Claude 3.5 Haiku is a new artificial intelligence model launched by Anthropic. While maintaining similar speed and cost to its predecessor, Claude 3 Haiku, it achieves improvements in various capabilities, including...
PUMA is an advanced multimodal large-scale language model (MLLM) designed to unify and enhance visual generation and understanding tasks based on integrated multi-granularity visual features. PUMA can handle tasks ranging from text-to-image generation, detailed image editing, and more...
Janus is an autoregressive framework developed by DeepSeek AI, aiming to unify multimodal understanding and generation tasks. It addresses the limitations of previous methods by separating visual encoding into different paths and uses a single transformer architecture for processing...
PaddleOCR 2.9 is an open-source optical character recognition (OCR) toolkit developed by Baidu PaddlePaddle. It offers a rich set of algorithms and models, supports multiple languages, and provides data annotation and synthesis tools. P...
Mellum is a Large Language Model (LLM) developed by JetBrains, designed specifically for developers to enhance AI-driven development tools. Through deep integration with the JetBrains IDE, it provides low-latency, high-accuracy code...
Easegen is an open-source AI-powered digital human course creation project that integrates a complete solution from course production and video management to intelligent courseware generation and intelligent question generation. The project supports batch automatic generation of PPT courseware, digital human cloning, voice...
sCM is a continuous-time consistency model introduced by OpenAI, based on the principle of diffusion model. sCM simplifies the theoretical framework and optimizes the sampling process, achieving a significant improvement in image generation speed. The sCM model can generate images in just two sampling steps...
Moonshine is a speech recognition model optimized for resource-constrained devices, providing fast and accurate real-time speech-to-text services. It is suitable for applications requiring immediate response, such as on-site transcription and voice command recognition. Moonshine...
Time-MoE is an innovative time series foundational model based on a Mixture of Experts (MoE) architecture, scaling the parameter scale of pre-trained time series models to the billion-level. The model uses a sparse activation mechanism...
SynthID is a technology tool developed by DeepMind that uses digital watermarks embedded in AI-generated content to aid in content identification. Watermarks are imperceptible to humans but detectable by machines, thus increasing the trustworthiness of information. Syn...
Whispo is an AI-powered speech transcription tool that allows users to record speech by holding down the Ctrl key and automatically inserting the transcribed text into any text-input-enabled application when the key is released. The tool's data processing is entirely...
AutoGLM-Web is an intelligent browser assistant that simulates user actions. Built on a large language model, it can perform tasks such as web page access, information retrieval, and content summarization. Based on simple text or voice commands, AutoGLM-Web can...
OmniGen is a novel diffusion model for unified image generation, capable of handling various image generation tasks within a single framework, such as text-to-image generation, image editing, topic-driven generation, and visual conditional generation. OmniGen involves...
CogAgent is a multimodal visual model jointly developed by Tsinghua University and Zhipu AI, focusing on the understanding and navigation of graphical user interfaces (GUIs). It perceives GUI interfaces through visual modalities, rather than traditional text-based modalities, making it more suitable for...
OMNE Multiagent is a large-scale multi-agent framework developed by the Tianqiao and Chrissy Chen Institute (TCCI). It is built upon Long Term Memory (LTM)...
DuoAttention is a novel framework proposed by Han Song's team at MIT to improve the inference efficiency of large language models (LLMs) when handling long contexts. It optimizes the model by distinguishing between two attention heads: a "retrieval head" and a "streaming head"...
HuggingChat for macOS is an open-source chat application from Hugging Face, designed specifically for macOS users. Based on a powerful open-source language model, it brings advanced AI conversational capabilities directly to the user's desktop. The application supports...
OmniParser is a screen parsing tool developed by Microsoft Research that converts user interface screenshots into structured data. The tool is specifically designed to improve the performance of UI proxy systems based on large language models (such as GPT-4V)...
Ditto is a Flask-based application generator that allows users to create multi-file web applications based on simple natural language descriptions. The tool automates the coding process using a simple LLM loop and some auxiliary tools, transforming the user's...
Flux.1 Lite is a lightweight AI model developed by the Freepik team, currently in alpha mode. It's based on an 8 billion parameter transformer architecture and is derived from the FLUX.1-dev model. Flux.1 Lite...
Phidata is an open-source AI agent framework that helps developers build intelligent agent systems with memory, knowledge, tools, and reasoning capabilities. It supports the creation of collaborative agent teams and provides a user interface for interacting with agents...
Allegro is an advanced text-to-video generation model developed by Rhymes AI. It can convert simple text input into high-quality video content up to 720p resolution, 15 frames per second, and up to 6 seconds in length. The model excels in the field of video generation...
FakeShield is a multimodal large-scale language model framework developed by researchers at Peking University, capable of detecting and locating image forgeries. The framework can assess the authenticity of images, generate masks of tampered regions, and provide pixel-level and image-based...
MaskGCT is a large-scale speech synthesis model developed by FunWan Technology in collaboration with the Chinese University of Hong Kong, Shenzhen. Based on a mask generation model and decoupled encoding of speech representation, it enables applications such as voice cloning, cross-language synthesis, and voice control.