Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

AnimateLCM: Accelerating the Animation of Personalized Diffusion Models and Adapters with Decoupled Consistency Learning (2024)
Consistent3D: Towards Consistent High-Fidelity Text-to-3D Generation with Deterministic Sampling Prior (2024)
PIXART-δ: Fast and Controllable Image Generation with Latent Consistency Models (2024)
Sequential Flow Straightening for Generative Modeling (2024)
The Surprising Effectiveness of Skip-Tuning in Diffusion Sampling (2024)

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2024-03-02T01:20:40.204Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":264}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7610328793525696},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2402.19159","authors":[{"_id":"65e15242b8517a94d137ccfd","user":{"_id":"641870135d6f3d15c64d074e","avatarUrl":"/avatars/35f4b33fd1d4db326e4ee4300b26db72.svg","isPro":false,"fullname":"Jianbin Zheng","user":"jabir-zheng","type":"user"},"name":"Jianbin Zheng","status":"claimed_verified","statusLastChangedAt":"2024-03-01T09:26:04.302Z","hidden":false},{"_id":"65e15242b8517a94d137ccfe","user":{"_id":"630b77f68b327c7b8b98c409","avatarUrl":"/avatars/2fe95c9ac95f34dbb031f2ec018f68b0.svg","isPro":false,"fullname":"Minghui Hu","user":"h1t","type":"user"},"name":"Minghui Hu","status":"claimed_verified","statusLastChangedAt":"2024-03-01T05:40:10.665Z","hidden":false},{"_id":"65e15242b8517a94d137ccff","user":{"_id":"642e74226a378e41aa551994","avatarUrl":"/avatars/3a3618a214dcf5913e8a3d2f4a1cd5b8.svg","isPro":false,"fullname":"Zhongyi Fan","user":"zyfan","type":"user"},"name":"Zhongyi Fan","status":"admin_assigned","statusLastChangedAt":"2024-03-01T10:43:45.711Z","hidden":false},{"_id":"65e15242b8517a94d137cd00","name":"Chaoyue Wang","hidden":false},{"_id":"65e15242b8517a94d137cd01","name":"Changxing Ding","hidden":false},{"_id":"65e15242b8517a94d137cd02","name":"Dacheng Tao","hidden":false},{"_id":"65e15242b8517a94d137cd03","name":"Tat-Jen Cham","hidden":false}],"publishedAt":"2024-02-29T13:44:14.000Z","submittedOnDailyAt":"2024-03-01T02:54:41.094Z","title":"Trajectory Consistency Distillation","submittedOnDailyBy":{"_id":"60f1abe7544c2adfd699860c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674929746905-60f1abe7544c2adfd699860c.jpeg","isPro":false,"fullname":"AK","user":"akhaliq","type":"user"},"summary":"Latent Consistency Model (LCM) extends the Consistency Model to the latent\nspace and leverages the guided consistency distillation technique to achieve\nimpressive performance in accelerating text-to-image synthesis. However, we\nobserved that LCM struggles to generate images with both clarity and detailed\nintricacy. To address this limitation, we initially delve into and elucidate\nthe underlying causes. Our investigation identifies that the primary issue\nstems from errors in three distinct areas. Consequently, we introduce\nTrajectory Consistency Distillation (TCD), which encompasses trajectory\nconsistency function and strategic stochastic sampling. The trajectory\nconsistency function diminishes the distillation errors by broadening the scope\nof the self-consistency boundary condition and endowing the TCD with the\nability to accurately trace the entire trajectory of the Probability Flow ODE.\nAdditionally, strategic stochastic sampling is specifically designed to\ncircumvent the accumulated errors inherent in multi-step consistency sampling,\nwhich is meticulously tailored to complement the TCD model. Experiments\ndemonstrate that TCD not only significantly enhances image quality at low NFEs\nbut also yields more detailed results compared to the teacher model at high\nNFEs.","upvotes":16,"discussionId":"65e15247b8517a94d137d178","ai_summary":"Trajectory Consistency Distillation (TCD) improves text-to-image synthesis by addressing errors in consistency models, leading to higher image quality and detail at low numerical flow evaluations.","ai_keywords":["Latent Consistency Model (LCM)","Consistency Model","latent space","guided consistency distillation","trajectory consistency function","strategic stochastic sampling","Probability Flow ODE","TCD","image quality","numerical flow evaluations (NFEs)"]},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"630b77f68b327c7b8b98c409","avatarUrl":"/avatars/2fe95c9ac95f34dbb031f2ec018f68b0.svg","isPro":false,"fullname":"Minghui Hu","user":"h1t","type":"user"},{"_id":"641870135d6f3d15c64d074e","avatarUrl":"/avatars/35f4b33fd1d4db326e4ee4300b26db72.svg","isPro":false,"fullname":"Jianbin Zheng","user":"jabir-zheng","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"63c5d43ae2804cb2407e4d43","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1673909278097-noauth.png","isPro":false,"fullname":"xziayro","user":"xziayro","type":"user"},{"_id":"6311bca0ae8896941da24e66","avatarUrl":"/avatars/48de64894fc3c9397e26e4d6da3ff537.svg","isPro":false,"fullname":"Fynn Kröger","user":"fynnkroeger","type":"user"},{"_id":"648eb1eb59c4e5c87dc116e0","avatarUrl":"/avatars/c636cea39c2c0937f01398c94ead5dad.svg","isPro":false,"fullname":"fdsqefsgergd","user":"T-representer","type":"user"},{"_id":"6362ddb7d3be91534c30bfd6","avatarUrl":"/avatars/dac76ebd3b8a08099497ec0b0524bc7c.svg","isPro":false,"fullname":"Art Atk","user":"ArtAtk","type":"user"},{"_id":"6064e095abd8d3692e3e2ed6","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1648966381588-6064e095abd8d3692e3e2ed6.jpeg","isPro":true,"fullname":"Radamés Ajna","user":"radames","type":"user"},{"_id":"6538119803519fddb4a17e10","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6538119803519fddb4a17e10/ffJMkdx-rM7VvLTCM6ri_.jpeg","isPro":false,"fullname":"samusenps","user":"samusenps","type":"user"},{"_id":"6495d5e8f1d3ee1d68de7721","avatarUrl":"/avatars/8d57ec468df68d1d1eea9f9b8eacac72.svg","isPro":false,"fullname":"Muhammad Maxalmina Magnum","user":"Maxyro33354","type":"user"},{"_id":"61848f9a62753793d7ffabaa","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1636077456577-noauth.jpeg","isPro":false,"fullname":"Hoyeong Heo","user":"hotohoto","type":"user"},{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0}">

Papers

arxiv:2402.19159

Trajectory Consistency Distillation

Published on Feb 29, 2024

· Submitted by

AK on Mar 1, 2024

Upvote

Authors:

Jianbin Zheng ,

Minghui Hu ,

Zhongyi Fan ,

Abstract

Trajectory Consistency Distillation (TCD) improves text-to-image synthesis by addressing errors in consistency models, leading to higher image quality and detail at low numerical flow evaluations.

AI-generated summary

Latent Consistency Model (LCM) extends the Consistency Model to the latent space and leverages the guided consistency distillation technique to achieve impressive performance in accelerating text-to-image synthesis. However, we observed that LCM struggles to generate images with both clarity and detailed intricacy. To address this limitation, we initially delve into and elucidate the underlying causes. Our investigation identifies that the primary issue stems from errors in three distinct areas. Consequently, we introduce Trajectory Consistency Distillation (TCD), which encompasses trajectory consistency function and strategic stochastic sampling. The trajectory consistency function diminishes the distillation errors by broadening the scope of the self-consistency boundary condition and endowing the TCD with the ability to accurately trace the entire trajectory of the Probability Flow ODE. Additionally, strategic stochastic sampling is specifically designed to circumvent the accumulated errors inherent in multi-step consistency sampling, which is meticulously tailored to complement the TCD model. Experiments demonstrate that TCD not only significantly enhances image quality at low NFEs but also yields more detailed results compared to the teacher model at high NFEs.