I2VGen-XL is an open-source image-to-video generation model launched by Alibaba DAMO Academy. It decouples text-to-video data from video structure through an innovative cascading diffusion method, while utilizing still images as key guiding elements...
MagicVideo-V2 is an AI video generation model developed by ByteDance. It integrates a text-to-image model, a video motion generator, a reference image embedding module, and a frame interpolation module into an end-to-end video...
Motionshop is an AI character animation framework developed by Alibaba Research Institute of Intelligent Computing. It can replace characters in a video with 3D avatars without altering other scenes or characters in the video. This framework utilizes video processing...
AnyText is a diffusion-based multilingual visual text generation and editing model developed by the Alibaba Research Institute of Intelligent Computing. It focuses on rendering accurate and coherent text in images. AnyText comprises two main elements...
InstantID is an image generation technology based on a diffusion model, focusing on achieving zero-shot identity-preserving personalized image synthesis. This technology allows users to use only one...
VideoCrafter2 is a video generation model developed by Tencent AI Labs. It aims to overcome the limitations of acquiring high-quality video data and train a model capable of generating high-quality videos. The core idea of this model is to...
Sscreenshot to Code is an open-source project that uses artificial intelligence technologies (GPT-4V and DALL·E 3) to convert users' screenshots into front-end web page code. The core function of the project is to automate the coding of web page designs...
DDColor is an open-source AI image colorization framework developed by researchers at Alibaba DAMO Academy. It can transform black and white images into full-color images with a single click. This method utilizes a dual-decoder architecture (pixel decoder and color decoder)...
Real-ESRGAN (Real-World Blind Super-Resolution with Pure Synthetic Data) is a deep learning model developed by Tencent's research team, specifically designed for processing and upscaling low-resolution images to...
DemoFusion is a technical framework designed for low-cost high-resolution image generation. It extends existing open-source generative AI models (such as Stable Diffusion) to enable these models to generate high-resolution images without additional training...
ActAnywhere is a video generation model developed jointly by researchers at Stanford University and Adobe Research. It aims to solve the problem of video background generation, especially when it's necessary to blend foreground subjects (such as people) with a new background...
Vary-toy is a small visual language model (LVLM) proposed by researchers from Megvii, the University of Chinese Academy of Sciences, and Huazhong University of Science and Technology of China. It aims to address the challenges of training and deploying large visual language models (LVLMs). For...
Lepton Search is an open-source conversational AI search engine developed by Jia Yangqing, former Vice President of Technology and AI scientist at Alibaba, from the Lepton AI application building platform. The project is based on leveraging cloud-based artificial intelligence on the Lepton platform...
Follow Your Pose is an open-source text-to-video generation framework developed by researchers from Tsinghua University, Hong Kong University of Science and Technology, Tencent AI Lab, and the Chinese Academy of Sciences. It allows users to generate videos by providing text descriptions and specified poses...
The IP-Adapter (Image Prompt Adapter) is an adapter specifically designed for pre-trained text-to-image diffusion models (such as Stable Diffusion), aiming to enable text-to-image models to generate...
OLMo (Open Language Model) is a fully open-source large-scale language model (LLM) framework developed by Allen AI (AI2, Allen Institute for AI). Its initial design aim to promote academic collaboration through open research...
DiffusionGPT is an open-source, large model (LLM)-driven text-to-image generation system developed by researchers from ByteDance and Sun Yat-sen University. It aims to address the limitations of text-to-image generation systems in handling diverse inputs or being restricted to certain types of data...
MetaGPT is an innovative metaprogramming framework that combines large language models and multi-agent collaborative systems, aiming to solve complex problems by simulating human workflows. At its core lies the standardization of operating procedures (SOPs)...
Make-A-Character (Mach for short) is an AI-powered 3D digital human generation framework developed by Alibaba Group's Intelligent Computing Research Institute. It aims to quickly create realistic 3D characters based on text descriptions. This system is particularly...
MotionCtrl is a unified and flexible motion controller designed for video generation models, developed by researchers from Tencent, the University of Hong Kong, and other institutions. It can independently control camera movement and object motion perspective within a video...
AnimateDiff is a framework developed by researchers from the Shanghai Artificial Intelligence Laboratory, the Chinese University of Hong Kong, and Stanford University. It extends personalized text-to-image models into animation generators, and its core feature is its ability to utilize...
Depth Anything is a deep learning model designed for monocular depth estimation, developed by researchers from TikTok, the University of Hong Kong, and Zhejiang University. It aims to handle images in various scenarios and estimate their depth information. The...
V-JEPA, developed by researchers at Meta, is a novel video self-supervised learning method that focuses on learning visual representations of videos through feature prediction. The core idea of this method is to enable the model to predict a...
Boximator is a video compositing technology developed by ByteDance's research team, designed to generate rich and controllable motion to enhance the quality and controllability of video compositing. This technology introduces two types of constraint boxes (hard boxes...