{"paper":{"title":"In-Context Edit: Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Large Diffusion Transformers perform precise instructional image editing via in-context generation without major retraining.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ji Xie, Yi Yang, Yu Lu, Zechuan Zhang, Zongxin Yang","submitted_at":"2025-04-29T12:14:47Z","abstract_excerpt":"Instruction-based image editing enables precise modifications via natural language prompts, but existing methods face a precision-efficiency tradeoff: fine-tuning demands massive datasets (>10M) and computational resources, while training-free approaches suffer from weak instruction comprehension. We address this by proposing ICEdit, which leverages the inherent comprehension and generation abilities of large-scale Diffusion Transformers (DiTs) through three key innovations: (1) An in-context editing paradigm without architectural modifications; (2) Minimal parameter-efficient fine-tuning for "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"ICEdit achieves state-of-the-art editing performance with only 0.1% of the training data and 1% trainable parameters compared to previous methods.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the inherent comprehension and generation abilities of large-scale Diffusion Transformers can be effectively leveraged for precise instructional editing through an in-context paradigm without any architectural modifications.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"ICEdit achieves state-of-the-art instructional image editing in Diffusion Transformers via in-context generation, requiring only 0.1% of prior training data and 1% trainable parameters.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Large Diffusion Transformers perform precise instructional image editing via in-context generation without major retraining.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"a30c54734c9e7ab2952e1db28f3caff75a0a24a0b731dd6cdb0869296cde5e7a"},"source":{"id":"2504.20690","kind":"arxiv","version":3},"verdict":{"id":"1e73ec36-2161-4fe0-b2f0-207517cbbcb5","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T16:04:06.609653Z","strongest_claim":"ICEdit achieves state-of-the-art editing performance with only 0.1% of the training data and 1% trainable parameters compared to previous methods.","one_line_summary":"ICEdit achieves state-of-the-art instructional image editing in Diffusion Transformers via in-context generation, requiring only 0.1% of prior training data and 1% trainable parameters.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the inherent comprehension and generation abilities of large-scale Diffusion Transformers can be effectively leveraged for precise instructional editing through an in-context paradigm without any architectural modifications.","pith_extraction_headline":"Large Diffusion Transformers perform precise instructional image editing via in-context generation without major retraining."},"references":{"count":60,"sample":[{"doi":"","year":2023,"title":"Instructpix2pix: Learning to follow image editing instructions","work_id":"fd7a4946-0389-427a-bb45-c75d5e7c48c3","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Magicbrush: A manually annotated dataset for instruction-guided image editing.Advances in Neural Information Processing Systems, 36, 2024","work_id":"da65b41a-11ec-470b-8042-e8459a3c6095","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Emu edit: Precise image editing via recognition and generation tasks","work_id":"041fd7a8-5b0e-41c8-9dd4-1750ebee5c81","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Ultraedit: Instruction-based fine-grained image editing at scale.Advances in Neural Information Processing Systems, 37:3058–3093, 2025","work_id":"63911165-3afc-4d06-8929-abf84f29ab58","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"arXiv preprint arXiv:2309.17102 (2023)","work_id":"2bb3b94c-6b63-43cc-8c25-25ed4a0de29c","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":60,"snapshot_sha256":"8df2c86e799a633ea6a0b76f35c0d47ce4df2644664e872c963ef45ccd0c2e84","internal_anchors":13},"formal_canon":{"evidence_count":1,"snapshot_sha256":"76fb4d9e70ba5fd058da1aa94c524198b6a56a3e291d4193f52fbbaab528c58e"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}