Hello everyone.
So, today’s generative AI news.
Microsoft’s Copilot has evolved significantly.
For example, Copilot announced a feature that processes Excel data and generates Python code.
It seems that it has also evolved in Excel, Word, and Outlook.
Also, a chatbot has appeared in Copilot Studio.
■Dream Machine API
An API that enables developers to build and scale creative products using Dream Machine without having to build complex tools in-app.
Luma Dream MachineDream Machine is an AI model that makes high quality, realistlumalabs.ai
■ComfyWeb
A simple interface for ComfyUI, with intuitive form-driven navigation and an integrated gallery view.
■stable-diffusion-webui-simple-manga-maker
It is a browser application specialized for manga production based on Stable Diffusion WebUI (Forge). It is equipped
with all the functions necessary for manga production, such as intuitive operation, panel division, speech bubble creation, image editing, text input, and effects.
■Consistory
NVIDIA has released a model that generates consistent images without fine-tuning or training.
However, I couldn’t access the site because it was too busy at the moment.
✨ 🖼️ Generate images with consistent characters without any fine tuning nor training. Consistory is a training-free approach to maintaining subject consistency between text-to-image generations on pretrained models. #NVIDIAResearch
— NVIDIA AI Developer (@NVIDIAAIDev) September 16, 2024
🎨 Test it here: https://t.co/ZcWewEJssA pic.twitter.com/Kiwl9974k6
■Seed-Music
It is a framework for a music generation system that can generate high-quality music with diverse style control.
It has features such as multilingual expressive vocal music generation, fine-grained note-level music audio editing, and zero-shot singing voice conversion.
■Fluxforge
Fluxforge updated to v0.2 with
resolution, prompt intensity, upscaling, UI fixes, and more.
■DrawingSpinUp
A workflow that generates lifelike 3D animation from a single character drawing.
Reconstruct a 3D model from the character drawing and retarget 3D motion data to it to achieve more three-dimensional animation.
■InstantDrag
A workflow that improves the interactivity and speed of drag-based image editing. It
only requires the image and drag instructions as input, and no additional inputs such as masks or text prompts, allowing for fast, realistic editing without sacrificing interactivity.
InstantDrag for fast and light Drag Editing. arXiv2024joonghyuk.com
■NPGA
A data-driven framework for creating high-quality, controllable avatars from multi-view video recordings.
■DualGS
A workflow for accurately tracking complex human movements and generating high-quality, immersive human-centric volumetric videos.
■Stable Pixel
An early preview of Stable Pixel (desktop) has been announced.
Microsoft Formsforms.office.com
■Saturated audio with Stable Audio
This is a workflow to generate sound effects from images using ComfyUI.
■Llama3.1-70B
Llama3.1-70B weight with AQLM-PV compression has been released.
https://huggingface.co/ISTA-DASLab/Meta-Llama-3.1-70B-AQLM-PV-2Bit-1×16
■sakana.ai
Sakana.ai announced that it has raised funds from 10 domestic companies, including Mitsubishi UFJ Financial Group, three megabank groups, and NEC.
It has secured a total of about 30 billion yen through a series of funding rounds, including from the US semiconductor giant NVIDIA.
Sakana AISakana AI Blogsakana.ai
That ‘s all for today.
See you next time.