Perception-as-Control is an image and animation framework developed by Alibaba Tongyi Labs, capable of achieving fine-grained motion control based on user intent. Perception-as-Control is based on building a 3D perceptual motion representation, and...
Motion Dreamer is a video generation framework proposed by researchers at the Hong Kong University of Science and Technology (Guangzhou) that generates videos with reasonable motion. Based on a two-stage generation method, it first generates an intermediate motion representation based on the input image and motion conditions, and then utilizes...
Sky-T1 is an open-source inference AI model released by NovaSky, a research team at the Sky Computing Lab at UC Berkeley. It is named Sky-T1-32B-Preview. It is the first open-source inference model, with both its training dataset and code...
FaceLift is a single-image to 3D head model conversion technology developed by Adobe and the University of California, Merced. It can reconstruct a 360-degree head model from a single face image. FaceLift is based on a two-stage process: ...
SVFR (Stable Video Face Restoration) is a unified framework for generalized video face restoration, jointly developed by Tencent YouTu Lab and Xiamen University. It integrates video face restoration (BFR), colorization, and restoration tasks...
TimesFM 2.0 is an open-source time series forecasting model developed by Google's research team. It boasts powerful forecasting capabilities, handling univariate time series with up to 2048 time points and supporting arbitrary forecast time spans. The model employs a solution-only approach...
NeuralSVG is an innovative text-driven vector graphics generation technology. It transforms text cues into hierarchical vector graphics using a small multilayer perceptron network. The network takes shape indices as input and outputs shape parameters...
ViTPose is a human pose estimation model based on the Transformer architecture. It uses a standard visual Transformer as its backbone network, extracting features by segmenting the input image and feeding it into Transformer blocks, then...
Nemotron-CC is a large-scale pre-trained dataset released by the NVIDIA team, which can transform Common Crawl data into high-quality corpora suitable for long-sequence pre-training. The Nemotron-CC dataset combines classifier ensembles, synthetic data...
NVIDIA Ingest is an open-source collection of microservices from NVIDIA for parsing complex, messy, unstructured PDFs and other enterprise documents. NVIDIA Ingest converts documents into metadata and text, making them easy to embed into retrieval systems...
Eko is a production-ready JavaScript framework developed by Fellou AI. Based on natural language, it helps developers easily create reliable intelligent agents ranging from simple instructions to complex processes. Eko supports all platforms, including...
Emotion-LLaMA is a multimodal emotion recognition and inference model that integrates audio, visual, and text inputs, combining information through a specific emotion encoder. The model is based on a modified version of LLaMA, adjusted with instructions to improve emotion recognition capabilities...
The GR00T-Teleop is part of the NVIDIA Isaac GR00T series and utilizes key technologies for capturing remote operation data. Based on NVIDIA CloudXR connectivity with the Apple Vision Pro headset, it uses a specially designed humanoid remote operation system...
MiniRAG is a novel retrieval-enhanced generative (RAG) system developed by the University of Hong Kong, designed for the efficient deployment of small language models (SLMs) in resource-constrained scenarios. MiniRAG achieves this goal based on two key technologies: ...
Riona-AI-Agent is an AI-driven automation tool built with Node.js and TypeScript that interacts with social media platforms such as Instagram, Twitter, and GitHub. Riona-AI-Agent can generate...
Prometheus is an innovative 3D-aware latent diffusion model specifically designed for rapidly generating text-to-3D scene content. It can complete object- and scene-level 3D generation in seconds while maintaining high-quality output and good generalization...
MinMo is a multimodal large-scale model developed by the FunAudioLLM team at Alibaba's Tongyi Lab, focusing on achieving seamless voice interaction. MinMo boasts approximately 8 billion parameters, trained through multi-stage training on 1.4 million hours of diverse voice data...
MiniMax-01 is a brand new series of models launched by MiniMax, including the basic language large-scale model MiniMax-Text-01 and the visual multimodal large-scale model MiniMax-VL-01. MiniMax-01 is the first to implement a linear attention mechanism on a large scale, achieving...
iFlytek Spark X1, released by iFlytek on January 15, 2025, is China's first large-scale model with deep thinking and reasoning capabilities trained on a fully domestically produced computing platform. In its problem-solving process, it more closely resembles the human "slow thinking" method, using only...
The Spark Simultaneous Voice Interpretation Model, released by iFlytek on January 15, 2025, is China's first large-scale model with end-to-end simultaneous voice interpretation capabilities. The model leads the industry in content completeness, information accuracy, and language quality...
Titans is a new neural network architecture introduced by Google that overcomes the memory bottleneck of Transformer when processing long sequences of data. Titans introduces a neural long-term memory module, simulating the human brain's memory mechanism, and is particularly enhanced to handle unexpected events...
moonshot-v1-vision-preview is a multimodal image understanding model developed by Dark Side of the Moon. It enhances the multimodal capabilities of the moonshot-v1 model series, possessing powerful image recognition abilities and the capacity to accurately distinguish complex images...
MiniCPM-o 2.6 is the latest and highest-performing multimodal large model in the MiniCPM-o series, featuring 8B parameters. MiniCPM-o 2.6 excels in multiple areas, including vision, speech, and multimodal live streaming, achieving performance comparable to GPT-4o...
Flowise is an open-source, low-code/no-code drag-and-drop tool that allows users to quickly build custom LLM applications by dragging and dropping visual components. It provides a set of easy-to-use tools and components, requiring little to no coding...