VtripGPT is an AI model developed by Vistrip Technology, focusing on the tourism industry. It provides intelligent support for generating tourism dialogues based on deep synthesis services. The model is based on the Transformer architecture, combining tourism industry data and general knowledge...
D-Edit is a versatile image editing framework based on both images and text. It utilizes a pre-trained diffusion model and unique prompts to achieve precise control and editing of specific items within an image. The framework can handle image-based...
PromptFix is an open-source AI image restoration tool developed by Microsoft. Based on diffusion model technology, it can process damaged images according to user instructions and remove unwanted elements. PromptFix supports various image processing tasks, including colorization...
Show-o is a unified Transformer model that integrates multimodal understanding and generation. By combining autoregressive and discrete diffusion modeling, it can flexibly handle tasks including visual question answering, text-to-image generation, text-guided repair and expansion, and more...
MimicTalk, jointly developed by Zhejiang University and ByteDance, is based on NeRF (Neural Radiation Field) technology. It can train a personalized and expressive 3D speaking face model in just 15 minutes. Mimi...
Wonder Animation is an innovative AI technology developed by Wonder Dynamics, a subsidiary of Autodesk. The beta version of Wonder Animation is now available in Wonder Studio. Wonder Animation is based on Video to 3D...
TimeSuite is a novel framework developed by the Shanghai AI Lab that enhances the performance of Multimodal Large Language Models (MLLMs) in long video understanding tasks. It leverages an efficient long video processing framework and the high-quality video dataset Tim...
EMMA is Waymo's end-to-end multimodal autonomous driving model based on the Gemini model. It directly maps raw camera sensor data to driving-specific outputs, such as planned trajectories, perceived objects, and road map elements. EMMA integrates non-traditional...
DreamVideo-2 is an innovative zero-sample video customization framework jointly developed by Fudan University, Alibaba Group, and other institutions. DreamVideo-2 can generate videos with specific themes and precise motion trajectories from a single image and a sequence of bounding boxes...
NotebookMLX is an open-source version of NotebookLM, integrating the functionality of NotebookLlama. It can convert PDF documents into easily understandable and shareable audio podcast formats. The project implements natural language processing capabilities based on MLX technology, including...
HOVER is a 1.5M miniature model launched by NVIDIA. Its full name is "Humanoid Versatile Controller," meaning a multi-functional, whole-body neural controller for humanoid robots. The model uses 1.5 million parameters to achieve complex robot movements...
Voice Changer is a new model from Cartesia that can convert the speech of any audio clip into other timbres while preserving the emotion and expression of the original audio. Users can choose from a variety of high-quality voice libraries provided by Cartesia...
Oasis is the world's first AI-generated game in real time, jointly developed by Decart and Etched. The game renders interactive video content in real time at 20 frames per second, generated directly by an AI model without a game engine. Players can...
Quark Intelligence is a brand-new intelligent learning model launched by Quark Technologies, possessing doctoral-level reasoning capabilities. Based on AI technology, it provides users with step-by-step explanations of problems and can answer questions at any time. It is particularly useful for solving problems in postgraduate entrance exam mathematics, etc...
OSAID 1.0 (The Open Source AI Definition – 1.0) is an official standard released by the Open Source Initiative (OSI) to clarify the conditions for AI systems to become open source. Those involved in developing OSAID 1...
MobileLLM is a large language model with fewer than one billion parameters, optimized by Meta for mobile device use cases, addressing issues of rising cloud costs and latency. MobileLLM is based on a deep-thin architecture, embedded sharing, and a grouped query attention mechanism...
SmolLLM2 is a compact, large-scale language model from Hugging Face for on-device applications. It offers three different parameter levels: 1.7B, 360M, and 135M, to adapt to various application scenarios and resource constraints. The model's understanding...
Genmoai-smol is an open-source video generation model, a fork of Genmoai's txt2video model, optimized for single-GPU devices to reduce memory usage and enable video creation on resource-constrained devices. The model uses high...
VirSci (Virtual Scientists) is a multi-agent AI scientific research tool launched by the Shanghai Artificial Intelligence Laboratory. It accelerates scientific innovation by simulating the collaborative process of a team of scientists. The system is based on large-scale language models (LLMs)...
Self-Lengthen is an innovative iterative training framework developed by Alibaba's Qianwen team, which enhances the ability of large language models (LLMs) to generate long texts. The framework is based on two roles: a generator and an expander, working together to generate...
Amphion is an open-source audio, music, and speech generation toolkit, jointly developed by the team of Associate Professor Wu Zhizheng at the Chinese University of Hong Kong, Shenzhen, in collaboration with the Shanghai Artificial Intelligence Laboratory and the Shenzhen Big Data Research Institute. The toolkit supports...
LongReward is a method jointly developed by Tsinghua University, the Chinese Academy of Sciences, and Zhipu AI to improve the performance of large language models (LLMs) for long texts based on AI feedback. LongReward evaluates usefulness, logicality, fidelity, and completeness...
Fish Agent is an innovative end-to-end speech processing model from FishAudio, integrating Automatic Speech Recognition (ASR) and Text-to-Speech (TTS) technologies. It achieves speech-to-speech processing without the need for traditional semantic encoders/decoders...
MeetingMind is an AI-powered meeting assistant that automatically transcribes meeting audio based on recording and file upload functions, extracting key information such as tasks, decisions, and questions to help users easily capture and analyze meeting content and take action...