Halo is an open-source DIY health tracking project that allows users to build their own personalized health monitoring applications based on a low-cost smart ring and open-source software. Halo supports features such as activity tracking, heart rate monitoring, and sleep analysis, and is completely free from...
JoyVASA is an open-source, audio-driven digital head project from JD Health International. Based on diffusion model technology, it generates facial and head movements synchronized with the audio signal. JoyVASA can achieve synchronized lip movements...
TIP-I2V is a large-scale dataset of real-world text and image cues used in image-to-video generation. TIP-I2V contains over 1.7 million unique user-generated text and image cues, along with corresponding videos generated by five state-of-the-art (SOTA) image-to-video models...
FLUX Tools is a suite of modeling tools developed by Black Forest Labs that enhances the controllability and operability of the basic text-to-image model FLUX.1. FLUX Tools includes FLUX.1 Fill (image repair and expansion), FLUX.1 D...
SmoothCache is a general inference acceleration technique for Diffusion Transformers (DiT) models, developed by a research team at Roblox and Queen's University. It is based on analyzing the similarity of layer outputs at adjacent diffusion time steps...
AutoConsis is an intelligent UI consistency detection tool jointly developed by Meituan's technology team and Fudan University. Based on deep learning and large-scale language models, the tool automatically identifies and extracts key data from the interface, detecting and recognizing data...
OmniEdit is an advanced image editing technology that trains a general-purpose model using supervision from multiple expert models to handle various image editing tasks. It can handle images with different aspect ratios and seven different image editing tasks...
In-Context LoRA is an image generation framework based on Diffusion Transformers (DiTs) developed by Alibaba Tongyi Labs. It leverages the model's inherent context learning capabilities to minimize the context generation ability of adjusting the activation model. This...
Add-it is a training-free image editing technology from NVIDIA that adds objects to images based on text instructions. This technology is based on the attention mechanism of the Extended Diffusion Model, integrating scene images, text prompts, and generated images...
DINO-X is a general-purpose large-scale vision model developed by IDEA Research, possessing open-world object detection and understanding capabilities. It supports text, visual, and custom cues, and can recognize any object in an image without user prompts. Based on over 100 million...
The Matrix is the first AI-based world simulator, named after the movie of the same name. It was developed by an all-Chinese team (the authors are from Alibaba, the University of Hong Kong, the University of Waterloo, and the Canadian AI research institution Vector Institute)...
DreamPolish is a text-to-3D generation model developed by Zhipu AI, Tsinghua University, and Peking University. It improves the generation of fine geometry and high-quality textures for complex objects using a two-stage approach. The first stage uses multiple neural representations...
OmniBooth is an image generation framework jointly developed by Huawei Noah's Ark Lab and the Hong Kong University of Science and Technology research team. It supports spatial control and instance-level customization based on text prompts or image references. The framework uses user-defined masks and related...
DynaMem is a dynamic spatial semantic memory system developed by New York University and Hello Robot, designed specifically for mobile operations in open worlds. It maintains a feature point cloud as the robot's memory to handle dynamic changes in the environment...
MVPaint is a 3D texture generation framework jointly developed by Tencent PCG, Shanghai AI Lab, Nanyang Technological University S-Lab, and Tsinghua University. It achieves high-resolution, seamless, and consistent 3D texture generation across multiple views based on synchronous multi-view diffusion technology...
LTX Video is an open-source AI video generation model from Lightricks that can generate 5-second high-quality videos in 4 seconds, exceeding the speed of watching videos. Based on a DiT architecture with 200 million parameters, it ensures smooth motion and structural consistency between frames...
BALROG evaluates the reasoning capabilities of Large Language Models (LLMs) and Visual Language Models (VLMs) in games, particularly their planning, spatial reasoning, and exploration abilities in dynamic environments. It is based on a series of challenging games...
AutoVFX is an advanced physical effects framework developed by a research team at the University of Illinois at Urbana-Champaign. It automatically creates realistic and dynamic visual effects (VFX) videos based on natural language commands. The framework integrates neural scene building...
Markdown-to-Image is an open-source Markdown editor that converts Markdown text into posters. As a React component, it transforms Markdown text content into images, making it suitable for creating social media posts, posters, and other visual content. Tools support...
TÜLU 3 is a series of open-source instruction conformance models released by the Allen Institute for Artificial Intelligence (Ai2), including versions 8B and 70B, with a planned 405B version in the future. The model outperforms Llama 3.1 Instruction in terms of performance...
EchoMimicV2 is a half-body AI digital human project launched by Alibaba's Ant Group. It generates high-quality animated videos based on reference images, audio editing, and hand gesture sequences, ensuring consistency between audio content and half-body movements. EchoMi...
FlipSketch, an innovative system developed by the University of Surrey, transforms static drawings into text-guided sketch animations. The technology is based on three key innovations: frame generation for fine-tuning sketch style, refining with noise while preserving the input sketch view...
StoryTeller is a system jointly developed by ByteDance, Shanghai Jiao Tong University, and Peking University. It improves the quality and consistency of long video descriptions based on audio-visual character recognition technology. The system combines low-level visual concepts with high-level narrative...
DELIFT (Data Efficient Language Model Instruction Fine-Tuning) is a novel algorithm used to optimize large language models (LLMs) across three key stages: instruction fine-tuning, task-specific fine-tuning, and continuous fine-tuning...