SMooDi - AI Generates Realistic and Stylized Human Motions
Today I read a paper titled SMooDi: Stylized Motion Diffusion Model, which is currently still
Recent Articles
Alibaba's EchoMimic - Generating Portrait Videos
EchoMimic - Lifelike Audio-Driven Portrait Animations through Editable Landmark Conditioning
Alibaba's semantic recognition model SenseVoice and voice generation model CosyVoice
Voice Understanding and Generation Foundation Models for Natural Interaction Between Humans and LLMs
Google's Still-Moving: Generating personalized video content through a few static reference images
A Research Exploration into Customized Video Generation without Customized Video Data
Google's internal tool Smart Paste - Simplifying the code writing workflow by automatically adjusting pasted code
Smart Paste for context-aware adjustments to pasted code
Google's Magic Insert achieves style-aware and realistic insertion effects by dragging into the target image
Magic Insert: Style-Aware Drag-and-Drop
Google DeepMind's Video-to-audio research - dubbing for videos
Generating audio for video
Dify - LLM Application Development Platform
Dify is an open-source LLM app development platform.
The author of ControlNet's new project PaintsUndo - Foundation Model for Human Drawing Behavior
Understand Human Behavior to Align True Needs