AI Art Weekly #128
Hello, my fellow dreamers, and welcome to issue #128 of AI Art Weekly! 👋
This week I kinda got addicted to Suno 4.5. Always wanted to produce my own music and their new 4.5 model feels like AI is finally catching up to realize that dream, so I’m brewing something new that also satisfies my Midjourney urges 😏
Creating music videos was actually something that got me into playing with AI three years ago. Even got commissioned one time by The Winter Sounds to create a deforum video. Something I just now realize I want to get back into again.
So I’m off vibe-creating music videos. Have a great weekend ✌️
Support the newsletter and unlock the full potential of AI-generated art with my curated collection of 325+ high-quality Midjourney SREF codes and 3000+ creative prompts.
News & Papers
Highlights
FLUX Kontext dev
Black Forest Labs released the weights for the FLUX.1 Kontext [dev] model. In case you missed this: the Kontext model series is for text-based image editing that retains the image structure and only changes the part you describes. The cool part compared to the Pro and Ultra model is that the dev model can be fine-tuned and can be combined with LoRAs.

FLUX.1 Kontext [dev] examples
3D
AnimaX: Animating the Inanimate in 3D with Joint Video-Pose Diffusion Models
AnimaX can animate rigged 3D characters using video-pose diffusion models based on text prompts.

AnimaX example
SeamGPT: Auto-Regressive Surface Cutting
SeamGPT can generate cutting seams for UV flattening and part decomposition.

SeamGPT example
Assembler: Scalable 3D Part Assembly via Anchor Point Diffusion
Assembler can reconstruct complete 3D objects from part meshes and a reference image.

Assembler example
BecomingLit: Relightable Gaussian Avatars with Hybrid Neural Shading
BecomingLit can create high-resolution head avatars that can be relit and animated from single videos.

BecomingLit example
Image
XVerse: Consistent Multi-Subject Control of Identity and Semantic Attributes via DiT Modulation
XVerse can create high-quality images with multiple subjects that can be edited. It allows precise control over each subject’s pose, style, and lighting, while also reducing issues like attribute entanglement and artifacts.

XVerse example
TAIR: Text-Aware Image Restoration with Diffusion Models
Text-Aware Image Restoration can restore images and retain the accuracy of text in them.

TAIR example
Video
Matrix-Game: Interactive World Foundation Model
Matrix-Game can generate high-quality interactive game worlds in Minecraft.

Matrix-Game example
OmniAvatar: Efficient Audio-Driven Avatar Video Generation with Adaptive Body Animation
OmniAvatar can generate lifelike full-body avatar videos from audio. It offers accurate lip-syncing and natural movements, and allows for precise control over emotions and backgrounds.

OmniAvatar example
FairyGen: Storied Cartoon Video from a Single Child-Drawn Character
FairyGen can turn a child’s drawing into a story-driven cartoon video while keeping its unique style.

FairyGen example
Hunyuan-GameCraft: High-dynamic Interactive Game Video Generation with Hybrid History Condition
Hunyuan-GameCraft can generate interactive game videos by combining keyboard and mouse inputs into a shared camera view.

Hunyuan-GameCraft example
SwapAnyHead: Controllable and Expressive One-Shot Video Head Swapping
SwapAnyHead can swap a human head from a still image into a video while keeping the original body and background.

SwapAnyHead example
SyncTalk++: High-Fidelity and Efficient Synchronized Talking Heads Synthesis Using Gaussian Splatting
SyncTalk++ can generate high-quality talking head videos with synchronized lip movements and facial expressions. It uses Gaussian Splatting for consistent subject identity and can render up to 101 frames per second.

SyncTalk++ example
Audio
Sonic4D: Spatial Audio Generation for Immersive 4D Scene Exploration
Sonic4D can generate spatial audio for 4D scenes by tracking sound sources from monocular video.

Sonic4D example

Enjoy the weekend my fellow dreamers! Infrachrome
style is now on Promptcache.
And that my fellow dreamers, concludes yet another AI Art weekly issue. If you like what I do, you can support me by:
- Sharing it 🙏❤️
- Following me on Twitter: @dreamingtulpa
- Buying me a coffee (I could seriously use it, putting these issues together takes me 8-12 hours every Friday 😅)
- Buying my Midjourney prompt collection on PROMPTCACHE 🚀
- Buying access to AI Art Weekly Premium 👑
Reply to this email if you have any feedback or ideas for this newsletter.
Thanks for reading and talk to you next week!
– dreamingtulpa