{"id":"arxiv:2310.04378","title":"Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference","date":"2023-10-06","section":"Self-supervised learning and vision","quality":0.195723,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/","source_url":"https://arxiv.org/abs/2310.04378","text_available":true,"authors":["Simian Luo","Yiqin Tan","Longbo Huang","Jian Li"],"outgoing":[],"incoming":[{"source":"arxiv:2608.00146","target":"arxiv:2310.04378","evidence":"title","id":"arxiv:2608.00146","title":"DiffusionGemma Technical Report","url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/90a7b084496cd65968627c115230fc3a5e78005a66b5ebb34547e874754bc82e/"}],"text_parts":[{"start":0,"end":15389,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-1.html"},{"start":15389,"end":31306,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-2.html"},{"start":31306,"end":46964,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html"},{"start":46964,"end":62723,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-4.html"}],"contents":[{"title":"Latent Consistency Models: Synthesizing High-Resolution Images with Few-step Inference","offset":0,"level":1,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-1.html#text"},{"title":"Abstract","offset":90,"level":6,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-1.html#text"},{"title":"1 Introduction","offset":1341,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-1.html#text"},{"title":"2 Related Work","offset":4971,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-1.html#text"},{"title":"3 Preliminaries","offset":7043,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-1.html#text"},{"title":"4 Latent Consistency Models","offset":14033,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-1.html#text"},{"title":"4.1 Consistency Distillation in the Latent Space","offset":15335,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-1.html#text"},{"title":"4.2 One-Stage Guided Distillation by solving augmented PF-ODE","offset":20242,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-2.html#text"},{"title":"4.3 Accelerating Distillation with Skipping Time Steps","offset":25279,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-2.html#text"},{"title":"4.4 Latent Consistency Fine-tuning for customized dataset","offset":30540,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-2.html#text"},{"title":"5 Experiment","offset":31289,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-2.html#text"},{"title":"5.1 Text-to-Image Generation","offset":31786,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html#text"},{"title":"5.2 Ablation Study","offset":35225,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html#text"},{"title":"5.3 Downstream Consistency Fine-tuning Results","offset":37475,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html#text"},{"title":"6 Conclusion","offset":38312,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html#text"},{"title":"Appendix A More Details on Diffusion and Consistency Models","offset":38892,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html#text"},{"title":"A.1 Diffusion Models","offset":38956,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html#text"},{"title":"A.2 More Details on Consistency Models in ( Song et al., 2023 )","offset":42512,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html#text"},{"title":"Appendix B Multistep Latent Consistency Sampling","offset":45400,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html#text"},{"title":"Appendix C Algorithm Details of Latent Consistency Fine-tuning","offset":46897,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-3.html#text"},{"title":"Appendix D Different ways to parameterize the consistency function","offset":48093,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-4.html#text"},{"title":"Appendix E Formulas of Other ODE Solvers","offset":50873,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-4.html#text"},{"title":"Appendix F Training Details of Latent Consistency Distillation","offset":55588,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-4.html#text"},{"title":"Appendix G Reproduction Details of Guided-Distill","offset":58149,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-4.html#text"},{"title":"Appendix H More few-step Inference Results","offset":61254,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-4.html#text"},{"title":"Instructions for reporting errors","offset":61759,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/part-4.html#text"}],"text_code_points":62723,"source_markdown":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/16b9865215d555f17a2d1abc00c0915667686f20c49898f63872327a416690db/source.md"}