Co-founder & CTO @pika_labs | ex @StanfordAILab @Stanford | Hiring: Head of Engineering and Senior Engineer
- Tweets 341
- Following 1,753
- Followers 9,165
- Likes 2,660
Super excited for the launch of Pika 1.0! I am extremely grateful to be working with such an amazing and talented team on this journey! ❤️
I am also very thankful for the support from our incredible investors, advisors, friends, and community members. We couldn't have achieved this without you! :)
Introducing Pika 1.0, the idea-to-video platform that brings your creativity to life.
Create and edit your videos with AI.
Rolling out to new users on web and discord, starting today. Sign up at pika.art
52
32
606
168,505
Excited to share our work "On distillation of guided diffusion models"! Our distillation approach allows classifier-free guided diffusion models to generate high-quality samples using as few as 1-4 sampling steps😮
9
85
566
Excited to share our work SDEdit @YSongStanford, @baaadas, @jiajunwu_cs, @junyanz89, @StefanoErmon. Based on stochastic differential equations, SDEdit allows various image editing and synthesis applications without task specific training and optimization.
github.com/ermongroup/SDEdit
2
55
248
Welcome @omerbartal 🎉🎉 Looking forward to the amazing journey ahead! 🦾
Thrilled to share that I've joined @pika_labs as Founding Scientist!
It's exciting times in video generation and AI, and we're committed to unlock new capabilities and empower creative content creation.
Delighted to work with the most talented @demi_guo_ @chenlin_meng🚀
12
105
110
7,857
Score matching has proven effective in modeling continuous data modalities. However, this approach is not applicable in discrete domains where the gradient is undefined. We propose an analogous score function called the “Concrete score” for discrete data. #NeurIPS2022
2
16
147
Super excited to announce our new round led by @sparkcapital !! Extremely grateful for our amazing investors and the entire Pika team!! Really lucky to work with all of you ❤️
Can't wait for the new journey ahead 🦾
Excited to announce our $80M Series B funding led by @sparkcapital.
Grateful to have our investors and team members joining us along this journey - can't do this without you all.
Stay tuned for more updates from us later this year! 🙌
Read more: bloomberg.com/news/articles/…
17
5
139
28,229
Excited to share our #ICLR2023 paper "Dual Diffusion Implicit Bridges for Image-to-Image Translation" (DDIB), an image-to-image translation approach with diffusion models 🐱 -> 🐯
github.com/suxuann/ddib
1
25
129
24,463
Super excited to share our update on PIKA 1.5. Check out our mind-blowing Pikaffects 🤯
Sry, we forgot our password.
PIKA 1.5 IS HERE.
With more realistic movement, big screen shots, and mind-blowing Pikaffects that break the laws of physics, there’s more to love about Pika than ever before.
Try it.
2
49
4,540
We will be presenting our work SDEdit at #ICLR2022
Tue 26 Apr 6:30 p.m. PDT - 8:30 p.m. PDT! Come to our poster session today and see how SDE-based generative models can be used for guided image synthesis and editing! 🎨👩🎨
iclr.cc/virtual/2022/poster/…
1
4
52
Check out the paper for more details:
arxiv.org/abs/2210.03142
1
5
50
Interested in making diffusion models fast? I’ll present our paper On Distillation of Guided Diffusion Models and other interesting works at #NeurIPS workshop on score-based methods this afternoon.
📍/🕜: room 293-294 / 2:30pm - 2:40pm and 3:00pm-4:30pm
Excited to share our work "On distillation of guided diffusion models"! Our distillation approach allows classifier-free guided diffusion models to generate high-quality samples using as few as 1-4 sampling steps😮
1
3
50
Pika 2.1 is dropping soon 😁
Want early access? Click here to apply yf9ce6yzh87.typeform.com/pik…
Are you ready for the big reveal? 👀🐰
Drumroll please…
🥁🥁🥁
🥁🥁🥁
🥁🥁🥁
Pika 2.1 is dropping soon.
Want early access? Apply via the link in thread.
2
9
50
7,253
Try out Pika 2.2 and Pikaframes! ✨ Turn any image into something entirely new—powered by pure imagination. Try it now at pika.art and pika.art/ios! 🪄
Pika 2.2 is HERE, with 10s generations, 1080p resolution, and Pikaframes— key frame transitions anywhere from 1-10s. More transformation, more imagination. Try it at Pika dot art
1
44
5,244
🚨 More editing results with SDEdit, an image synthesis and editing framework based on stochastic differential equations @StefanoErmon 😎😲
Joint work with @YSongStanford @baaadas @jiajunwu_cs @junyanz89 @StefanoErmon
github.com/ermongroup/SDEdit 🙂
7
38
Thank you so much for everyone's contribution to our tutorial! 😍😎🚀 There are so many amazing works but we are only able to cover part of them 😢Here are some more papers github.com/cvpr2023-tutorial… Please feel free to add any related works to the repo!! ❤️
Sharing a list of very useful references from @chenlin_meng’s wonderful tutorial on the latest diffusion models for text-to-image synthesis, semantically-guided image editing, and controllable generation! 🎑🌁🌅🏞️🌃🌄
#CVPR2023
6
41
13,978
This approach is also effective for video generation and has been used in #ImagenVideo: it can distill video diffusion models down to just 8 sampling steps without noticeable loss in perceptual quality!
Work with @RuiqiGao, @dpkingma, @StefanoErmon, @hojonathanho, @TimSalimans
1
3
38
Here are some ImageNet samples using only one sampling step!
(w denotes the guidance strength for a classifier-free guided diffusion model)
1
2
35
Pika 2.1 is available to public! It supports 1080p resolution, sharp details, smooth/dynamic motion, lifelike human characters and many more. Want to unlock more potential, test it out yourself at pika.art 💗💗
Getting really real with you…
Pika 2.1 is HERE—crystal-clear 1080p resolution, razor-sharp details, seamless motion, lifelike human characters, and more. If you see it, you’ll believe it.
Try it out now at pika dot art.
5
38
3,199
תודה על התמיכה ב-Pika 2.1 🩷 זה באמת מרגש לראות. אם יש לך רעיונות או פידבק לשיפור, נשמח לשמוע
4
1
36
1,740
Super grateful for our incredible community, our amazing investors, and the entire Pika team!!❤️
It’s been a hell of a year. And today we’ve capped it off with a humbling fundraise, led by @sparkcapital. Huge thanks to our incredible community. We wouldn’t have gotten here without you!
4
4
36
6,853
Our #ICML2023 Workshop on Structured Probabilistic Inference & Generative Modeling will take place at Hawaii Convention Center, Room 323 on Friday, July 28
Come to our workshop to learn more about generative modeling and meet our amazing speakers and panelists! 😎☺️
2
4
31
10,226
Welcome @EladRichardson to the Pika Family!! 🥳☺️ So excited! @pika_labs
2
15
4,829
We also explore distilling diffusion models with guidance for style transfer applications.
Orange🍊 -> Bell pepper 🫑 (16 sampling steps)
Orange 🍊 -> Acorn squash 🍈 (16 sampling steps)
(w denotes the guidance strengths)
1
24
Excited to share our NeurIPS 2020 paper on training autoregressive score models:
arxiv.org/abs/2010.12810
joint work with @yulantao1996, @YSongStanford, @baaadas and @StefanoErmon
5
26
Come to our poster session on Thu, Dec 1 at 09:00 AM -- 11:00 AM (PST) @ Hall J 115 #NeurIPS2022
Paper: arxiv.org/abs/2211.00802
Joint work with @kristyechoi @baaadas @StefanoErmon
3
23
@pika_labs will be at ICLR! Ping us if you’re there #ICLR2024 🥳 Our swags will never let you down ☺️
Flying tomorrow to ICLR!
Happy to chat about video generation, diffusion models, and @pika_labs !
(this bag contains only Pika swag)
2
1
21
2,449
Pikadditions is here! You can insert anything you want into a video 😄 Try it out at pika.art 🥳😄
Wait…WHAT? Introducing Pikadditions, the easiest way to make your content stand out.
Add anyone or anything to any video, whether that’s a video you shoot yourself, or a favorite clip. Special surprise: get fifteen free Pikadditions generations when you sign up!
Go try it at pika dot art
3
22
2,624
Our approach works well for a wide range of guidance strengths using only one distilled model per dataset. We also develop a stochastic sampler for the distilled model. We observe that the stochastic sampler often outperforms the deterministic sampler in terms of Inception scores
1
1
21
We have just launched our mobile app. Check it out at pika.art/ios 😍😚
Introducing the official Pika app!
Now our reality-bending features are available right where you post & share.
Download it now on the iOS app store. Link in comments.
1
18
2,397
Organizing our workshop on Structured Probabilistic Inference and Generative Modeling at ICML 2023 😎
If you are interested in reviewing and learning the frontiers of probabilistic inference and generative models 🪄, then please fill out the form below😊
Interested in reviewing and learning the frontiers of probabilistic inference and generative models? Sign up this form forms.gle/jd6xGWGBgF8ukXLA6 for ICML 2023 workshop Structured Probabilistic Inference and Generative Modeling! #ICML2023
2
18
5,513
A new app
Some news: We're building the next big thing — the first-ever AI-only social video app, built on a highly expressive human video model. Over the past few weeks, we’ve been testing it in private beta. Now, we’re opening early access: download the iOS app to join the waitlist, or skip the line with an invite code. Codes are limited — if you know someone with early access, they might have one for you.
apps.apple.com/gb/app/pika-s…
19
3,020
One of the most exciting works I have ever seen🥳
Please check out at pika.art 😄
Excited to introduce Pikadditions!
Upload any video and make it more engaging by adding any object/character to it, based on an image reference.
Check it out at pika dot art! @pika_labs
17
1,064
Pikadditions and Pikaswaps are now supporting 1080p. Try it out yourself 🥳🙌
Pikadditions and Pikaswaps are now available in stunning 1080p resolution, with enhanced accuracy and precision for an ultra-realistic experience.
Available now at Pika dot art for all Pika paid plans. Coming soon to the app.
1
16
1,676
Wild examples with Pikadditions 😍🪄
Try it out yourself at pika.art
16
1,601
Here is the schedule of our #SIGGRAPH2023 Course on Diffusion Models with the amazing
@baaadas,@ShuangL13799063,@junyanz89, @karsten_kreis, @chenhsuan_lin, @TsungYiLinCV, @StefanoErmon 😎
Looking forward to seeing everyone tomorrow in Room 403 AB! 😊
Schedule:
🔸9:00am-10:00am Introduction to Diffusion Models
🔸10:00am-10:30am Latent Diffusion Models and Their Applications
🔸10:30am-11:00am Compositional Visual Generation
🔸11:00am-11:30am Customizing Diffusion Models
🔸11:30am-12:00am Diffusion Models for 3D Asset Generation
1
1
13
9,750
We will be hosting an event at @siggraph 🥂
Send me a DM if you are around 😄
warming up for @siggraph in august - @chenlin_meng and me will host another @pika_labs salon on august 11.
let us know if you'll be around - would love to meet with technical artists or artsy researchers :)
1
11
2,620
Come to our ICML workshop on July 28! ☺️🏖️
spigmworkshop.github.io/
Don't miss out on our SPIGM workshop (spigmworkshop.github.io/) at #ICML2023 in Hawaii🏖️ next week!
For better communication, please feel free to join our Slack channel: join.slack.com/t/probabilist… to connect with like-minded individuals and say hi to the community👩💻!
2
10
11,506
What if you could add your favorite items to any meme or video clip? Let’s make this a never-ending thread of fun and creativity! 😍 Drop a reply below, and let’s see what wild ideas we can bring to life with Pikadditions! 🪄
7
1,019
Thank you so much Aravid! Super grateful to have you as our early investor. Really thankful for your support!
7
1,172
Work led by Xuan Su, an amazing master's student at Stanford 😎. Joint work with @baaadas and @StefanoErmon
[2/7]
1
6
943
Really impressive work from @FreddyChavezO !!
FADING LIGHTS- A surreal experimental piece exploring the fragility of mortality, created using Pika 2.1, Runway Frames, Magnific and Suno.
3
6
1,237
Empirically, we demonstrate the efficacy of CSM on density estimation tasks on a mixture of synthetic, tabular, and high-dimensional image datasets, and demonstrate that it performs favorably relative to existing baselines for modeling discrete data.
1
6
Welcome @yuvalalaluf !! Looking forward to the amazing journey ahead 🚀
Very happy to share that I've joined @pika_labs as a Research Scientist! 😁
I'm excited about this new adventure and the opportunity to work with such an amazing team! 📽️🚀
ALT Ideogram
2
6
2,786
“Concrete score” is a generalization of the (Stein) score for discrete settings. Given a predefined neighborhood structure, the Concrete score of any input is defined by the rate of change of the probabilities with respect to local directional changes of the input.
1
5
My amazing collaborators @robrombach @RuiqiGao will be presenting our work at @CVPR ! @robrombach will also be giving the Award Candidate Talk this Thursday afternoon. Feel free to stop by and ask any questions! Looking forward to the discussion! 😎
#CVPR2023
1
5
1,744
Developing cutting-edge research and turning it into real, usable products is where the magic happens. 🪄
4
381
Super grateful for your support from Day 1! We are really lucky to have you as our investor ☺️ Couldn't ask for more
4
546
Our distillation approach allows classifier-free guided diffusion models to generate high-quality samples using as few as 1-4 sampling steps 😎 It is also effective for data modalities beyond images 😮🚀
1
4
882
This formulation allows us to recover the (Stein) score in continuous domains when measuring such changes by the Euclidean distance, while using the Manhattan distance leads to our novel score function in discrete domains.
1
4
Our proposed approach DDIB solves both problems:
✅ It does not require diffusion models to be trained jointly on two domains.
✅ It can be easily adapted to other domain pairs by leveraging pre-trained off-the-shelf diffusion models.
[4/7]
1
2
4
2,351
Joint work with my amazing collaborators and advisors
@electronickale @YSongStanford @baaadas @jiajunwu_cs @junyanz89 @StefanoErmon😊
3
Based on stochastic differential equations, SDEdit allows various image editing and synthesis applications without task-specific training and optimization.
1
3
Finally, we introduce a new framework to learn such scores from samples called Concrete Score Matching (CSM), and propose an efficient training objective to scale our approach to high dimensions.
1
3
Theoretically, we can interpret the translation procedure as two concatenated Schrödinger Bridges: an optimal transport process where the diffusion models minimize the pixel-wise distance between the translated images (see animation below)
[6/7]
1
3
1,504
Current image translation methods such as GANs:
⛔️ require the presence of both datasets during training, which betrays data separation and data privacy.
⛔️️cannot be easily adapted to translations between other source-target pairs.
[3/7]
1
3
794
We provide a wide variety of results on both synthetic and ImageNet datasets.
Check out the details in our paper and project page!
arxiv.org/abs/2203.08382
suxuann.github.io/ddib/
[7/7]
1
2
1,118