geometry-guided self-attention for coherence and visual fidelity.

DSR Suite enhances vision-language models with dynamic spatial reasoning through automated data generation and a geometry selection module that integrates geometric priors.

High-resolution video generation, while crucial for digital media and film, is computationally bottlenecked by the quadratic complexity of diffusion models, making practical inference infeasible. To address this, we introduce HiStream, an efficient autoregressive framework that systematically reduces redundancy across three axes: i) Spatial Compression: denoising at low resolution before refining at high resolution with cached features; ii) Temporal Compression: a chunk-by-chunk strategy with a fixed-size anchor cache, ensuring stable inference speed; and iii) Timestep Compression: applying fewer denoising steps to subsequent, cache-conditioned chunks. On 1080p benchmarks, our primary HiStream model (i+ii) achieves state-of-the-art visual quality while demonstrating up to 76.2x faster denoising compared to the Wan2.1 baseline and negligible quality loss. Our faster variant, HiStream+, applies all three optimizations (i+ii+iii), achieving a 107.5x acceleration over the baseline, offering a compelling trade-off between speed and quality, thereby making high-resolution video generation both practical and scalable.

FlashPortrait is a diffusion-based video transformer for long-portrait animation that ensures ID consistency and achieves 6x acceleration through a dynamic sliding-window scheme and higher-order latent derivatives.

WorldPlay is a streaming video diffusion model that achieves real-time, interactive world modeling with long-term geometric consistency by using a Dual Action Representation, Reconstituted Context Memory, and Context Forcing.

GLM-4.5, a Mixture-of-Experts large language model with 355B parameters, achieves strong performance across agentic, reasoning, and coding tasks using multi-stage training and reinforcement learning.

This survey provides an updated overview of agent memory research, distinguishing its forms, functions, and dynamics, and highlights emerging research directions.

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.

Agent S, a framework for autonomous GUI interactions, enhances task automation with experience-augmented hierarchical planning and Multimodal Large Language Models.

Agent S2, a compositional framework using Mixture-of-Grounding and Proactive Hierarchical Planning, achieves state-of-the-art performance in computer use automation across various benchmarks and operating systems.

Behavior Best-of-N (bBoN) improves the reliability and success rates of computer-use agents by generating and selecting among multiple rollouts using behavior narratives, achieving state-of-the-art performance on OSWorld and strong generalization to different operating systems.

3D-RE-GEN reconstructs single images into modifiable 3D textured mesh scenes with comprehensive backgrounds, achieving top performance through compositional generation and scene optimization.

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.

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 framework for Scientific General Intelligence (SGI) is presented, evaluated using SGI-Bench, and improved with Test-Time Reinforcement Learning, highlighting gaps in existing models' scientific capabilities.

An automated pipeline using layout-aware OCR and LLM-based semantic parsing extracts high-quality QA and VQA pairs from educational documents, enhancing LLM training without synthetic data.

Single-stream Policy Optimization (SPO) improves policy-gradient training for Large Language Models by eliminating group-based issues and providing a stable, low-variance learning signal, leading to better performance and efficiency.

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.

PersonaLive is a diffusion-based framework for real-time portrait animation that enhances speed and efficiency through multi-stage training, hybrid implicit signals, appearance distillation, and autoregressive micro-chunk streaming.

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.

HunyuanVideo 1.5 is a lightweight video generation model with state-of-the-art visual quality and motion coherence, using a DiT architecture with SSTA and an efficient video super-resolution network.

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.

Pixio, an enhanced masked autoencoder, demonstrates competitive performance across various downstream tasks using pixel-space self-supervised learning, outperforming latent-space approaches.

RAG-Anything is a unified framework that enhances multimodal knowledge retrieval by integrating cross-modal relationships and semantic matching, outperforming existing methods on complex benchmarks.

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.

A unified visual tokenizer pre-training framework (VTP) improves generative performance by optimizing image-text contrastive, self-supervised, and reconstruction losses, leading to better scaling properties and higher zero-shot accuracy and faster convergence.