{"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/arxiv-2606-07345","title":"TabSwift: An Efficient Tabular Foundation Model with Row-Wise Attention","arxiv_id":"2606.07345","date":"2026-06-05","proceeding":"ICML","authors":["Si-Yang Liu","Han-Jia Ye"],"abstract":"Tabular foundation models, exemplified by TabPFN, perform prediction via in-context learning, inferring test labels directly from labeled training examples. They have demonstrated competitive performance, particularly on small-to-medium datasets. However, recent tabular foundation models often improve accuracy with increasingly complex architectures, incurring higher inference cost and limiting practical deployment. In this work, we revisit the original TabPFN design and show that a lightweight row-wise attention-only backbone can remain highly competitive with two simple enhancements: a gated attention stabilization mechanism and a small set of learnable register tokens that provide global context and improve pretraining quality. The resulting model, TabSwift, supports both classification and regression, and is competitive with stronger tabular foundation models (e.g., TabPFN v2 and TabICL) while being more efficient at inference. For latency-sensitive serving, we further introduce an adaptive layer-wise early-exit mechanism that dynamically adjusts inference depth per sample. Overall, TabSwift enables efficient and anytime tabular in-context learning for practical deployments.","url_abs":"https://arxiv.org/abs/2606.07345","url_pdf":"https://arxiv.org/pdf/2606.07345","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2606.07345","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2606.07345"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+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":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/LAMDA-Tabular/TabSwift","reach":null}],"summary":{"ran":4,"unverified":2},"by_repo_kind":{"found_in_text":{"samples":6,"ran":4,"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":"c6990cbd011cf0ff","entry":"check_softmax","repo":"LAMDA-Tabular/TabSwift","repo_kind":"found_in_text","path":"TALENT/model/methods/tabswift.py","file_url":"https://github.com/LAMDA-Tabular/TabSwift/blob/HEAD/TALENT/model/methods/tabswift.py","link_basis":"first_harvest_node","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":"c6990cbd011cf0ff"}},{"code_sha256_prefix":"a9dd34f2bf5c94fd","entry":"multi_head_attention_forward","repo":"LAMDA-Tabular/TabSwift","repo_kind":"found_in_text","path":"TALENT/model/lib/tabswift/model/attention.py","file_url":"https://github.com/LAMDA-Tabular/TabSwift/blob/HEAD/TALENT/model/lib/tabswift/model/attention.py","link_basis":"first_harvest_node","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":"a9dd34f2bf5c94fd"}},{"code_sha256_prefix":"db2fe357a8af4965","entry":"rmse","repo":"LAMDA-Tabular/TabSwift","repo_kind":"found_in_text","path":"TALENT/model/utils.py","file_url":"https://github.com/LAMDA-Tabular/TabSwift/blob/HEAD/TALENT/model/utils.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":"db2fe357a8af4965"}},{"code_sha256_prefix":"fa8166d67efadbad","entry":"sdpa_with_flattened_batch","repo":"LAMDA-Tabular/TabSwift","repo_kind":"found_in_text","path":"TALENT/model/lib/tabswift/model/attention.py","file_url":"https://github.com/LAMDA-Tabular/TabSwift/blob/HEAD/TALENT/model/lib/tabswift/model/attention.py","link_basis":"plan_row","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":"fa8166d67efadbad"}},{"code_sha256_prefix":"e248737ecd5c07ad","entry":"load_config","repo":"LAMDA-Tabular/TabSwift","repo_kind":"found_in_text","path":"TALENT/model/utils.py","file_url":"https://github.com/LAMDA-Tabular/TabSwift/blob/HEAD/TALENT/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":"e248737ecd5c07ad"}},{"code_sha256_prefix":"a1e5bc5092674628","entry":"sample_parameters","repo":"LAMDA-Tabular/TabSwift","repo_kind":"found_in_text","path":"TALENT/model/utils.py","file_url":"https://github.com/LAMDA-Tabular/TabSwift/blob/HEAD/TALENT/model/utils.py","link_basis":"first_harvest_node","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":"a1e5bc5092674628"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}