This paper can be understood from two perspectives:</p>\n<ol>\n<li>A paradigm for multi-teacher distillation on image editing. </li>\n<li>A ready-to-use recipe for slashing LoRA serving costs.</li>\n</ol>\n<p>TL;DR: 50 effects, 1 LoRA, zero interference.</p>\n","updatedAt":"2026-05-29T04:20:36.317Z","author":{"_id":"637c941588699fba70e29f70","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/637c941588699fba70e29f70/b6G_QZkT-MhE47dx87i0d.png","fullname":"LIU JIAMING","name":"jamesliu1217","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":133,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8813140988349915},"editors":["jamesliu1217"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/637c941588699fba70e29f70/b6G_QZkT-MhE47dx87i0d.png"],"reactions":[],"isReport":false}},{"id":"6a1979b88ca2d54fe221bd4c","author":{"_id":"6960eca92f7ad9b043b5cbe0","avatarUrl":"/avatars/e68dcc7fd04f143d849d40414866e633.svg","fullname":"Noah","name":"noahml","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false},"createdAt":"2026-05-29T11:34:16.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"Made an audio walkthrough of this paper for anyone who wants to skim it on the go:\nhttps://researchpod.app/episode/a5a7fa29-dfde-4bfc-81f7-299f3e706bbc\n\nGenerated automatically by ResearchPod — happy to take feedback from the authors.","html":"<p>Made an audio walkthrough of this paper for anyone who wants to skim it on the go:<br><a href=\"https://researchpod.app/episode/a5a7fa29-dfde-4bfc-81f7-299f3e706bbc\" rel=\"nofollow\">https://researchpod.app/episode/a5a7fa29-dfde-4bfc-81f7-299f3e706bbc</a></p>\n<p>Generated automatically by ResearchPod — happy to take feedback from the authors.</p>\n","updatedAt":"2026-05-29T11:34:16.924Z","author":{"_id":"6960eca92f7ad9b043b5cbe0","avatarUrl":"/avatars/e68dcc7fd04f143d849d40414866e633.svg","fullname":"Noah","name":"noahml","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8522096872329712},"editors":["noahml"],"editorAvatarUrls":["/avatars/e68dcc7fd04f143d849d40414866e633.svg"],"reactions":[],"isReport":false}},{"id":"6a1a4169447ed909ef242f60","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":359,"isUserFollowing":false},"createdAt":"2026-05-30T01:46:17.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models](https://huggingface.co/papers/2605.05204) (2026)\n* [Flow-OPD: On-Policy Distillation for Flow Matching Models](https://huggingface.co/papers/2605.08063) (2026)\n* [Diffusion-APO: Trajectory-Aware Direct Preference Alignment for Video Diffusion Transformers](https://huggingface.co/papers/2605.07503) (2026)\n* [FREE-Switch: Frequency-based Dynamic LoRA Switch for Style Transfer](https://huggingface.co/papers/2604.10023) (2026)\n* [ACE-LoRA: Adaptive Orthogonal Decoupling for Continual Image Editing](https://huggingface.co/papers/2605.14948) (2026)\n* [Mamoda2.5: Enhancing Unified Multimodal Model with DiT-MoE](https://huggingface.co/papers/2605.02641) (2026)\n* [CPC-VAR:Continual Personalized and Compositional Generation in Visual Autoregressive Models](https://huggingface.co/papers/2605.19750) (2026)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"<p>This is an automated message from the <a href=\"https://huggingface.co/librarian-bots\">Librarian Bot</a>. I found the following papers similar to this paper. </p>\n<p>The following papers were recommended by the Semantic Scholar API </p>\n<ul>\n<li><a href=\"https://huggingface.co/papers/2605.05204\">D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2605.08063\">Flow-OPD: On-Policy Distillation for Flow Matching Models</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2605.07503\">Diffusion-APO: Trajectory-Aware Direct Preference Alignment for Video Diffusion Transformers</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2604.10023\">FREE-Switch: Frequency-based Dynamic LoRA Switch for Style Transfer</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2605.14948\">ACE-LoRA: Adaptive Orthogonal Decoupling for Continual Image Editing</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2605.02641\">Mamoda2.5: Enhancing Unified Multimodal Model with DiT-MoE</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2605.19750\">CPC-VAR:Continual Personalized and Compositional Generation in Visual Autoregressive Models</a> (2026)</li>\n</ul>\n<p> Please give a thumbs up to this comment if you found it helpful!</p>\n<p> If you want recommendations for any Paper on Hugging Face checkout <a href=\"https://huggingface.co/spaces/librarian-bots/recommend_similar_papers\">this</a> Space</p>\n<p> You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: <code><span class=\"SVELTE_PARTIAL_HYDRATER contents\" data-target=\"UserMention\" data-props=\"{&quot;user&quot;:&quot;librarian-bot&quot;}\"><span class=\"inline-block\"><span class=\"contents\"><a href=\"/librarian-bot\">@<span class=\"underline\">librarian-bot</span></a></span> </span></span> recommend</code></p>\n","updatedAt":"2026-05-30T01:46:17.719Z","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":359,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7087069153785706},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2605.25378","authors":[{"_id":"6a154d11b57a1823d5708d5c","name":"Fangtai Wu","hidden":false},{"_id":"6a154d11b57a1823d5708d5d","name":"Hailong Guo","hidden":false},{"_id":"6a154d11b57a1823d5708d5e","name":"Shijie Huang","hidden":false},{"_id":"6a154d11b57a1823d5708d5f","user":{"_id":"67051ad602b48c3fce373fe7","avatarUrl":"/avatars/ebf10382f7c34f41b77f30eb1eea784e.svg","isPro":false,"fullname":"sjy","user":"sjy92","type":"user","name":"sjy92"},"name":"Jiayi Song","status":"claimed_verified","statusLastChangedAt":"2026-05-29T09:40:22.699Z","hidden":false},{"_id":"6a154d11b57a1823d5708d60","name":"Yubo Huang","hidden":false},{"_id":"6a154d11b57a1823d5708d61","name":"Mushui Liu","hidden":false},{"_id":"6a154d11b57a1823d5708d62","name":"Zhao Wang","hidden":false},{"_id":"6a154d11b57a1823d5708d63","name":"Yunlong Yu","hidden":false},{"_id":"6a154d11b57a1823d5708d64","user":{"_id":"637c941588699fba70e29f70","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/637c941588699fba70e29f70/b6G_QZkT-MhE47dx87i0d.png","isPro":false,"fullname":"LIU JIAMING","user":"jamesliu1217","type":"user","name":"jamesliu1217"},"name":"Jiaming Liu","status":"claimed_verified","statusLastChangedAt":"2026-05-29T09:40:20.390Z","hidden":false},{"_id":"6a154d11b57a1823d5708d65","name":"Ruihua Huang","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/637c941588699fba70e29f70/wl1Ai6AIysZy1uC0v9Nlo.png"],"publishedAt":"2026-05-25T00:00:00.000Z","submittedOnDailyAt":"2026-05-29T00:00:00.000Z","title":"CollectionLoRA: Collecting 50 Effects in 1 LoRA via Multi-Teacher On-Policy Distillation","submittedOnDailyBy":{"_id":"637c941588699fba70e29f70","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/637c941588699fba70e29f70/b6G_QZkT-MhE47dx87i0d.png","isPro":false,"fullname":"LIU JIAMING","user":"jamesliu1217","type":"user","name":"jamesliu1217"},"summary":"Customized image editing aims to equip pre-trained diffusion models with specific visual effects using limited paired data, typically via Low-Rank Adaptation (LoRA). As the number of desired effects grows, storing and dynamically loading numerous these effect LoRAs significantly increases deployment overhead. Furthermore, current pipelines typically cascade these effect LoRAs with acceleration modules for fast generation, which triggers severe parameter interference and results in concept bleeding and style degradation. We propose CollectionLoRA, a multi-teacher on-policy distillation framework capable of distilling the concepts of up to 50 different effect LoRAs along with few-step generation capabilities into a single LoRA. This fundamentally resolves the feature interference issue and significantly reduces deployment costs. Specifically, the method introduces (i) a Probabilistic Dual-Stream Routing mechanism that enables the model to randomly switch between data sources during training, effectively enhancing its generalization in unseen scenarios; (ii) an Asymmetric Orthogonal Prompting strategy to achieve concept isolation within the prompt space; (iii) a Coarse-to-Fine Distillation Objective to mitigate the distribution gap between the teacher and student models. Extensive evaluations show that CollectionLoRA distills all customized effects and few-step generation into a single LoRA, reducing deployment overhead while achieving concept fidelity comparable to or better than independently trained teacher models.","upvotes":50,"discussionId":"6a154d11b57a1823d5708d66","projectPage":"https://collectionlora.github.io/","githubRepo":"https://github.com/Qwen-Applications/CollectionLoRA","githubRepoAddedBy":"user","ai_summary":"CollectionLoRA enables efficient deployment of multiple customized image editing effects by distilling numerous LoRAs into a single model through multi-teacher distillation and specialized mechanisms for concept isolation and generation.","ai_keywords":["Low-Rank Adaptation","LoRA","diffusion models","multi-teacher on-policy distillation","Probabilistic Dual-Stream Routing","Asymmetric Orthogonal Prompting","Coarse-to-Fine Distillation Objective","concept isolation","parameter interference","concept bleeding","style degradation"],"githubStars":14},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"637c941588699fba70e29f70","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/637c941588699fba70e29f70/b6G_QZkT-MhE47dx87i0d.png","isPro":false,"fullname":"LIU JIAMING","user":"jamesliu1217","type":"user"},{"_id":"6812f7f3508697fa11f1708b","avatarUrl":"/avatars/a036cf42eac4119bd32739b4fb5291c8.svg","isPro":false,"fullname":"Yi Chen Liu","user":"starry0929","type":"user"},{"_id":"660a645d5814b3dfc3a95e83","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/660a645d5814b3dfc3a95e83/S-yLw7-QFXg7fX94bfbtn.jpeg","isPro":false,"fullname":"Bingnan Li","user":"Cuttle-fish-my","type":"user"},{"_id":"67051ad602b48c3fce373fe7","avatarUrl":"/avatars/ebf10382f7c34f41b77f30eb1eea784e.svg","isPro":false,"fullname":"sjy","user":"sjy92","type":"user"},{"_id":"688cceac29c69ec01eb28644","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/wyL1U86pB5GIi4x3JflOa.png","isPro":false,"fullname":"Chuyue Li","user":"woody-woody","type":"user"},{"_id":"6605149000b4754d2b9d20ac","avatarUrl":"/avatars/5c4d91a9006a2d5ca1310f5751d1d3f9.svg","isPro":false,"fullname":"loong","user":"loooooong","type":"user"},{"_id":"692d7843a918d2fbb2893b15","avatarUrl":"/avatars/d68856827c5f1cd0aecc28398fc84757.svg","isPro":false,"fullname":"yang","user":"Newsyshi","type":"user"},{"_id":"681db91055699177c70a225b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/_v2L0W4VngDO2bync8ZfL.png","isPro":false,"fullname":"kangdi","user":"kangdi94120","type":"user"},{"_id":"6759582c44ec15bfe4bcf6a2","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/kSkj4OIjHXbYV7qhIhH78.png","isPro":false,"fullname":"li","user":"wang0422","type":"user"},{"_id":"69ce1551ac6abc0c359d87db","avatarUrl":"/avatars/a52dbcd7afdc4d5d5e5249987483e020.svg","isPro":false,"fullname":"shiwei yang","user":"Yangshiwei","type":"user"},{"_id":"69c5256dc134e4878dafa4c3","avatarUrl":"/avatars/a9b08cfef3a03f584734eb9792e75c02.svg","isPro":false,"fullname":"wu","user":"jiesong18","type":"user"},{"_id":"65e954e3588ff8b17899f05d","avatarUrl":"/avatars/44b5f4f12102b19d2bb374ad838ab6de.svg","isPro":false,"fullname":"L","user":"Shui-VL","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2605/2605.25378.md"}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2605.25378 in a model README.md to link it from this page.

Cite arxiv.org/abs/2605.25378 in a dataset README.md to link it from this page.

Cite arxiv.org/abs/2605.25378 in a Space README.md to link it from this page.

Add this paper to a collection to link it from this page.