Nataniel Ruiz @natanielruizg Twitter profile

Last Seen Profiles

@hurtwoodbooks

@Grandaddy

@husnuarkan

@ChooseYourself2

@DrStrange

@g9_generation

@22Pick

@alkhailabdulwah

@trackstarjoshCS

@HeronHockey

@AmeClub_

@infoffmc45

@Sabrina_Sparks2

@itemhjs

@notthateric

@shidamonze_

@CHOILABENT

@ShivyBabi

@ptaehvyung

@Yagmur818

@standrewsfemsoc

@harumame2

@succubusD20

@ElizabethHoyt

@ninuexe_

@SensoCrtico1

@djjimmyjatt

@joseanusalicari

@Somninous

@dhanjukulbir

@_MATTERR_

@thawhitegirl_

@mozzerllabih

@Ciize155cm

@FirstCow_JP

@AgroBrasil22

Nataniel Ruiz

@natanielruizg

2 years

Today, along with my collaborators at @GoogleAI , we announce DreamBooth! It allows a user to generate a subject of choice (pet, object, etc.) in myriad contexts and with text-guided semantic variations! The options are endless. (Thread 👇) webpage: 1/N

46

422

2K

Nataniel Ruiz

@natanielruizg

10 months

Today, with collaborators at @Google , we're excited to announce 🥳🥳HyperDreamBooth🥳 🥳! It's like DreamBooth, but smaller, faster and better. 25x faster. Think of 30 minutes vs. 14 hours for 100 models. And works on a single image! (Thread 👇) webpage:

42

304

2K

Nataniel Ruiz

@natanielruizg

6 months

With collaborators @Google we're announcing 💫 ZipLora 💫! Merging LoRAs has been a big thing in the community, but tuning can be an onerous process. ZipLora allows us to easily combine any subject LoRA with any style LoRA! Easy to reimplement 🥳 link:

33

237

1K

Nataniel Ruiz

@natanielruizg

11 months

Today, along with collaborators at @GoogleAI , we’re excited to announce StyleDrop! It allows a user to generate new images that follow a specific style of their choice given only a single style reference image 🤯 (Thread 👇) webpage:

49

253

1K

Nataniel Ruiz

@natanielruizg

7 months

Today, with collaborators at Google, we're announcing 🤩RealFill🤩! A generative AI approach to fill missing regions of an image with the content that should have been there. The best way to turn almost perfect pictures into invaluable memories! page:

27

145

668

Nataniel Ruiz

@natanielruizg

2 years

We made the first comic fully generated by an AI using DreamBooth finetuning of Imagen at @GoogleAI . We use only 4 input images of our cute little character Anselmo (drawn by my friend @IsaArduz ). (thread) 1/N #DreamBooth #AIArt #Imagen #stablediffusion #midjourney #AIArtwork

16

78

430

Nataniel Ruiz

@natanielruizg

1 month

AI generated writing *feels* AI-generated at a visceral level, and even if you ask an LLM to make the writing feel or read less AI-generated it horrifically fails and makes it feel even more AI-generated. Any tricks that can help? Any prompts to share?

232

8

370

Nataniel Ruiz

@natanielruizg

1 year

Passed my dissertation defense and obtained my PhD 🥳 a super fun journey, would definitely do it again if I could. Video will be uploaded soon.

39

12

350

Nataniel Ruiz

@natanielruizg

9 months

We are 🔥super excited🔥 to release the Platypus family of finetuned LLMs 🥳🥳. Platypus achieves the top score in the Hugging Face Open LLM Leaderboard 🏆! The main focus of our work is to achieve cheap, fast and powerful refinement of base LLMs. page:

14

85

308

Nataniel Ruiz

@natanielruizg

6 years

@cpicciolini @SamHarrisOrg Can you address his explanation, namely that you attributed specific positions to those two people which they do not actually hold?

5

2

253

Nataniel Ruiz

@natanielruizg

2 years

@bradesposito Interestingly she describes exactly what is wrong with streaming now. Many times I have to see an album cover to remember that it’s great and I should listen to it again!

3

7

262

Nataniel Ruiz

@natanielruizg

1 year

Super happy to announce that I will be joining @Google as a Research Scientist and will be starting tomorrow! Extremely excited by this new step and very grateful for everyone that made this possible. 🥳🥳🥳

31

4

258

Nataniel Ruiz

@natanielruizg

1 year

Super happy to announce that DreamBooth has been selected as an award candidate at CVPR 2023 (0.51% award rate). 🥳🥳🥳 link:

10

30

242

Nataniel Ruiz

@natanielruizg

1 year

🥳 DreamBooth has been accepted to CVPR 2023. And with this comes a *big update* to the paper including the largest evaluation dataset for subject driven generation and an evaluation protocol! Find it in the project webpage: (a thread) #Dreambooth 1/N

3

26

189

Nataniel Ruiz

@natanielruizg

7 months

Our team is looking for student researchers doing a PhD starting in January either full-time or part-time (prefer full-time). If you want to work on new exciting applications and methods like I did with DreamBooth, then please reach out. DMs open.

6

32

169

Nataniel Ruiz

@natanielruizg

6 years

Trump and his special assistant for infrastructure policy presenting the deepest neural network.

4

32

159

Nataniel Ruiz

@natanielruizg

6 months

Ok. We will release a very cool project soon. Easy to reimplement 🥳

13

5

159

Nataniel Ruiz

@natanielruizg

10 months

Excited 🥳🥳🥳to release my first senior author work, done while still a student at BU, with a start studded lineup of collaborators and an incredible student first-author @ArielNLee 🌻🙌- it's all about differences between Vision Transformers & CNNs 👇

4

31

133

Nataniel Ruiz

@natanielruizg

2 years

What we've been waiting for. I'm extremely excited for this paper.

AK

@_akhaliq

2 years

On Distillation of Guided Diffusion Models abs: On ImageNet 64x64 and CIFAR-10, approach is able to generate images visually comparable to that of the original model using as few as 4 sampling steps

5

129

630

3

15

122

Nataniel Ruiz

@natanielruizg

1 year

Today, at NeurIPS, we announce counterfactual simulation testing, a new framework for comparing vastly different network architectures using counterfactuals. We use it to compare the robustness of modern ConvNets and Transformers. (Thread 👇) webpage:

4

18

116

Nataniel Ruiz

@natanielruizg

2 years

Our method has some surprising capabilities inherited from large diffusion models. For example it can generate novel art renditions of a subject! Here are some renditions of a specific dog in the style of famous painters. 4/N

3

10

113

Nataniel Ruiz

@natanielruizg

2 years

@MIT_CSAIL GPT-3 says: "A neural network is a computer system that is modeled after the brain." It didn't understand the instruction.

5

1

104

Nataniel Ruiz

@natanielruizg

2 years

Public service announcement: Don’t use “sks” as a token for dreambooth. SKS is a type of rifle.

maderix

@originalmaderix

2 years

The runwayml version of #stablediffusion 1.5 model with Dreambooth tends to produce lot of guns by default lol 🤖

6

3

52

9

106

Nataniel Ruiz

@natanielruizg

6 months

I think this has to be some sort of speed record 🏅

mkshing

@mk1stats

6 months

So, I quickly implemented the ZipLoRA by 🤗🧨 (Some people have already noticed though) code: I hope it helps somehow and feel free to drop your comments and feedback~ Big thanks to the authors for their awesome work 🙌

5

60

323

3

9

102

Nataniel Ruiz

@natanielruizg

7 months

First RealFill open-source re-implementation finished 🔥🔥 link:

GitHub - thuanz123/realfill

Contribute to thuanz123/realfill development by creating an account on GitHub.

github.com

Thuan Hoang Nguyen

@thuanz123

7 months

@natanielruizg I think my job with RealFill is done @natanielruizg

2

27

0

16

102

Nataniel Ruiz

@natanielruizg

11 months

DreamBooth won an Honorable Mention Award at #CVPR23 (6 out of more than 8000 submissions, 0.08% rate) 🥳🥳🥳

#CVPR2024

@CVPR

11 months

1

2

22

4

100

Nataniel Ruiz

@natanielruizg

4 years

@marwilliamson Marianne, Venezuela tanked its own economy with 15 years of irresponsible economic policy.

25

3

86

Nataniel Ruiz

@natanielruizg

3 months

the more I look at the videos, the more the motions feel like a video game (the walking here), but the appereance of only some videos looks like video game footage. Maybe this model is trained on a lot of game footage? models are good at learning to change style simulated->real

OpenAI

@OpenAI

3 months

Introducing Sora, our 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. Prompt: “Beautiful, snowy…

10K

33K

141K

19

6

96

Nataniel Ruiz

@natanielruizg

11 months

@PatentlyApple There’s almost a 100% chance they knew this and named it Vision Pro anyway. They have a plan

1

0

87

Nataniel Ruiz

@natanielruizg

2 months

Some really cool work by our soon-to-be-intern @JialuLi96 link:

0

12

85

Nataniel Ruiz

@natanielruizg

2 years

Text-to-image diffusion models are extremely powerful and allow for flexible generation of images with complex user captions. One limitation is that controlling the subject’s appearance and identity using text is very hard. 2/N

3

5

80

Nataniel Ruiz

@natanielruizg

2 years

We can even do realistic viewpoint changes for some subjects which have a strong class prior! Here are some examples of different viewpoints for a cat. Notice the detailed fur patterns in the forehead are conserved. 🤯 7/N

1

7

78

Nataniel Ruiz

@natanielruizg

2 years

By finetuning a model (Imagen here) with few images of a subject (~3-5), a user can generate variations of the subject. E.g. by controlling the environment and context of the subject. Ever wanted to have a high-quality picture of your dog in Paris (no travel required)? 3/N

1

4

77

Nataniel Ruiz

@natanielruizg

2 years

And another! 🚀 @GoogleAI

AK

@_akhaliq

2 years

UniTune: Text-Driven Image Editing by Fine Tuning an Image Generation Model on a Single Image abs:

7

148

714

1

8

75

Nataniel Ruiz

@natanielruizg

2 years

We can also change semantic attributes of a subject. Re-coloring, chimeras, material changes, etc. 5/N

3

5

72

Nataniel Ruiz

@natanielruizg

9 months

@andrewmccalip

1

0

69

Nataniel Ruiz

@natanielruizg

10 months

My first paper as senior author (done while I was still a PhD student at BU!). So proud of Ariel and grateful for all coauthors 🙏🌸. I feel blessed. Thread coming out tomorrow 🔥

AK

@_akhaliq

10 months

Hardwiring ViT Patch Selectivity into CNNs using Patch Mixing paper page: Vision transformers (ViTs) have significantly changed the computer vision landscape and have periodically exhibited superior performance in vision tasks compared to convolutional…

3

19

101

0

7

72

Nataniel Ruiz

@natanielruizg

4 years

@TectonixGEO @vdbDennis @xmodesocial I mean how anonymized is it really if you can track a phone location? You can easily figure out where people live, and identifying the person is one-step away (maybe even a Google search away)

3

0

63

Nataniel Ruiz

@natanielruizg

2 years

What about accessorization? Given a few images of your pet you could accessorize them with extreme flexibility. Imagination is the limit! 6/N

3

9

70

Nataniel Ruiz

@natanielruizg

7 months

Cool work which proposes a very similar "lower-rank" LoRA like Lightweight DreamBooth that we proposed in our HyperDreamBooth work () but for LLMs. 10x reduction in size, just like in our case!

AK

@_akhaliq

7 months

VeRA: Vector-based Random Matrix Adaptation paper page: Low-rank adapation (LoRA) is a popular method that reduces the number of trainable parameters when finetuning large language models, but still faces acute storage challenges when scaling to even…

6

62

270

0

17

68

Nataniel Ruiz

@natanielruizg

6 months

Another record - someone *cough* @mk1stats *cough* made a @Gradio demo in no time 🤍

AK

@_akhaliq

6 months

ZipLoRA-pytorch with @Gradio demo by @mk1stats local demo: Methods for finetuning generative models for concept-driven personalization generally achieve strong results for subject-driven or style-driven generation. Recently, low-rank adaptations (LoRA)…

2

85

371

0

10

69

Nataniel Ruiz

@natanielruizg

2 years

Finally, our method can generate new images of a subject with different expressions/emotions. Note that the original images of the subject dog here did not exhibit any of these expressions. 8/N

1

4

65

Nataniel Ruiz

@natanielruizg

10 months

Very impressive work, a must read.

AK

@_akhaliq

10 months

Domain-Agnostic Tuning-Encoder for Fast Personalization of Text-To-Image Models paper page: Text-to-image (T2I) personalization allows users to guide the creative image generation process by combining their own visual concepts in natural language…

1

68

258

1

7

66

Nataniel Ruiz

@natanielruizg

5 months

This is seriously crazy. No-finetuning styledrop-like generation

Andrey Voynov

@kusichan

5 months

Happy to announce StyleAligned – our new work from @GoogleAI : Style Aligned Image Generation via Shared Attention 📜 arXiv: 👀 project page: (with quiz game!) 💻 code: Details below! ⬇️

19

114

516

1

3

67

Nataniel Ruiz

@natanielruizg

9 months

🚀 Presenting our latest SOTA LLM: OpenOrca-Platypus2-13B 🚀. Kudos to @ArielNLee and @ColeJHunter and the great people of @alignment_lab for topping the Hugging Face leaderboard in the 13B parameter category! Excited by this collaboration. link:

2

17

65

Nataniel Ruiz

@natanielruizg

1 year

I’m defending my PhD thesis tomorrow 🎉 at 3pm EST. It’s called: Simulating to Learn. Such a fun journey. Will post the video afterwards. If you want the zoom link send me a dm.

12

2

64

Nataniel Ruiz

@natanielruizg

2 years

@JxckSweeney @elonmusk So you run a Twitter account that tracks Musk's jet purportedly because it is "of service" and "interesting", yet here you are offering to take it down if the amount they pay you is enough? I don't understand.

7

1

61

Nataniel Ruiz

@natanielruizg

2 years

@MIT_CSAIL If I constrain to 5 words it says: "A neural network is a" lol

1

64

Nataniel Ruiz

@natanielruizg

9 months

@wangzjeff Haha well not drinking does not mean you have to be boring!

3

0

62

Nataniel Ruiz

@natanielruizg

4 years

@JamesTodaroMD @elonmusk James, there has been a lot of criticism of the Santa Clara study and it might overestimate positive cases because of the biased sample and the false positive rate of antibody tests. The IFR that I computed of 0.1% with that data would mean prevalence of more than 100% in NYC.

3

1

56

Nataniel Ruiz

@natanielruizg

2 years

We have many other details on the method in the paper. Feel free to check it out! arxiv: 11/N

DreamBooth: Fine Tuning Text-to-Image Diffusion Models for...

Large text-to-image models achieved a remarkable leap in the evolution of AI, enabling high-quality and diverse synthesis of images from a given text prompt. However, these models lack the ability...

arxiv.org

2

3

58

Nataniel Ruiz

@natanielruizg

6 months

Thank you @_akhaliq for the retweet! 🙏

Nataniel Ruiz

@natanielruizg

6 months

With collaborators @Google we're announcing 💫 ZipLora 💫! Merging LoRAs has been a big thing in the community, but tuning can be an onerous process. ZipLora allows us to easily combine any subject LoRA with any style LoRA! Easy to reimplement 🥳 link:

33

237

1K

0

6

55

Nataniel Ruiz

@natanielruizg

11 months

But here is a result I really didn't expect. What surprises me is how well it handles the translation of ideas into arbitrary styles, changing the object shape to fit the style - and following stylistic flourishes and geometrical style components.

2

3

57

Nataniel Ruiz

@natanielruizg

1 month

Improved how

OpenAI

@OpenAI

1 month

Majorly improved GPT-4 Turbo model available now in the API and rolling out in ChatGPT.

447

755

5K

14

0

55

Nataniel Ruiz

@natanielruizg

5 months

DreamBooth SDXL Turbo works it seems

Emad

@EMostaque

5 months

Yeah so SDXL Turbo tunes. Upcoming version tunes on consumer graphics cards nicely… Let’s just say things are just getting started…

8

5

129

2

3

54

Nataniel Ruiz

@natanielruizg

11 months

Congratulations to @kihyuk_sohn , @dilipkay and to all authors involved in this work! The list is long and can be found below. For more amazing examples go to the project page. paper: project webpage:

StyleDrop: Text-to-Image Generation in Any Style

styledrop.github.io

4

53

Nataniel Ruiz

@natanielruizg

2 years

Another great Google work on diffusion models! 🚀🚀

AK

@_akhaliq

2 years

Imagic: Text-Based Real Image Editing with Diffusion Models abs:

22

422

2K

1

7

50

Nataniel Ruiz

@natanielruizg

4 months

Some incredibly cool work by Google. Don’t miss it!

AK

@_akhaliq

4 months

Google announces PALP Prompt Aligned Personalization of Text-to-Image Models paper page: Content creators often aim to create personalized images using personal subjects that go beyond the capabilities of conventional text-to-image models. Additionally,…

2

110

455

1

3

50

Nataniel Ruiz

@natanielruizg

5 months

With some cool people from @huggingface and @weights_biases ( @altryne , @linoy_tsaban , @multimodalart ) 🥳

2

4

49

Nataniel Ruiz

@natanielruizg

1 year

DreamBooth featured at Google I/O 🥳 on an insane concept: a card game with 7+ Million unique generated characters! Amazing work by the I/O Flip team! 🤯 The first instance of such a card game? (clip linked)

✂️ IO Flip

60 seconds · Clipped by Nataniel Ruiz · Original video "Developer keynote (Google I/O '23)" by Google for Developers

www.youtube.com

2

8

49

Nataniel Ruiz

@natanielruizg

11 months

Yuanzhen Li presenting DreamBooth, DreamBooth3D and StyleDrop at #CVPR2023 !

3

5

50

Nataniel Ruiz

@natanielruizg

2 years

One main difficulty in finetuning a diffusion model using few images is overfitting. We tackle this problem by presenting an autogenous class-specific prior preservation loss. More details in the paper. 9/N

1

2

49

Nataniel Ruiz

@natanielruizg

9 months

In order to do so we propose an optimized, small, yet very powerful dataset named Open-Platypus, which is a curated subset of open datasets and focuses on enhancing LLMs' STEM and logic proficiency. We release this dataset to the public.

garage-bAInd/Open-Platypus · Datasets at Hugging Face

huggingface.co

2

6

47

Nataniel Ruiz

@natanielruizg

11 months

Before diving into technical details, let's explore some impressive examples. StyleDrop can extract the color palette and overall style from this watercolor cat painting, and generate almost anything one can imagine in that same style.

1

3

48

Nataniel Ruiz

@natanielruizg

8 months

Anyone can now use StyleDrop 🥳 - announced at the Google Cloud Next '23 event! link:

5

9

47

Nataniel Ruiz

@natanielruizg

3 months

I think @Scenario_gg are pushing the limits of DreamBooth in crazy ways. They really are alchemists working with the original DreamBooth idea to make it much stronger and to be able to do more things with it.

Mint

@araminta_k

3 months

We just made creating your next Consistent Character waaaaaaay easier :D Workflow 1/3 I am sharing THREE workflows this week to using the new "Character Base" LoRAs that we just added to Scenario to: - Use as a consistent character - Create a new consistent character from -…

5

48

238

4

5

47

Nataniel Ruiz

@natanielruizg

10 months

🚀 NeRFs are getting better and faster

Jon Barron

@jon_barron

10 months

Our freshly minted ICCV2023 paper: The nice anti-aliasing of mip-NeRF 360, but with most of the speed of Instant NGP. Error rate reductions of 8%-77% compared to either prior technique, and 24x faster than the most accurate NeRF baseline we tried.

16

93

692

0

1

45

Nataniel Ruiz

@natanielruizg

2 years

We are able to alleviate overfitting using this approach. We show that finetuning without this loss term leads to accelerated overfitting of subject pose and appearance, or context. This decreases generation variability and incorrect scenes. 10/N

2

4

44

Nataniel Ruiz

@natanielruizg

2 years

We also thank the Imagen team for lending us access to their incredible model. And we deeply thank all of the great people who helped with reviews and feedback (all acknowledged in the paper). Again, our project website is: 13/13 (END)

7

0

45

Nataniel Ruiz

@natanielruizg

1 year

Train CLIP using non-industrial scale resources! Some amazing work by a good friend of min @cihangxie !

AK

@_akhaliq

1 year

An Inverse Scaling Law for CLIP Training abs: paper page:

3

39

220

0

13

44

Nataniel Ruiz

@natanielruizg

4 years

@afneil The study is hard to read. From what I saw it 1. is a retrospective study 2. treats patients that are severely ill, probably later in the course of the disease. HCQ has in vitro antiviral effects against SARS-CoV-2 and should be used EARLY. Not effective to use it late!

2

35

Nataniel Ruiz

@natanielruizg

11 months

This also happens with 3D styles. Here, even the isometric viewpoint is captured. As well as the rounded edges, layout and color palette.

1

42

Nataniel Ruiz

@natanielruizg

1 year

@thatfollowed @sama It’s pretty awesome to have a CEO that is tethered to reality. A lot of respect!

0

42

Nataniel Ruiz

@natanielruizg

2 years

@Bryce_Nickels @R_H_Ebright @cshperspectives Quick question, is Markolin a portmanteau of Market and Pangolin?

6

5

39

Nataniel Ruiz

@natanielruizg

2 months

I need to retweet this again because I feel like it needs to get more attention. Awesome read

Robin Rombach

@robrombach

2 months

Party time! The SD3 paper made it to arxiv: Key takeaways: - flow matching is very nice. - back to work with @pess_r and a fantastic team ♥️ The paper is full of details on improved flow matching, scaling and engineering. Enjoy!

10

38

267

2

39

Nataniel Ruiz

@natanielruizg

5 years

@TheAnnaGat @clairlemon Now I understand why the bay area has so many Libertarians

2

0

31

Nataniel Ruiz

@natanielruizg

4 years

@marwilliamson The sanctions were primarily targeted towards the regime (who wine and dine at expensive restaurants while the people starve). I just don’t agree with this specific example.

13

1

33

Nataniel Ruiz

@natanielruizg

4 years

@TectonixGEO @vdbDennis @xmodesocial What do you think about authoritarian government use of this technology?

5

0

30

Nataniel Ruiz

@natanielruizg

9 months

Collecting Google offices Google Austin ✅

7

1

38

Nataniel Ruiz

@natanielruizg

11 months

I'll be revealing something super super cool tomorrow 🥳🤯

8

0

37

Nataniel Ruiz

@natanielruizg

3 years

@alexandrosM @R_H_Ebright This letter is pretty startling I have to say. As scientists, how could they have been so certain about the origins of the virus about a month after the news of the outbreak? It's always important to have a little bit of doubt when the evidence is not fully there yet

1

34

Nataniel Ruiz

@natanielruizg

7 months

"This Should Be Impossible!" 🥳🥳 Our RealFill work at @Google made it into Two Minute Paper ( @twominutepapers 🙏). Truly great presentation of the work!

Google’s AI: This Should Be Impossible!

❤️ Check out Weights & Biases and sign up for a free demo here: https://wandb.me/papers📝 The paper "RealFill - Reference-Driven Generation for Authentic Ima...

www.youtube.com

2

6

37

Nataniel Ruiz

@natanielruizg

4 months

Some truly beautiful work by colleagues at Google 🌹🌻🌸

Hila Chefer@ ICLR’24

@hila_chefer

4 months

TLDR: Meet ✨Lumiere✨ our new text-to-video model from @GoogleAI ! Lumiere is designed to create entire clips in just one go! Seamlessly opening up possibilities for many applications: Image-to-video 🖼️ Stylized generation 🖌️ Video editing 🪩 and beyond. See 🧵👇

77

207

954

5

0

37

Nataniel Ruiz

@natanielruizg

8 months

Google Paris ✅

7

1

36

Nataniel Ruiz

@natanielruizg

1 year

Wow. No wayyy. Super cool!

AK

@_akhaliq

1 year

New simple @Gradio dreambooth web ui by @mirage_ml on @huggingface Spaces upload images, add concept prompt and click start training demo:

3

29

157

1

4

35

Nataniel Ruiz

@natanielruizg

10 months

The core idea is to train a HyperNetwork that predicts weight deltas for the diffusion model in order to make it personalized. This initialization is strong enough that, given fast finetuning, we can achieve great identity preservation with impressive editability and variety 🔥

2

3

33

Nataniel Ruiz

@natanielruizg

6 months

We observe impressive subject detail preservation with a lot of fidelity to the user-provided style. I find these very cool. 🧙‍♂️

1

3

35

Nataniel Ruiz

@natanielruizg

1 month

Some insane work by colleagues at Google. Really impressive

AK

@_akhaliq

1 month

Google presents ObjectDrop Bootstrapping Counterfactuals for Photorealistic Object Removal and Insertion Diffusion models have revolutionized image editing but often generate images that violate physical laws, particularly the effects of objects on the scene, e.g.,

6

140

616

2

1

34

Nataniel Ruiz

@natanielruizg

2 years

Thank you for your time. And thank you to all of my collaborators @AbermanKfir , Yuanzhen Li, @jampani_varun , @MikiRubinstein , Yael Pritch. I had an amazing time working on this with you and am looking forward to future uses of this technology and more research! 12/N

1

0

33

Nataniel Ruiz

@natanielruizg

3 months

I’m kinda done with people posting screenshots of a single example of a failed LLM query and going “absolutely trash model, XYZ model is much better” Have seen this happen for every single popular model out there. You would think they’re all bad and never give good answers

5

3

32

Nataniel Ruiz

@natanielruizg

8 months

We have something very cool coming out soon 🤩🤩🤩 (suspense)

1

0

33

Nataniel Ruiz

@natanielruizg

2 years

These are the input images for our character Anselmo. We generate a fully-fledged comic with Anselmo in new poses, with different accessories and even with text and speech bubbles automatically drawn by the diffusion model! (just prompt for "a [V] cartoon saying XYZ"!) 2/N

2

1

31

Nataniel Ruiz

@natanielruizg

10 months

Ok big thing coming out tomorrow 🚀

3

0

30

Nataniel Ruiz

@natanielruizg

11 months

How does it work? We use MUSE, a masked Transformer for Text-to-Image Synthesis. (project: ). MUSE seems to have some properties that make it excel at learning and reproducing style.

1

30

Nataniel Ruiz

@natanielruizg

6 months

Insane speedups (and available for use) by @neuralmagic and @mgoin_ Very impressive… we need this for diffusion models.

AK

@_akhaliq

6 months

Fast Llama 2 on CPUs With Sparse Fine-Tuning and DeepSparse by @neuralmagic with @Gradio demo demo: run with docker: duplicate space for private use: blog:

3

52

206

0

2

29

Nataniel Ruiz

@natanielruizg

11 months

Imagic 🤝 DreamBooth @bahjat_kawar #CVPR2023

1

30

Nataniel Ruiz

@natanielruizg

10 months

The first key idea is Lightweight DreamBooth (LiDB), a customized model that is only ~100KB instead of more than 1GB for a typical Stable Diffusion model. This makes it 10k times smaller.