Mathias Parger

[Linqi (Alex) Zhou @linqi_zhou]

Thanks to my incredible collaborators @MParger @ayaanzhaque @baaadas for making it happen.

Blog: https://t.co/M8N66waQFN
ArXiv: https://t.co/LACoUwpj8Y
Code: https://t.co/CawGtlcig3

[Jiaming Song @baaadas]

Our new method, Terminal Velocity Matching (TVM), works on large-scale models while enjoying all the advantages of IMM — training stability, high-quality samples, efficient inference. This moves us closer to a genuine evolutionary step beyond today’s diffusion models.

[Luma @LumaLabsAI]

Introducing Ray2, a new frontier in video generative models. Scaled to 10x compute, #Ray2 creates realistic videos with natural and coherent motion, unlocking new freedoms of creative expression and visual storytelling. Available now. Learn more https://t.co/jGI6KmRQpR.

[Luma @LumaLabsAI]

Introducing the all-new Luma Photon text-to-image models, now available in the Luma API. Photon and Photon Flash are the most creative, personalizable, and intelligent visual models for creatives, bringing a step-function change in the cost of high-quality image generation.

[Luma @LumaLabsAI]

Say hello to the all-new #DreamMachine. 🚀 Your visual thought partner—where ideas become reality. Ideate, visualize, and share your ideas with the world. Available now.

[Luma @LumaLabsAI]

🚀 Welcome to the era of Hyperfast video generation: with 10x faster inference, you can now generate full-quality Dream Machine v1.6 clips in under 20 seconds. No "turbo" or "distilled" models - just uncompromised quality. Available today to all subscribers and API customers.

[Luma @LumaLabsAI]

🚀 Introducing the Dream Machine API. Developers can now build and scale creative products with the world's most popular and intuitive video generation model without building complex tools in their apps. Start today https://t.co/rtDKtZ5kTW #LumaDreamMachine

[Mathias Parger @MParger]

Are you annoyed by popping artifacts with 3DGS?
Check out our new Siggraph paper StopThePop. StopThePop adds per-pixel sorting with minimal performance impact.

[Lukas Sebastian Radl @rafourdl]

🚀 Excited to announce our latest research: “StopThePop: Sorted Gaussian Splatting for View-Consistent Real-Time Rendering”, which will appear in #SIGGRAPH2024! 🎉
Finally, fast and view-consistent rendering of 3D Gaussians, without popping artifacts! https://t.co/bsOHSn0peX

[Shimon Vainer @esx2ve]

We’re excited to introduce our group's new research paper, “Collaborative Control for Normal-Conditioned PBR Image Generation”, in which we tackle high quality single-view PBR materials! 🧵

arxiv: https://t.co/f2wp9p6ywz
project page (with live demo!): https://t.co/5x4jZiAByb

[Thomas Neff @thomasneff93]

Excited to announce our #ECCV2022 paper "AdaNeRF"! Together with @LvZhaoyang and @MZollhoefer , we show that our adaptive, fully neural representation can compete with (and even outperform) modern hybrid approaches in terms of quality, compactness and rendering speed. 1/4

[Mathias Parger @MParger]

@DaoudiMed64 @CVPR @DaoudiMed64 same problem here. I think it will only get worse, because soon the first session starts and even more people will have that google docs open at the same time to find the posters they are looking for.

[Mathias Parger @MParger]

Happy to introduce DeltaCNN (#CVPR2022), a CNN framework for sparse video inference. By reusing previous results for unchanged pixels, we accelerate CNN inference up to 7x compared to cuDNN.
Project page: https://t.co/g31BO1h0U1

[Mathias Parger @MParger]

@Lane_Kerr Thanks. But unfortunately we only had few active players in the last years and the maintenance cost too much time and money to keep Splitter alive.