{"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/mixture-of-transformers-a-sparse-and-scalable","title":"Mixture-of-Transformers: A Sparse and Scalable Architecture for Multi-Modal Foundation Models","arxiv_id":"2411.04996","date":"2024-11-07","proceeding":null,"authors":["Weixin Liang","Lili Yu","Liang Luo","Srinivasan Iyer","Ning Dong","Chunting Zhou","Gargi Ghosh","Mike Lewis","Wen-tau Yih","Luke Zettlemoyer","Xi Victoria Lin"],"abstract":"The development of large language models (LLMs) has expanded to multi-modal systems capable of processing text, images, and speech within a unified framework. Training these models demands significantly larger datasets and computational resources compared to text-only LLMs. To address the scaling challenges, we introduce Mixture-of-Transformers (MoT), a sparse multi-modal transformer architecture that significantly reduces pretraining computational costs. MoT decouples non-embedding parameters of the model by modality -- including feed-forward networks, attention matrices, and layer normalization -- enabling modality-specific processing with global self-attention over the full input sequence. We evaluate MoT across multiple settings and model scales. In the Chameleon 7B setting (autoregressive text-and-image generation), MoT matches the dense baseline's performance using only 55.8\\% of the FLOPs. When extended to include speech, MoT reaches speech performance comparable to the dense baseline with only 37.2\\% of the FLOPs. In the Transfusion setting, where text and image are trained with different objectives, a 7B MoT model matches the image modality performance of the dense baseline with one third of the FLOPs, and a 760M MoT model outperforms a 1.4B dense baseline across key image generation metrics. System profiling further highlights MoT's practical benefits, achieving dense baseline image quality in 47.2\\% of the wall-clock time and text quality in 75.6\\% of the wall-clock time (measured on AWS p4de.24xlarge instances with NVIDIA A100 GPUs).","url_abs":"https://arxiv.org/abs/2411.04996v2","url_pdf":"https://arxiv.org/pdf/2411.04996v2.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":"mixture-of-transformers-a-sparse-and-scalable","repo_url":"https://github.com/allenzren/open-pi-zero","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"image-generation","task_name":"Image Generation"}],"methods":[{"method_slug":"attention","method_name":"Attention"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2411.04996","atlas_url":"https://app.syntology.ai/?focus=2411.04996","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2411.04996"}},"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/allenzren/open-pi-zero","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran":10,"ran_fixture":1,"ran_honours":1,"unverified":1},"by_repo_kind":{"listed":{"samples":13,"ran":12,"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":"d949d0d0bc5e63bc","entry":"conditional_decorator","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/utils/decorator.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/utils/decorator.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d949d0d0bc5e63bc"}},{"code_sha256_prefix":"7f39687304604d5d","entry":"get_action_accuracy","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/utils/metric.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/utils/metric.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"7f39687304604d5d"}},{"code_sha256_prefix":"c309b4c8e3254bb9","entry":"get_num_params_in_billions","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/utils/optim.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/utils/optim.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c309b4c8e3254bb9"}},{"code_sha256_prefix":"ee388693e9bdd485","entry":"isrotation","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/utils/geometry.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/utils/geometry.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ee388693e9bdd485"}},{"code_sha256_prefix":"1c57d6dfeacfc394","entry":"log_execution_time","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/utils/monitor.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/utils/monitor.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1c57d6dfeacfc394"}},{"code_sha256_prefix":"52614cdfc99cf150","entry":"lora_state_dict","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/model/lora.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/model/lora.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"52614cdfc99cf150"}},{"code_sha256_prefix":"8d01c3eb9b8242f3","entry":"main_rank_only","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/utils/decorator.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/utils/decorator.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8d01c3eb9b8242f3"}},{"code_sha256_prefix":"eed6f0c4289630b1","entry":"mat2euler","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/utils/geometry.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/utils/geometry.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"eed6f0c4289630b1"}},{"code_sha256_prefix":"901875b18f961eef","entry":"quat2mat","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/utils/geometry.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/utils/geometry.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"901875b18f961eef"}},{"code_sha256_prefix":"4e762613ee907cdb","entry":"repeat_kv","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/model/utils.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/model/utils.py","link_basis":"harvester_set","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4e762613ee907cdb"}},{"code_sha256_prefix":"cec833a99feb528c","entry":"rotate_half","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/model/utils.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/model/utils.py","link_basis":"harvester_set","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"cec833a99feb528c"}},{"code_sha256_prefix":"67a5b7072d402f6f","entry":"transpose","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/model/lora.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/model/lora.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"67a5b7072d402f6f"}},{"code_sha256_prefix":"354c67f169da36b9","entry":"apply_rotary_pos_emb","repo":"allenzren/open-pi-zero","repo_kind":"listed","path":"src/model/utils.py","file_url":"https://github.com/allenzren/open-pi-zero/blob/HEAD/src/model/utils.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":"354c67f169da36b9"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}