Imagen3 Released
Google has officially launched Imagen3, their image-generating AI. It’s now available to all Gemini users around the world on Gemini Advanced. Imagen3 produces very high-quality images and is more accurate than other image-generating AIs.
Veo for Video Generation
In addition, Google has introduced “Veo,” an AI that generates videos.
SyncTweedies Framework
SyncTweedies is a new system that helps create different types of images using pre-trained models. It works by syncing several image diffusion processes to create fuzzy images, panoramic views, textures on 3D shapes, and more.
https://synctweedies.github.io/
https://synctweedies.github.io/static//synctweedies.pdf
For more info, check out SyncTweedies here:
https://github.com/KAIST-Visual-AI-Group/SyncTweedies
DressRecon Technology
DressRecon is a tool that builds realistic 3D human body models from regular videos. It focuses on how people interact with loose clothing and objects. By combining general body shape data with specific video details, it achieves high-quality 3D reconstructions.
https://jefftan969.github.io/dressrecon/
Learn more about DressRecon here:
https://github.com/jefftan969/dressrecon
Pyramidal Flow Matching
Pyramidal Flow Matching is a fast video generation model that uses open-source datasets for training. It helps create videos more efficiently.
https://pyramid-flow.github.io/
You can find the Pyramidal Flow Matching code here:
https://github.com/jy0205/Pyramid-Flow
https://huggingface.co/rain1011/pyramid-flow-sd3
Gmail Update
Gmail has added a new feature! You can now reply to and summarize emails right on the same screen, thanks to Gemini’s chat feature.
LLavaImageTagger
LLavaImageTagger is a helpful tool that automatically organizes your images. If you have many images without names, it analyzes them, generates keywords, and adds them to the metadata.
Check it out here:
https://github.com/jabberjabberjabber/LLavaImageTagger
Introducing ARIA
ARIA is a new AI model that can handle different types of information, like images, text, and code. It excels in understanding videos and documents.
Find out more about ARIA here:
https://huggingface.co/rhymes-ai/Aria
FluxBooru Updates
FluxBooru v0.1 focuses on generating “non-aesthetic” content. It’s been improved to create more interesting images.
See it here: FluxBooru Hugging Face.
https://huggingface.co/spaces/bghira/FluxBooru-CFG3.5
https://huggingface.co/terminusresearch/flux-booru-CFG3.5
https://huggingface.co/terminusresearch/flux-booru-CFG3.5
Flux Fusion Update
The new version of Flux Fusion Merge has been released, making it faster to generate images.
https://civitai.com/models/630820
Llama 3.1 in Medicine
Llama 3.1 is being trained with synthetic data from doctors. This makes it better at tasks in the medical field, like answering questions and summarizing medical research.
Kaiber Update
Kaiber, the AI video generation service used in a Linkin Park music video, is getting an update. Keep an eye out for the preview!
https://kaiber.ai/comingsoon
Sign up for updates here: Kaiber.
https://kaiber.ai/
Bolt Update
The AI tool Bolt, which helps create web applications in Japanese, has been updated. You can now purchase tokens to use the service.
Blinkshot Consistency Mode
Blinkshot, which generates images in real-time, now has a consistency mode. This keeps the background and subject the same while generating images.
Check it out: Blinkshot.
Fotographer AI Demo Video
Here’s a demo video of “Fotographer AI,” an automatic product image generation service that recreates scenes from Tokyo.
Learn more about Fotographer.ai: Fotographer.
https://fotographer.ai/en/
LM Studio 0.3.4 Link
You can find the link to “LM Studio 0.3.4,” a tool that runs large language models quickly on Apple Silicon Macs.
Check it out here: LM Studio.
https://lmstudio.ai/
Generative AI Exhibition
The “Generative AI Anything Exhibition” is looking for exhibitors. You can apply in different categories like Audio, Image, Language, and more.
Find more information here: Exhibition Form.
https://docs.google.com/forms/d/e/1FAIpQLScdUKqCq1sDbelwYmUEp0VDwamksIFEhjPE-7qxCjAGr-SSNQ/viewform
Meta’s AI Expansion
Meta has announced that its conversational AI is now available in 21 new markets.
That’s all for today! It seems like Google is really stepping up! Keep in mind, you need to sign up for Gemini Advanced, which costs 2,900 yen a month, but you can try it for free for the first month. It might be worth checking out!
