{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/omnimamba-efficient-and-unified-multimodal","title":"OmniMamba: Efficient and Unified Multimodal Understanding and Generation via State Space Models","arxiv_id":"2503.08686","date":"2025-03-11","proceeding":null,"authors":["Jialv Zou","Bencheng Liao","Qian Zhang","Wenyu Liu","Xinggang Wang"],"abstract":"Recent advancements in unified multimodal understanding and visual generation (or multimodal generation) models have been hindered by their quadratic computational complexity and dependence on large-scale training data. We present OmniMamba, the first linear-architecture-based multimodal generation model that generates both text and images through a unified next-token prediction paradigm. The model fully leverages Mamba-2's high computational and memory efficiency, extending its capabilities from text generation to multimodal generation. To address the data inefficiency of existing unified models, we propose two key innovations: (1) decoupled vocabularies to guide modality-specific generation, and (2) task-specific LoRA for parameter-efficient adaptation. Furthermore, we introduce a decoupled two-stage training strategy to mitigate data imbalance between two tasks. Equipped with these techniques, OmniMamba achieves competitive performance with JanusFlow while surpassing Show-o across benchmarks, despite being trained on merely 2M image-text pairs, which is 1,000 times fewer than Show-o. Notably, OmniMamba stands out with outstanding inference efficiency, achieving up to a 119.2 times speedup and 63% GPU memory reduction for long-sequence generation compared to Transformer-based counterparts. Code and models are released at https://github.com/hustvl/OmniMamba","url_abs":"https://arxiv.org/abs/2503.08686v1","url_pdf":"https://arxiv.org/pdf/2503.08686v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"omnimamba-efficient-and-unified-multimodal","repo_url":"https://github.com/hustvl/omnimamba","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":null,"task_name":"GPU"},{"task_slug":"mamba","task_name":"Mamba"},{"task_slug":"state-space-models","task_name":"State Space Models"},{"task_slug":"text-generation","task_name":"Text Generation"},{"task_slug":"multimodal-generation","task_name":"multimodal generation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2503.08686","atlas_url":"https://app.syntology.ai/?focus=2503.08686","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2503.08686"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/hustvl/omnimamba","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran_draft_wrong":4,"ran_fixture":1,"unverified":3},"by_repo_kind":{"official":{"samples":8,"ran":5,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"c3a6b977022957cb","entry":"Normalize","repo":"hustvl/omnimamba","repo_kind":"official","path":"llamagen_tokenizer/vqgan/layer.py","file_url":"https://github.com/hustvl/omnimamba/blob/HEAD/llamagen_tokenizer/vqgan/layer.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c3a6b977022957cb"}},{"code_sha256_prefix":"01b5cf639b6bac89","entry":"Normalize","repo":"hustvl/omnimamba","repo_kind":"official","path":"llamagen_tokenizer/tokenizer_image/vq_model.py","file_url":"https://github.com/hustvl/omnimamba/blob/HEAD/llamagen_tokenizer/tokenizer_image/vq_model.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"01b5cf639b6bac89"}},{"code_sha256_prefix":"774c95b77d7621f3","entry":"compute_entropy_loss","repo":"hustvl/omnimamba","repo_kind":"official","path":"llamagen_tokenizer/tokenizer_image/vq_model.py","file_url":"https://github.com/hustvl/omnimamba/blob/HEAD/llamagen_tokenizer/tokenizer_image/vq_model.py","link_basis":"harvester_set","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"774c95b77d7621f3"}},{"code_sha256_prefix":"22a83df891d750c6","entry":"hinge_d_loss","repo":"hustvl/omnimamba","repo_kind":"official","path":"llamagen_tokenizer/tokenizer_image/vq_loss.py","file_url":"https://github.com/hustvl/omnimamba/blob/HEAD/llamagen_tokenizer/tokenizer_image/vq_loss.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"22a83df891d750c6"}},{"code_sha256_prefix":"3137073275f8c21a","entry":"nonlinearity","repo":"hustvl/omnimamba","repo_kind":"official","path":"llamagen_tokenizer/tokenizer_image/vq_model.py","file_url":"https://github.com/hustvl/omnimamba/blob/HEAD/llamagen_tokenizer/tokenizer_image/vq_model.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3137073275f8c21a"}},{"code_sha256_prefix":"546aeb3f73727064","entry":"create_training_arguments","repo":"hustvl/omnimamba","repo_kind":"official","path":"train_stage2.py","file_url":"https://github.com/hustvl/omnimamba/blob/HEAD/train_stage2.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"546aeb3f73727064"}},{"code_sha256_prefix":"dd2278aa669b6313","entry":"non_saturating_d_loss","repo":"hustvl/omnimamba","repo_kind":"official","path":"llamagen_tokenizer/tokenizer_image/vq_loss.py","file_url":"https://github.com/hustvl/omnimamba/blob/HEAD/llamagen_tokenizer/tokenizer_image/vq_loss.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"dd2278aa669b6313"}},{"code_sha256_prefix":"519cb60c0e3fbcb2","entry":"vanilla_d_loss","repo":"hustvl/omnimamba","repo_kind":"official","path":"llamagen_tokenizer/tokenizer_image/vq_loss.py","file_url":"https://github.com/hustvl/omnimamba/blob/HEAD/llamagen_tokenizer/tokenizer_image/vq_loss.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"519cb60c0e3fbcb2"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}