GLM-4-long is a 2 million-character long text model launched by Zhipu AI, and its API is now available on the Zhipu AI Open Platform BigModel. It supports extremely long inputs and is designed for handling very long texts and memory-based tasks, with a context length of...
Deep-Live-Cam is an open-source AI real-time face-swapping tool that allows users to swap faces in videos in real time using only a single image. It supports multiple hardware platforms including CPUs, NVIDIA CUDA, Apple Silicon, and Core ML...
Qwen2-Audio is the latest open-source AI speech model released by Alibaba's Tongyi Qianwen team. It supports direct voice input and multilingual text output. It features voice chat and audio analysis capabilities and supports more than 8 languages. Qwen2-Audio...
Imagen 3 is an AI image generation model from Google that can generate high-quality, realistic images. Imagen 3 has better understanding of cue words, captures more details, and generates a wider range of visual styles. Based on Google's text-based image generation technology...
EAFormer is an AI text segmentation framework developed by Fudan University. It excels at finding text in photos and can accurately separate text from the background. Even with blurry text edges or complex backgrounds, EAFormer can perform well...
WiseFlow is an open-source AI information mining tool that intelligently extracts and classifies data from multiple sources, including websites, WeChat official accounts, and social media platforms. It combines statistical learning and large-scale language models (LLM) and supports various news pages...
Agents is an AI agent development tool developed by AIWaves. It allows users to fine-tune and optimize AI models, building more intelligent and personalized AI agents. Through the Agents platform, users can customize AI models...
MiniCPM-V is an open-source, multimodal large-scale model developed by Wallfacer Intelligence. It boasts 8 billion parameters and excels in image and video understanding. MiniCPM-V surpasses models like GPT-4V in single-image understanding and is the first to support implementation on devices such as iPads...
VideoDoodles is an AI video editing framework developed by Adobe in collaboration with several universities. It allows users to easily insert hand-drawn animations into videos, achieving seamless integration with the video content. Through preprocessing video frames, the system provides 2D...
CharacterFactory is an AI character creation tool developed by the IIAU team at Dalian University of Technology. Using a diffusion model, it can quickly generate characters with consistent features without requiring reference images, and only needs 2.5GB of GPU memory and 10 minutes of training...
UniBench is a visual language model (VLM) evaluation framework developed by MetaFair, providing a comprehensive evaluation of VLMs. UniBench includes over 50 benchmark tests covering object recognition, spatial understanding, reasoning, and more...
Retinex-Diffusion is an AI-based image lighting control technology based on Retinex theory, resulting in more natural, nuanced, and layered image brightness and darkness. Retinex-Diffusion does not require model retraining; it works by recognizing elements in the image...
EXAONE 3.0 is an open-source AI model from LG AI Research, featuring 780 million parameters and designed specifically for English and Korean. EXAONE 3.0 performs exceptionally well in both English and Korean language tests, and also excels in real-world use cases and mathematical coding...
DressCode is a 3D clothing generation framework jointly developed by ShanghaiTech University, the University of Pennsylvania, and Deemos Technologies. It allows users to automatically generate 3D clothing models of various styles and materials based on text descriptions.
ReSyncer is an AI video editing tool jointly developed by Tsinghua University and Baidu. It generates high-quality lip-motion videos synchronized with sound through audio-driven processing. ReSyncer uses Style-SyncFormer to analyze sound and create 3D facial models...
MotionClone is a text-driven AI video motion cloning framework that clones actions from reference videos using a temporal attention mechanism and generates new videos by combining text cues. It can handle complex global camera movements and fine-grained local limb movements...
Mini-Monkey is a lightweight multimodal AI model jointly developed by Huazhong University of Science and Technology and South China University of Technology. It employs a multi-scale adaptive segmentation strategy (MSAC) and a scale compression mechanism (SCM) to effectively solve the problems of traditional image segmentation...
MDT-A2G is an AI model jointly developed by Fudan University and Tencent YouTu, specifically designed to synchronously generate corresponding gestures based on speech content. MDT-A2G mimics the natural gestures humans make during communication, enabling computers to create more vivid and...
ASAM (Adversarial Adjustment of Segment Anything Model) is an AI image segmentation model developed by vivo. It enhances the performance of the original SAM model through adversarial adjustments. ASAM uses natural adversarial examples...
Falcon Mamba 7B is an open-source AI model developed by the Technology Innovation Institute (TII) in the UAE, outperforming models such as Meta's Llama 3.1-8B. Falcon Mamba 7B employs an encoder-decoder architecture and multi-head attention...
Clapper is a free and open-source visual AI video editing tool developed by Julian Bilcke, an AI front-end engineer at HuggingFace. It integrates various generative AI technologies, allowing users to interact, iteratively, and intuitively...
UniTalker is an audio-driven 3D facial animation generation model that can generate realistic facial movements based on input audio. It employs a unified multi-head architecture model, uses datasets with different annotations, and supports multiple languages and audio formats...
Grok-2 is a next-generation AI model from xAI, offering superior chat, programming, and reasoning capabilities. In academic benchmark tests, Grok-2 outperforms its predecessor, Grok-1, in areas such as GPQA, MMLU, MMLU-Pro, and MATH...
MagicPose is an AI video generation model jointly developed by the University of Southern California and ByteDance. It generates realistic videos of human movements and facial expressions without any fine-tuning. MagicPose employs a novel two-stage training strategy...