Anton Obukhov @AntonObukhov1 Twitter profile

Pinned Tweet

Anton Obukhov

1 month

Introducing Marigold-LCM 🌼 - a FAST version of the now-popular state-of-the-art depth estimator! Check out the new 🤗 demo (now accepts videos): New on the team: Tianfu Wang ( @TianfuWang2 ) Original Marigold:

Anton Obukhov

@AntonObukhov1

5 months

Introducing Marigold 🌼 - a universal monocular depth estimator, delivering incredibly sharp predictions in the wild! Based on Stable Diffusion, it is trained with synthetic depth data only and excels in zero-shot adaptation to real-world imagery. Check it out: 🌐 Website:…

47

278

1K

12

103

588

Last Seen Profiles

@antoine_anfre

@rarcilain

@jollybeggars

@SenryuReplector

@broy38

@EuCitSci

@cocoro_0808_

@softbrowarnik

@AnarchyHabitat

@MilkyNrco

@key878_tomas

@sugoiibiarmy

@Cinzia_Bianco

@kennyvaccaro

@YangsiboHuang

@shekuwow

@openhealthhub

@Thintallmarcus

@Kijima_Lear

@ethereumJoseph

@MaxRovensky

@reiq

@jin_he_hun

@Markie_Gee

@dev_bcn

@Kosha_tmn

@PalauanR

@reiq

@IMQueenzy_

@little_jiaxinn_

@ShacharLab

@Shiv_NMapp

@ItsBarlo

@GroovyOwls

@MagnetSidi

@lucio_xix

Anton Obukhov

@AntonObukhov1

5 months

Introducing Marigold 🌼 - a universal monocular depth estimator, delivering incredibly sharp predictions in the wild! Based on Stable Diffusion, it is trained with synthetic depth data only and excels in zero-shot adaptation to real-world imagery. Check it out: 🌐 Website:…

47

278

1K

Anton Obukhov

@AntonObukhov1

4 months

Point2CAD: A leap in mechanical CAD model reconstruction from point clouds, blending semantic segmentation with a novel neuro-analytic surface representation. Discover more: 🔗 Website: ⚙️ Paper: 🛠️ Code:

8

82

433

Anton Obukhov

@AntonObukhov1

5 months

Marigold Depth got an update! 🖨️ Introducing a NEW 3D-printing demo feature! 🚀 Demo now 2x faster, with fine controls 🔓 Switched to Apache-2 license 💻 Code enhancements: fp16 & batched inference === 🤗 Demo: 🌐 Website:

6

76

397

Anton Obukhov

@AntonObukhov1

4 months

Want to learn how to lift Midjourney v6 (and before) into 3D, and 3D-print it? Follow the link to our free online tool: , and follow me to stay updated!

3

31

214

Anton Obukhov

@AntonObukhov1

5 months

Since Marigold depth launch yesterday, we've seen amazing uses: 3D printing, holographic images, UE5 assets, ControlNet, and optical illusion disentanglement. Share your creations with us! Use our free tool and follow @KBingxin for updates!

Marigold Depth Estimation - a Hugging Face Space by prs-eth

huggingface.co

6

30

140

Anton Obukhov

@AntonObukhov1

2 months

#CVPR2024 results are in -- time for the humblebrag fest! Happy to share that Marigold Depth and Point2CAD are heading to Seattle this year! Congrats to the amazing PRS-ETH team! Marigold: Point2CAD:

14

8

147

Anton Obukhov

@AntonObukhov1

4 months

High-quality monocular depth estimation enables conversion of regular videos into VR-ready 3D video formats! #MarigoldDepth

Ali

@lirezaBahremand

4 months

@pvncher Hey Eric, I took this idea and ran with it, learning a lot along the way. I made a simple Jupyter Notebook that people can run to generate and upload a sbs video to Quest. Dev can use whatever depth model (I tested with PatchFusion and Marigold).

3

6

45

1

13

123

Anton Obukhov

@AntonObukhov1

2 years

Excited to share that my semester student @hexagon_prime did unimaginable by shaping his thesis (4 months of part-time) into a #CVPR2022 submission, which got accepted and already gained a lot of attention in the community. FYI: he is on the lookout for PhD positions!

AK

@_akhaliq

2 years

Pix2NeRF: Unsupervised Conditional π-GAN for Single Image to Neural Radiance Fields Translation abs:

1

50

254

1

6

75

Anton Obukhov

@AntonObukhov1

3 years

I will present "torch-fidelity: High-fidelity performance metrics for generative models in PyTorch" at PyTorch Ecosystem Day #PTED21 on April 21, 2021. Register at Project and code:

0

7

61

Anton Obukhov

@AntonObukhov1

2 years

Excited to share that I successfully passed my Ph.D. defense on the topic "Tensor Decompositions in Deep Learning." Huge thanks to @lucvangool , @CVL_ETH , and everyone involved for the amazing journey!

4

0

49

Anton Obukhov

@AntonObukhov1

7 months

I met the legend @_akhaliq at @ICCVConference (both slightly retouched)

0

3

47

Anton Obukhov

@AntonObukhov1

2 years

This year, any #neurips2022 reviewer who did not vanish into thin air during the discussion period deserves an outstanding reviewer award.

0

3

45

Anton Obukhov

@AntonObukhov1

5 months

@kwea123 We focus on the quality first - speed will come later through research and improvements in diffusion samplers. Arguably, Stable Diffusion is also slower than most of the GANs :)

1

0

45

Anton Obukhov

@AntonObukhov1

5 months

Marigold 3D printing demo will convert your image into a watertight 3D asset in STL format, readily compatible with any 3d printer software or online printing services. It's a great opportunity to make customized gifts for your loved ones! 🎁

Anton Obukhov

@AntonObukhov1

5 months

Marigold Depth got an update! 🖨️ Introducing a NEW 3D-printing demo feature! 🚀 Demo now 2x faster, with fine controls 🔓 Switched to Apache-2 license 💻 Code enhancements: fp16 & batched inference === 🤗 Demo: 🌐 Website:

6

76

397

3

5

41

Anton Obukhov

@AntonObukhov1

7 months

🎨 Excited to share our latest paper: "Breathing New Life into 3D Assets with Generative Repainting"! We present a novel approach to repaint 3D models using text prompts, pretrained Stable Diffusion, and NeRF. A step towards assisting content creators! 🖌️ 📄 Paper:…

1

8

38

Anton Obukhov

@AntonObukhov1

3 years

1/ Our paper “Spectral Tensor Train Parameterization of Deep Learning Layers” about end-to-end neural network compression and stability of training in the GAN setting is live at #AISTATS2021 this week! PDF: Project page:

1

10

28

Anton Obukhov

@AntonObukhov1

3 years

1/1 accepted at #CVPR2021 , more details after (the overall) acceptance happiness cools down a notch :)

0

25

Anton Obukhov

@AntonObukhov1

7 months

Matthew Tancik talks about all the cool stuff from @nerfstudioteam and @LumaLabsAI at the @ICCVConference AI for 3D Content Creation workshop

0

2

24

Anton Obukhov

@AntonObukhov1

3 months

This is just crazy. Given the pace of generative imaging, open models with such quality should become available in 2024

AK

@_akhaliq

3 months

Open AI introducing Sora text-to-video model Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions.

37

255

1K

1

21

Anton Obukhov

@AntonObukhov1

5 months

Marigold depth estimator works great with images created by Adobe Firefly and other generative AI tools! #MarigoldDepth #AdobeFirefly

Kris Kashtanova

@icreatelife

5 months

TIP OF THE DAY Add “cinematic shot” to your prompt prompt: Close up of a fire breathing flying dragon, cinematic shot Tested with Adobe Firefly

23

10

138

3

1

23

Anton Obukhov

@AntonObukhov1

4 years

Our paper “T-Basis: a compact representation for neural networks” is live at ICML! We learn compressed neural networks by reparameterizing weights twice: through tensor ring (TR) decomposition, and each TR core through a shared basis. Project page:

1

6

21

Anton Obukhov

@AntonObukhov1

4 months

3D printing with Marigold depth taken to the next level 🤯

Jukka Seppänen

@Kijaidesign

4 months

Example of what AI generation can do in practice currently, from initial image (which is also generated) to a very detailed printable model, the level of detail mostly made possible with the Marigold depth estimation. #ai #3dprinting #dragon #technology

6

36

116

0

1

21

Anton Obukhov

@AntonObukhov1

3 months

@creatine_cycle You forgot __pycache__

1

0

19

Anton Obukhov

@AntonObukhov1

3 years

Searching (and finding) @ICCV_2021 paper id in the accepted list is a form of contrastive learning: at first annoying, then immediately joyful

1

0

19

Anton Obukhov

@AntonObukhov1

3 years

Our #ICCV2021 paper on Adaptive Task-Relational Context, leveraging neural architecture search and attention mechanisms in the multi-task learning setting, is now paired with source code! @menelaoskanakis @stam_g paper: code:

1

5

19

Anton Obukhov

@AntonObukhov1

3 years

@CSProfKGD Thanks for pointing out! As a reviewer, I noted that CMT did not send me a notification about #CVPR2021 rebuttals reveal. As an author, it makes me worry the reviewers may not remember to check my rebuttal.

1

0

19

Anton Obukhov

@AntonObukhov1

2 years

Starting in November, I will be a postdoc in Prof. Konrad Schindler's Photogrammetry and Remote Sensing lab at ETHZ. Looking forward to more 3D work!

0

17

Anton Obukhov

@AntonObukhov1

1 year

95% of @CVPRConf proceedings text will be written/edited in the remaining 12 hours

2

17

Anton Obukhov

@AntonObukhov1

2 years

#ICVSS2022 is over; recap time. Where are we now, @GMFarinella ? ;) We thank you, @robertocipolla , @sebattiato , for delivering a unique experience for over a decade! It was a pleasure to meet speakers and connect with peers. Kudos to @vincesitzmann for coffee break discussions!

1

0

16

Anton Obukhov

@AntonObukhov1

4 years

Check out my new Python package for evaluation of GANs in @PyTorch with - simplicity (import or CLI command) - extensibility (swap Inception with your own feature extractor) - fidelity (epsilon relative errors compared with original codes) in mind.

0

7

16

Anton Obukhov

@AntonObukhov1

2 years

My library of generative models performance evaluation, torch-fidelity, which I started as a pet project, reached 1M downloads today!

GitHub - toshas/torch-fidelity: High-fidelity performance metrics for generative models in PyTorch

High-fidelity performance metrics for generative models in PyTorch - toshas/torch-fidelity

github.com

1

0

13

Anton Obukhov

@AntonObukhov1

2 months

@CVPR Hit that Enter!

0

13

Anton Obukhov

@AntonObukhov1

3 years

Glad to see that tensor methods are re-emerging in model compression. Even more glad, knowing that our paper "Spectral Tensor Train Parameterization of Deep Learning Layers" was recently accepted at AISTATS'2021. Links to follow

AI at Meta

@AIatMeta

3 years

Introducing TT-Rec, a new method to dramatically shrink memory-intensive Deep Learning Recommendation Models and make them easier to deploy at scale.

4

139

600

0

2

13

Anton Obukhov

@AntonObukhov1

7 months

@huggingface reception at @ICCVConference

0

1

11

Anton Obukhov

@AntonObukhov1

5 months

Marigold depth can boost performance of many downstream applications. Mesmerizing results from UE5 here 😍

麻場健司／MocapFor(Streamer|All)開発

@KenjiASABA

5 months

Marigold🌼で、手持ちの写真の奥行を推定してUE5のNanite Tessellationで描画してみた。これまで試した単眼深度推定モデルの中でも群を抜いてきれいな気がする。近景はほとんど破綻がない一方で、遠景メインの風景はあまり学習対象になっていないのかも。

0

57

265

0

3

12

Anton Obukhov

@AntonObukhov1

5 months

More 3D assets from your generative pipeline with Marigold Depth! Thread 🧵⬇️

Daniel Skaale

@DSkaale

5 months

Wow I'am really impressed by Marigold Depth Estimation just a quick extraction in Zbrush 4k texture #madewithunity #zbrush #huggingface #AIimage #midjouney #HDRP

0

1

11

2

3

12

Anton Obukhov

@AntonObukhov1

4 months

Nice qualitative comparison of Marigold depth with other SOTA methods!

patricio gonzalez vivo

@patriciogv

4 months

Testing SOTA for Depth Estimation. Marigold vs PathFusion vs DepthZoe

8

14

139

0

1

12

Anton Obukhov

@AntonObukhov1

3 years

Exceeding expectations as an #ICCV reviewer since 2021!

1

0

12

Anton Obukhov

@AntonObukhov1

3 years

Code packaging frenzy continues! Check out my latest python package, democratizing orthogonal transformations in #PyTorch . Goomba has nothing to do with the Householder transformation; your attention is all I need.

GitHub - toshas/torch-householder: Efficient Householder Transformation in PyTorch

Efficient Householder Transformation in PyTorch. Contribute to toshas/torch-householder development by creating an account on GitHub.

github.com

0

5

11

Anton Obukhov

@AntonObukhov1

10 months

As #ICCV2023 accepted papers promotions slow down, here's a shameless plug: I'm delighted to share that 2/3 of our submissions made the cut. The 1st, "DiffDreamer," enables dream-like generative experiences of nature's trailing views. Follow @prime_cai for updates!

Shengqu Cai

@prime_cai

10 months

Freshly accepted to #ICCV2023 ! Trained on **images only**, we can freely extrapolate into or out of a scene. Joint effort with an amazing team @ericryanchan @songyoupeng @MShahbazi_ @AntonObukhov1 Luc Van Gool @GordonWetzstein

0

30

169

1

0

12

Anton Obukhov

@AntonObukhov1

5 months

Marigold seeding continues!

fofr

@fofrAI

5 months

The excellent Marigold depth model is now on Replicate: 🙏 @alaradirik for adding

1

4

32

0

1

11

Anton Obukhov

@AntonObukhov1

3 years

Ever wondered if there is anything more efficient for #RL returns than a #PyTorch loop, #TF .scan, or #SciPy .lfilter? Check out my algorithm for parallel discounted cumulative sum computation! `pip install torch-discounted-cumsum`

torch-discounted-cumsum

Fast Discounted Cumulative Sums in PyTorch

www.obukhov.ai

0

5

10

Anton Obukhov

@AntonObukhov1

7 months

@RanaHanocka talks about learning in 3D by lifting 2D generative priors at @ICCVConference AI for 3D Content Creation workshop

0

1

11

Anton Obukhov

@AntonObukhov1

2 years

Anyone else checking the #NeurIPS2022 author console in hopes of ripping off the band-aid 1-2 hours earlier?

0

10

Anton Obukhov

@AntonObukhov1

4 years

First time that I got a NeurIPS reviewer saying I must have referenced a mildly unrelated ICML2020 paper. I thought these stories are hoax!

0

10

Anton Obukhov

@AntonObukhov1

7 months

1st day of @ICCVConference poster session. With 6000 posters over 3 days, my routine is to run all the rows and later revisit some. It worked great during @CVPRConf , but the narrow spacing between the rows makes it impossible this time. 🤞 poster spacing standardized

0

1

10

Anton Obukhov

@AntonObukhov1

4 months

I can't get enough of seeing how Marigold Depth fits like a puzzle piece into so many creative workflows and ultimately makes people happy!

Barber Shop Globes

@MinyardCathy

4 months

@AntonObukhov1 Thank you for creating this app. I can upload AI images into Marigold. Create STL file. Download to my computer. Upload STL into my cnc router software and within 5 minutes of adjusting image, I have a 3D bas relief suitable for wood carving. I'm overwhelmed with joy!

1

3

0

1

10

Anton Obukhov

@AntonObukhov1

2 years

@RealAAAI what is the threshold score for NeurIPS fast track this year, given the changed scoring system? Previously 4.9 corresponded to the first grade of reject. The same threshold this year translates to the first grade of accept. #aaai2023 #NeurIPS2022

0

2

8

Anton Obukhov

@AntonObukhov1

5 months

Tianfu presenting his work at @BMVCconf

0

2

9

Anton Obukhov

@AntonObukhov1

5 months

miyamotty

@miyamotty

5 months

なるほど

0

5

1

9

Anton Obukhov

@AntonObukhov1

2 years

Mixed (1/2 got in) feelings about #NeurIPS2022

0

8

Anton Obukhov

@AntonObukhov1

5 months

Reading updates about OpenAI be like

0

8

Anton Obukhov

@AntonObukhov1

3 years

I will present our latest #AISTATS work about sparse and stable CNN parameterizations (pinned) in the sparsity workshop today in 4 hrs #SNN21

0

3

8

Anton Obukhov

@AntonObukhov1

11 months

I'm at #CVPR this week! DM or reply if you want to chat and/or reconnect!

0

7

Anton Obukhov

@AntonObukhov1

3 years

@junyanz89 @sayantandas_ @GauravTParmar @rzhang88 Great work! Some of the interpolation issues in the context of FID calculation are discussed and implemented here:

GitHub - toshas/torch-fidelity: High-fidelity performance metrics for generative models in PyTorch

High-fidelity performance metrics for generative models in PyTorch - toshas/torch-fidelity

github.com

1

0

7

Anton Obukhov

@AntonObukhov1

1 year

Our #NeurIPS2022 paper on reducing the computational complexity of controlling the _real_ singular values of convolutional layers is online: None of us could go present in person, but feel free to contact Alexandra (email in pdf) or me to discuss! 🧵

1

0

7

Anton Obukhov

@AntonObukhov1

10 months

2M downloads just half a year later! Thanks to the huge interest in generative models and a large community of researchers. An upcoming version 0.4.0 is planned to include more recent metrics, including FID_CLIP and more. Leave a feature request in this thread! #pytorch #metrics

0

7

Anton Obukhov

@AntonObukhov1

4 months

A new release of @huggingface @diffuserslib includes our Marigold depth estimation pipeline!

Sayak Paul

@RisingSayak

4 months

How can we end the year without a new release of 🧨 diffusers to celebrate it? 1️⃣ A brand new model, aMUSEd (more soon from @psuraj28 and @williamlberman ), which is NOT based on diffusion 😅 2️⃣ 3x faster SDXL (extends to other pipelines) 🏎️ 3️⃣ Transitioning to PEFT for LoRA…

3

14

81

0

7

Anton Obukhov

@AntonObukhov1

3 years

That escalated quickly - just received a notification of being an outstanding (top 5%) reviewer! Thanks for the free registration #ICCV2021

0

7

Anton Obukhov

@AntonObukhov1

7 months

Fast forward 5 years. AGI is around us, making lives better, connecting people. Conference badges still flip, printed one-sided 🥲

0

7

Anton Obukhov

@AntonObukhov1

2 years

Another highlighted #ICLR2022 reviewer here! #ShamelessSelfPromo Also congrats @Pseudomanifold , @bayesgroup , @songyoupeng !

0

7

Anton Obukhov

@AntonObukhov1

4 months

Ever wanted to explore image scenes from a new perspective? Our cutting-edge depth estimator breaks through the fourth wall, bringing every pixel closer to you! DIY:

0

1

7

Anton Obukhov

@AntonObukhov1

2 years

First outdoor experience with @LumaLabsAI

Turtle - Created by @antonobukhov with Luma

October 24, 2022

lumalabs.ai

0

1

5

Anton Obukhov

@AntonObukhov1

5 months

A.I.Warper

@AIWarper

5 months

Image to depth using Marigold and ConfyUI Then depth to 3D geometry in Blender Super interesting stuff 🤯

31

223

2K

1

0

6

Anton Obukhov

@AntonObukhov1

3 months

@ylecun These generative world models are nevertheless very useful for distilling into sota perception models. We showed it on the task of monocular depth estimation in our Marigold paper:

Marigold Depth

Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation

marigoldmonodepth.github.io

0

6

Anton Obukhov

@AntonObukhov1

4 months

Point2CAD operates seamlessly on top of any point cloud segmentation or clustering method, with results directly fed into analytic models. Our approach utilizes a novel implicit neural representation for freeform surface fitting. The team: Yujia Liu ( @YujiaLiu14 ), Anton Obukhov…

1

2

5

Anton Obukhov

@AntonObukhov1

5 months

@BoyuanChen0 Relative

0

7

Anton Obukhov

@AntonObukhov1

3 months

Will Smith enjoying spaghetti might just be the ultimate benchmark for generative video quality. We need a fair comparison!

Sam Sheffer

@samsheffer

3 months

🚨text to video progress in one year🤯 2023: 2024:

1K

7K

66K

0

6

Anton Obukhov

@AntonObukhov1

5 months

Great new work from @lukashoyer3 ; Follow him for news in semantic perception, autonomous driving, and beyond!

Lukas Hoyer

@lukashoyer3

5 months

Introducing SemiVL - a semi-supervised semantic segmentation model exploiting priors from vision-language models. Paper: Code: Team: @lukashoyer3 , David J. Tan, @ferjadnaeem , Luc Van Gool, @fedassa #ETHZurich #Google internship

0

11

77

1

0

6

Anton Obukhov

@AntonObukhov1

3 years

My colleague @CSakaridis , who works on making self-driving cars safer to use in harsh weather conditions, has just released a #cityscapes -inspired benchmark, a paper, and joined Twitter. Welcome Christos!

Christos Sakaridis

@CSakaridis

3 years

ACDC released: a new large-scale driving dataset for training and testing semantic segmentation algorithms on adverse visual conditions, such as fog, nighttime, rain, and snow. The dataset and benchmarks are available at #computervision #autonomouscars

0

7

27

1

6

Anton Obukhov

@AntonObukhov1

7 months

@_akhaliq @huggingface Thank you for listing our paper, means a lot!

0

6

Anton Obukhov

@AntonObukhov1

3 months

I have been learning German for more than a year now, and recently, I realized that I need a principled way to learn new words from the textbook as I progress through it. Unfortunately, neither Anki nor Quizlet had the sets I needed, so I decided to make them myself with AI 🧵⬇️

1

0

6

Anton Obukhov

@AntonObukhov1

5 months

Impressively detailed scenes extruded and rendered in Unreal Engine 5:

麻場健司／MocapFor(Streamer|All)開発

@KenjiASABA

5 months

Marigold🌼で、手持ちの写真の奥行を推定してUE5のNanite Tessellationで描画してみた。これまで試した単眼深度推定モデルの中でも群を抜いてきれいな気がする。近景はほとんど破綻がない一方で、遠景メインの風景はあまり学習対象になっていないのかも。

0

57

265

1

6

Anton Obukhov

@AntonObukhov1

5 months

@mattwallace @yacineMTB @Tesla Marigold is tailored for single-image scene inference. In contrast, cars handle video streams, analyzing temporal data. They use multiple cameras and advanced multiview geometry algorithms for precise distance estimation.

1

0

6

Anton Obukhov

@AntonObukhov1

1 year

I will be presenting our latest work on tensorial neural fields compression at the 3rd workshop on neural low‑dimensionality #slowdnn organized by @mbzuai

0

6

Anton Obukhov

@AntonObukhov1

3 years

Here is a game to spice up the coming review periods: Reviewer Bingo! Tick out a row/col/diagonal, mutter "BS!", and proceed with the review: "The paper is well-written and easy to follow..."

0

6

Anton Obukhov

@AntonObukhov1

2 years

Hey @overleaf and @grammarly , we need you integrate with each other, badly. We write TeX, we appreciate great spell-checking, and we wish to unlearn the tag-stripping skill!

2

0

5

Anton Obukhov

@AntonObukhov1

7 months

@Jimantha talks about learning motion and 3D in the @ICCVConference workshop on Frontiers of Monocular 3D Perception

0

5

Anton Obukhov

@AntonObukhov1

5 months

This is so exciting, seeing what can be done with Marigold! Thanks for sharing!

1

0

5

Anton Obukhov

@AntonObukhov1

1 year

@heykahn We already saw how it panned out for self-driving cars: who is going to be held accountable? Before this question is answered, jobs aren't going anywhere

0

4

Anton Obukhov

@AntonObukhov1

7 months

First day of @ICCVConference - workshop and tutorials. I will be spending the whole day in AI for 3D content creation workshop, come by to say hi!

0

5

Anton Obukhov

@AntonObukhov1

3 years

torch-fidelity got some recognition!

Lightning AI ⚡️

@LightningAI

3 years

️🎉⚡️ Now available, #TorchMetrics v0.4! We've made improvements to the library and added new Audio and Image domains with new metrics! Check out Nicki Skafte's blogpost to learn more:

2

6

24

0

5

Anton Obukhov

@AntonObukhov1

2 years

Happy that torch-fidelity was found useful in #stablediffusion code!

0

5

Anton Obukhov

@AntonObukhov1

5 months

@taiyasaki @jon_barron If overleaf had a global search and replace like any normal IDE, the monofile idea won't have legs

1

0

4

Anton Obukhov

@AntonObukhov1

2 years

Something very familiar to all NeRF enthusiasts, made with @LumaLabsAI

Lego (real) - Created by @antonobukhov with Luma

October 28, 2022

lumalabs.ai

2

0

5

Anton Obukhov

@AntonObukhov1

11 months

Heading to CVPR and thinking what to download into Netflix? New season of Black Mirror is out, with the first episode about generative AI

1

0

4

Anton Obukhov

@AntonObukhov1

2 years

@mentzer_f While leaky ReLU is bad enough, check out the default batchnorm momentums! ;)

0

4

Anton Obukhov

@AntonObukhov1

3 years

Achievement "you again" unlocked: get invited to review a journal submission of a paper you just recently reviewed for a conference

0

4

Anton Obukhov

@AntonObukhov1

1 year

@francoisfleuret @ducha_aiki @ChrisJReiser @szeliski @dorverbin @_pratul_ @BenMildenhall @jon_barron @PeterHedman3 Many low-rank weight or activation decomposition methods work well, but ultimately fail to convert FLOPs improvement promises into wallclock speedups. Beating cudnn is really hard. Things might be changing to the better with operator fusion..

0

4

Anton Obukhov

@AntonObukhov1

5 months

The first seed of Marigold:

Anton Obukhov

@AntonObukhov1

5 months

Introducing Marigold 🌼 - a universal monocular depth estimator, delivering incredibly sharp predictions in the wild! Based on Stable Diffusion, it is trained with synthetic depth data only and excels in zero-shot adaptation to real-world imagery. Check it out: 🌐 Website:…

47

278

1K

1

4

Anton Obukhov

@AntonObukhov1

4 months

@sameerishwar After you create the depth map in our demo, scroll down below the examples -- you will find three sliders on the left. Create 3D, clear, adjust, repeat. HTH!

0

2

Anton Obukhov

@AntonObukhov1

4 years

Are conference deadlines getting any less thrilling?.. Has anyone tried submitting a paper a week in advance? I would be mildly interesting to see a histogram of uploads to cmt on the range from the call for papers to the deadline

2

0

4

Anton Obukhov

@AntonObukhov1

5 months

@popcornflix Not out of the box -- it is for single image inference. Still, will probably do better than other non-temporal algorithms, such as MiDaS/LeRes. If you have good findings, please share!

0

4

Anton Obukhov

@AntonObukhov1

3 years

Wouldn't it be great if (online) conference registration fee covered international shipping of swag? My last conference mug is from @NeurIPSConf '19, and every day feels more like a collectible @ICCV_2021 @CVPRConf

0

4

Anton Obukhov

@AntonObukhov1

3 years

Now that #PyTorch 1.9 is out, I got to play with the revamped linalg module. I'm glad I convinced the maintainers to rescue the Householder transformation fn initially slated for removal - ORGQR. It got a new name (householder_product), GPU, batch, and autograd support. We show

1

4

Anton Obukhov

@AntonObukhov1

2 years

#ICVSS2022 poster session and wine tasting: in progress

1

0

4

Anton Obukhov

@AntonObukhov1

5 months

@noodlecakeai @KotLesny Can you please create an issue on our github page to troubleshoot the 24GB memory issue? Please mention the command line you used and the input image resolution so that we can try to reproduce it. For the speed, yes, it is currently slow -- stay tuned for the next releases!

2

0

4

Anton Obukhov

@AntonObukhov1

7 months

3 days left until this gem of reusing pretrained Stable Diffusion is live at AI for 3D content creation workshop in ICCV in Paris!

Anton Obukhov

@AntonObukhov1

7 months

🎨 Excited to share our latest paper: "Breathing New Life into 3D Assets with Generative Repainting"! We present a novel approach to repaint 3D models using text prompts, pretrained Stable Diffusion, and NeRF. A step towards assisting content creators! 🖌️ 📄 Paper:…

1

8

38

0

4