Ceyuan Yang @CeyuanY Twitter profile

Last Seen Profiles

@CYBERWARCON

@USChinaBusiness

@JasonMBrodsky

@MarkReckons

@hermaissajesty

@RushTourism

@stwmaniax

@Da_VinChief

@sac4nagem

@stwmaniax

@BinorRaja

@IkoumNaomi

@RJayousi

@Spirit_2401

@StasiHoll

@princepark5252

@Single_hun_yr

@theryujinpics

@trenton_smith7

@quemendoooo

@kizakisato

@almighty_ouch

@BaWubba

@my_shsyh

@acaldermusic

@stwmaniax

@bokeplokalmalam

@amatory

@MstSumaiya42628

@Ibrahim98151131

@Jas_2101

@Nicolejqzhang

@steamy_crusader

@dodaveggiedance

@cotachi_books

@manamiogaki

Ceyuan Yang

@CeyuanY

1 year

Personalization means a lot. We thus introduce the lightweight motion LoRA into #AnimateDiff . Fine-grained motions like the following camera control are enabled. Models are available now at GitHub. Just create more motions as you want.

11

91

494

Ceyuan Yang

@CeyuanY

11 months

Update of #AnimateDiff : now can avoid watermarks and more importantly, support #SDXL i.e., produce high resolution videos (ie 1024x1024x16 frames with various aspect ratios) with/without personalized models (~13GB VRAM for inference). Try it out on .

6

52

280

Ceyuan Yang

@CeyuanY

10 months

Yuwei ( @GuoywGuo ) just released #AnimateDiff v3 and #SparseCtrl which allows to animate ONE keyframe, generate transition between TWO keyframes and interpolate MULTIPLE sparse keyframes. RGB images and scribbles are supported for now. Github:

14

57

318

Ceyuan Yang

@CeyuanY

1 year

Wanna make your personalized T2I (\eg, LoRA) move? Try our AnimateDiff which animates multiple models once for all without specific tuning. Page:

9

41

176

Ceyuan Yang

@CeyuanY

10 months

Controllability matters for video generation. We thus propose SparseCtrl that enables keyframe (sketch, depth, img) animation/transition/prediction/interpolation by feeding TEMPORALLY SPARSE condition maps. It’s also compatible with #AnimateDiff . Page:

6

32

171

Ceyuan Yang

@CeyuanY

1 year

If your GAN related research (e.g., image, video and 3D) is also bottlenecked by its scale, just like me, try our Aurora, a GAN based text-to-image generator. Surely, code and model will be publicly available soon. ArXiv:

1

13

59

Ceyuan Yang

@CeyuanY

3 years

Thank @ak92501 for sharing our recent work! Our GenDA could substantially improve the synthesis quality and diversity with few-shot even one-shot target image. Code is coming soon. Stay tuned!

AK

@_akhaliq

3 years

One-Shot Generative Domain Adaptation abs: project page: method works well even with large domain gaps, and robustly converges within a few minutes for each experiment

0

17

92

0

7

40

Ceyuan Yang

@CeyuanY

1 year

Finally came out

AK

@_akhaliq

1 year

LAVIE: High-Quality Video Generation with Cascaded Latent Diffusion Models paper page: This work aims to learn a high-quality text-to-video (T2V) generative model by leveraging a pre-trained text-to-image (T2I) model as a basis. It is a highly desirable

9

83

332

1

2

38

Ceyuan Yang

@CeyuanY

1 year

@_akhaliq Try it now on huggingface

AnimateDiff - a Hugging Face Space by guoyww

huggingface.co

1

5

31

Ceyuan Yang

@CeyuanY

2 years

Glad to share our #NeurIPS2022 work 'Improving GANs with A Dynamic Discriminator' which could easily improve 2D and 3D-aware image synthesis under various data regimes. Paper: Work with Yujun Shen, @YinghaoXu1 , Deli Zhao, @doubledaibo , @zhoubolei

1

3

24

Ceyuan Yang

@CeyuanY

11 months

Nice work by @YinghaoXu1

AK

@_akhaliq

11 months

DMV3D: Denoising Multi-View Diffusion using 3D Large Reconstruction Model paper page: propose DMV3D, a novel 3D generation approach that uses a transformer-based 3D large reconstruction model to denoise multi-view diffusion. Our reconstruction model

0

50

225

0

7

20

Ceyuan Yang

@CeyuanY

4 years

Check our #CVPR2020 work 'Temporal Pyramid Network for Action Recognition', a generic module to capture the visual tempos of action instances Paper: Github: Joint work with @YinghaoXu1 Jianping Shi @doubledaibo and @zhoubolei

1

4

22

Ceyuan Yang

@CeyuanY

10 months

Check out Qihang @QihangZhang0224 's previous work BerfScene that designs BEV-conditioned equivariant radiance fields for large-scale 3D scene generation and manipulation. Page:

1

5

26

Ceyuan Yang

@CeyuanY

3 years

Check our work 'Data-Efficient Instance Generation from Instance Discrimination', which advances image synthesis with limited data by a clear margin. Github: Paper: Joint work with Yujun Shen, @YinghaoXu1 , @zhoubolei

1

3

18

Ceyuan Yang

@CeyuanY

4 years

Glad to share our Generative Hierarchical Features (GHFeat) got accepted by #CVPR2021 as an ORAL presentation, which learns hierarchical representations from a well-trained GAN. Many thanks to my collaborators @YinghaoXu1 , Yujun Shen, Jiapeng Zhu and @zhoubolei .

0

1

17

Ceyuan Yang

@CeyuanY

4 years

Our Instance localization is accepted by #CVPR2021 😆. Code and models are coming soon. Stay tuned! Joint work with Zhirong Wu, @zhoubolei , Stephen Lin.

0

1

17

Ceyuan Yang

@CeyuanY

10 months

Cool stuff from the community!

Nathan Shipley

@CitizenPlain

10 months

Roll your own Motion Brush 🖌️ Here are 3 examples using simple masks, prompts, and the in-painting controlnet with #AnimateDiff to create (looping!) motion on still images. #stablediffusion #aivfx #aivideo #ComfyUI

27

175

1K

0

16

Ceyuan Yang

@CeyuanY

11 months

Speed up #AnimateDiff via LCM. Nice try.

toyxyz

@toyxyz3

11 months

1

3

29

1

0

10

Ceyuan Yang

@CeyuanY

1 year

Glad to see AnimateDiff enables Vid2Vid as well. Great job!

toyxyz

@toyxyz3

1 year

1

18

84

1

10

Ceyuan Yang

@CeyuanY

11 months

Thanks for the efforts from diffusers team and thank you, Dhruv. #AnimateDiff is now available in the latest diffusers!

Dhruv Nair

@_DhruvNair_

11 months

The latest version of diffusers 🧨 ships with AnimateDiff and Motion LoRAs from @CeyuanY , @GuoywGuo and team. And integrates with PEFT Create animations using text prompts with AnimateDiff and any SD 1.5 checkpoint. Control the movements using Motion LoRAs + PEFT. Details👇🏽

1

18

0

1

11

Ceyuan Yang

@CeyuanY

1 year

Hi All, with @GuoywGuo ’s hardworking, our animatediff now requires around 12G memory, that is, can run on a single RTX3090 without sacrificing quality due to reduced #frames or resolutions.

2

0

10

Ceyuan Yang

@CeyuanY

1 year

Inference code and weights are available now. Try it on

GitHub - zhujiapeng/Aurora: Official implementation of Aurora

Official implementation of Aurora. Contribute to zhujiapeng/Aurora development by creating an account on GitHub.

github.com

Ceyuan Yang

@CeyuanY

1 year

If your GAN related research (e.g., image, video and 3D) is also bottlenecked by its scale, just like me, try our Aurora, a GAN based text-to-image generator. Surely, code and model will be publicly available soon. ArXiv:

1

13

59

0

1

9

Ceyuan Yang

@CeyuanY

10 months

Qihang's #SceneWiz3D that enables text-to-3D scene generation and object-level manipulation!

Qihang Zhang

@QihangZhang0224

10 months

Excited to share our recent attempt for #3D scene generation🪐. Given text prompt, "SceneWiz3D" can generate high-fidelty 3D scenes and support flexible object control. Project webpage: Paper: Code:

3

48

204

0

11

Ceyuan Yang

@CeyuanY

1 year

@camenduru @raoanyi @doubledaibo The latest animatediff optimizes memory, requiring around 12G memory for now. So we don’t have reduce the number of frames or resolutions :)

2

1

8

Ceyuan Yang

@CeyuanY

1 year

@GuoywGuo Available on huggingface now.

AnimateDiff - a Hugging Face Space by guoyww

huggingface.co

0

2

7

Ceyuan Yang

@CeyuanY

3 years

Our InsGen is accepted by #NeurIPS2021 ✈️✈️ Code and pre-trained weights are available at . Grateful to my collaborators and reviewers!

GitHub - genforce/insgen: [NeurIPS 2021] Data-Efficient Instance Generation from Instance Discrim...

[NeurIPS 2021] Data-Efficient Instance Generation from Instance Discrimination - genforce/insgen

github.com

Ceyuan Yang

@CeyuanY

3 years

Check our work 'Data-Efficient Instance Generation from Instance Discrimination', which advances image synthesis with limited data by a clear margin. Github: Paper: Joint work with Yujun Shen, @YinghaoXu1 , @zhoubolei

1

3

18

0

1

6

Ceyuan Yang

@CeyuanY

4 years

The devil is in the details. It is really takes our time and efforts to design the framework, figure out the differences between various platforms, and debug day and night. Great teamwork! Hope this codebase could help everyone in generative modeling in future🎉🎉🎉

Bolei Zhou

@zhoubolei

4 years

No accepted papers, I am still happy to announce we open-source GenForce, an efficient PyTorch library for generative modeling. StyleGAN and PG-GAN are perfectly reproduced. We also collect a zoo of 60+ pretrained StyleGAN models, with Colab to play.

10

207

976

0

5

Ceyuan Yang

@CeyuanY

1 year

@Drex1165890669 Surely, training code is coming soon.

1

0

6

Ceyuan Yang

@CeyuanY

2 years

What if sekiro becomes a jedi master, perhaps holding a lightsaber? Amazing synthesis produced by @huggingface 's 🤣🤣🤣 Demo link:

0

1

6

Ceyuan Yang

@CeyuanY

1 year

@CiaraRowles1 Very impressive! Btw, a new checkpoint supporting 512x320 resolution and the same number of frames, based on SD1.5) will be available tomorrow. Stay tuned!

1

0

6

Ceyuan Yang

@CeyuanY

3 years

3D shape generation and 3D-aware image generation have recently drawn a lot of attention. Check out the reading list collected by @YinghaoXu1 , which includes awesome 3D generation papers.

GitHub - justimyhxu/awesome-3D-generation: A curated list of awesome 3d generation papers

A curated list of awesome 3d generation papers. Contribute to justimyhxu/awesome-3D-generation development by creating an account on GitHub.

github.com

0

3

Ceyuan Yang

@CeyuanY

4 years

It includes the first try of my research career when I was an undergraduate student. Be grateful to Prof. Gong Cheng.

arXiv Daily

@Arxiv_Daily

4 years

Remote Sensing Image Scene Classification Meets Deep Learning: Challenges, Methods, Benchmarks, and Opportunities by Gong Cheng et al. #Autoencoder #ConvolutionalNeuralNetworks

0

5

0

3

Ceyuan Yang

@CeyuanY

11 months

cool

AK

@_akhaliq

11 months

Great waves off KanimateDiff by ai_maak

4

122

608

0

3

Ceyuan Yang

@CeyuanY

11 months

@HuanLing6 More details generated by SDXL. For appearance, training video matters.

0

Ceyuan Yang

@CeyuanY

10 months

@EMostaque @huggingface I personally appreciate the efforts of opensource. Stability AI really helps advance the field of AIGC.

0

2

Ceyuan Yang

@CeyuanY

1 year

@OrenElbaum @a_liveankur @doubledaibo @camenduru wow, 720x480？

0

Ceyuan Yang

@CeyuanY

11 months

@peteromallet Appealing results!

1

0

1

Ceyuan Yang

@CeyuanY

1 year

@jon_barron thanks for sharing！

0

1

Ceyuan Yang

@CeyuanY

1 year

@jfischoff @camenduru @raoanyi @doubledaibo Flash attention (Implemented by xformer) helps a lot. Besides, inference frames by frames for SD decoder also matters

1

0

1

Ceyuan Yang

@CeyuanY

1 year

@raoanyi @Stanford @UCBerkeley @Caltech 🐮

0

1

Ceyuan Yang

@CeyuanY

1 year

@_akhaliq Thanks for sharing!!

0

1

Ceyuan Yang

@CeyuanY

11 months

@pcuenq Looking forward to testing it on video generations

0

1

Ceyuan Yang

@CeyuanY

1 year

@jianyuan_wang See u in Paris.

0

1

Ceyuan Yang

@CeyuanY

1 year

@TDS_95514874 amazing results！

0

1

Ceyuan Yang

@CeyuanY

11 months

@YinghaoXu1 @_akhaliq @KaiZhang9546 @HaoTan5 @zexiangxu @GordonWetzstein Well done🐮

1

0

1

Ceyuan Yang

@CeyuanY

11 months

@the_marconi @netflix @NetflixBrasil I really like this one🤣

1

0

1

Ceyuan Yang

@CeyuanY

3 years

I will present our #NeurIPS2021 work InsGen. Please visit our poster session on 12/8 12:30 am - 2 am PT! Poster: Paper: Code:

Ceyuan Yang

@CeyuanY

3 years

Check our work 'Data-Efficient Instance Generation from Instance Discrimination', which advances image synthesis with limited data by a clear margin. Github: Paper: Joint work with Yujun Shen, @YinghaoXu1 , @zhoubolei

1

3

18

2

0

1