- 15
- 42 560
Berkeley Synthetic
เข้าร่วมเมื่อ 24 มิ.ย. 2022
Berkeley Synthetic is an AI research group working with generative AI, simulations and 3D asset portability.
Texture: Text to 3D textures
🚀TEXTure: Text-Guided 3D Texturing
👉 Anonymous Authors unveil a new method for creating new textures using an iterative diffusion-based process
🔹Generate HQ Textures using prompts
🔹Leveraging pre trained depth-to-image diffusion
🔹Novel elaborated diffusion sampling process
🔹Edit Existing textures using prompts
Project Source Code:
→Project Page: lnkd.in/g4PfbAnM
→Paper: lnkd.in/gqbFd2qY
→Code: lnkd.in/gtngxXFr
→Hugging Face App: lnkd.in/gusabEuv
👉 React and follow for more unique posts on generative AI applications
Follow Berkeley Synthetic Page for more unique posts on generative AI applications: lnkd.in/gjKJPbxA
#generativeai #berkeleysynthetic #3d #3dmodeling #texture #py
👉 Anonymous Authors unveil a new method for creating new textures using an iterative diffusion-based process
🔹Generate HQ Textures using prompts
🔹Leveraging pre trained depth-to-image diffusion
🔹Novel elaborated diffusion sampling process
🔹Edit Existing textures using prompts
Project Source Code:
→Project Page: lnkd.in/g4PfbAnM
→Paper: lnkd.in/gqbFd2qY
→Code: lnkd.in/gtngxXFr
→Hugging Face App: lnkd.in/gusabEuv
👉 React and follow for more unique posts on generative AI applications
Follow Berkeley Synthetic Page for more unique posts on generative AI applications: lnkd.in/gjKJPbxA
#generativeai #berkeleysynthetic #3d #3dmodeling #texture #py
มุมมอง: 866
วีดีโอ
Who wants pineapple?
มุมมอง 97ปีที่แล้ว
Outputs of Stable DreamFusion a PyTorch implementation of DreamFusion using Stable Diffusion. Follow Berkeley Synthetic Page for more unique posts on generative AI applications: lnkd.in/gjKJPbxA
Snoop Dogg AI Animation| Giving a speech about Zombies
มุมมอง 649ปีที่แล้ว
Our team at Berkeley Synthetic decided to have some fun with Generative AI tools, Here is a video of Snoop Dogg talking about a Zombie Epidemic Caused by fungi and Global warming. Tools used: → Midjourney: lnkd.in/gPSN8HMU → ChatGPT: lnkd.in/g8yNrcqw → Speechify: speechify.com → MakeitTalk: lnkd.in/gGHCuhfv 👉 React and follow for more unique posts on generative AI applications Follow Berkeley S...
🚀Rodin: 3D Avatars Using Diffusion
มุมมอง 1.4Kปีที่แล้ว
You can learn more about it from here: www.linkedin.com/feed/update/urn:li:activity:7024622043669176320 This content is uploaded only for educational purpose Follow Berkeley Synthetic Page for more unique posts on generative AI applications: lnkd.in/gjKJPbxA
🚀Online Real-Time Volumetric Nerf + Slam
มุมมอง 1.1Kปีที่แล้ว
🚀Online Real-Time Volumetric Nerf Slam 👉MIT unveils a novel geometric/photometric 3D mapping pipeline powered by NeRF SLAM. 🔹First dense monocular SLAM hierarchical V-NeRF 🔹Photometric accuracy / great geometric accuracy 🔹NO poses or depths as input, runs in real-time 🔹New SOTA performance on the Replica dataset Project Source Code: → Nerf Slam Paper: lnkd.in/g6AhjuxV → Nerf Slam Code:lnkd.in/g...
Ryan Reyonlds Toonified using VToonify
มุมมอง 717ปีที่แล้ว
🚀VToonify: Controllable High-Resolution Portrait Video Style Transfer 👉VToonify leverages the mid- and high-resolution layers of StyleGAN to render high-quality artistic portraits based on the multi-scale content features extracted by an encoder to better preserve the frame details. 🔹Supporting unaligned faces & various video sizes 🔹Flexible Style Control 🔹Optical-flow-free flicker loss for con...
🚀Tune-A-Video : Text 2 Video Algorithm
มุมมอง 845ปีที่แล้ว
🚀Tune-A-Video: One-Shot Tuning of Image Diffusion Models for Text-to-Video Generation 🔹One-Shot Video Generation 🔹generates videos from text prompts via an efficient one-shot tuning of pretrained T2I diffusion models. 🔹First ever video via one-shot pre-trained T2I DMs 🔹capable of producing temporally-coherent videos over various applications such as change of subject or background, attribute ed...
Muse: Text2Image Demo
มุมมอง 385ปีที่แล้ว
🚀Muse:Text 2 Image Generation via Masked Generative Transformers 👉Google unveils the new SOTA (faster and better) for text-to-image generation. 👉Muse is a fast, state-of-the-art text-to-image generation and editing model 🔹Text-to-image Generation 🔹Zero-shot, Mask-free editing 🔹 Almost Twice as Faster than Stable Diffusion 1.4 🔹Zero-shot Inpainting/Outpainting Project Source Code: → Project page...
Creating 3D Models using Get3D from NVIDIA
มุมมอง 5Kปีที่แล้ว
GET3D: A Generative Model of High Quality 3D Textured Shapes Learned from Images #Nvidia open-sourced GET3D: textured 3D meshes with complex topology, rich geometry, and HQ textures Here is an example of 3D models generated using shapenet car model from Get3D Project Source Code: Project Page: lnkd.in/eXskY62B Get 3D Code: lnkd.in/gpXuDRdc Paper: lnkd.in/gpxpweWr React and follow for more uniqu...
Text to 3D Model using SJC
มุมมอง 880ปีที่แล้ว
🚀 (Score Jacobian Chaining): A Text to 3D model using Pretrained Stable Diffusion Model SJC is a Text to a 3D algorithm which uses 2D stable diffusion model for generating 3D assets, it can help tremendously for creating new 3D models and assets for various different industries Prompt: “ A high quality picture of cheese burger,8k,insane details ” Training Time with RTX 3080: 3H 14M Project Sour...
City 3D model using Instant Nerf
มุมมอง 1.5Kปีที่แล้ว
🚀 NeRF with Google Earth Engine 👉🏼 Here is an example of how instantNeRF can be used with Google Earth Engine data to construct a 3D cityscape which can be used as an asset in simulations, games and film. 🔹NeRF produces 3D multiview synthesis from 2D images 🔹Training time on 1 RTX 3090: 1 Hour 12 M Project Source Code: InstantNeRF: lnkd.in/e7KxvndG Instant Nerf Code: lnkd.in/evB82Wgy Paper: lnk...
Metaverse For All - Building an Open and Inclusive Metaverse
มุมมอง 28Kปีที่แล้ว
Berkeley Synthetic's CEO Matt White presents on building an open and inclusive Metaverse during the ARVR Association's Metaverse 2.0 event on March 10th, 2022. #Metaverse #Ethics #inclusion #ARVR #SyntheticReality
That’s quite underwhelming 😢
Now we know the process of "Creating 3D Models using Get3D from NVIDIA" ...
Hello, I have a question about how you extracted the 2D images from Google Earth.
Hi! I would like to know if you are running it on Windows or Linux, im trying to get the installation for Linux but Im breaking my head against the wall. I just can get it right. Thanks!
How ?
Would be impressive if they didn't use the exact same frame for all three of the fish. If the product was something that actually generated a 3D structure that aligned with the image, it'd be cool. But this seems to be a way to import Dall-E or some other dime-a-dozen image generator onto premade models. Hard pass. :(
Why anonymous, is it real ?
i'd assume for blind submission to conferences or journals
it is real, we have listed down the links in the description, do check them out
You can also find the paper on arxiv with the same title which will tell you what the authors are ;)
Bummer story
Zombie Generator.
I don't know if it has that level of flexiblity but it could always be a possibly to build custom models similar to that of stable diffusion Generative AI is gonna create many more amazing tools
Those results look amazing
It also has features for Image to model generation, it could be really cool to implement once this technology is open-sourced
@@berkeleysynthetic8542 Nice. But no working online demo yet?
@@apatsa_basiteni not yet
YEAH!
That look interesting, is it working on windows with anaconda ?
It is working in Windows, do try it out and let me know your thoughts
Own nothing, be happy….AI communism.
Imaginations, I keep hearing that….vain in their own imaginations. So a great big digital prison. What if you miss behave, use wrong think, will you been thrown in some dark part of the meta web? Maybe a zombie game, a demon realm or some other horror? A counterfeit world where you can be anything you want, have super powers……what if you run out of money…….where will you go? If you think this is a good thing, your out of your mind.