{"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/delight-very-deep-and-light-weight","title":"DeLighT: Deep and Light-weight Transformer","arxiv_id":"2008.00623","date":"2020-08-03","proceeding":"ICLR 2021 1","authors":["Sachin Mehta","Marjan Ghazvininejad","Srinivasan Iyer","Luke Zettlemoyer","Hannaneh Hajishirzi"],"abstract":"We introduce a deep and light-weight transformer, DeLighT, that delivers similar or better performance than standard transformer-based models with significantly fewer parameters. DeLighT more efficiently allocates parameters both (1) within each Transformer block using the DeLighT transformation, a deep and light-weight transformation, and (2) across blocks using block-wise scaling, which allows for shallower and narrower DeLighT blocks near the input and wider and deeper DeLighT blocks near the output. Overall, DeLighT networks are 2.5 to 4 times deeper than standard transformer models and yet have fewer parameters and operations. Experiments on benchmark machine translation and language modeling tasks show that DeLighT matches or improves the performance of baseline Transformers with 2 to 3 times fewer parameters on average. Our source code is available at: \\url{https://github.com/sacmehta/delight}","url_abs":"https://arxiv.org/abs/2008.00623v2","url_pdf":"https://arxiv.org/pdf/2008.00623v2.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":"delight-very-deep-and-light-weight","repo_url":"https://github.com/sacmehta/delight","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"delight-very-deep-and-light-weight","repo_url":"https://github.com/pranay185417/delight","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"translation","task_name":"Translation"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"adaptive-softmax","method_name":"Adaptive Softmax"},{"method_slug":"dextra","method_name":"DExTra"},{"method_slug":"delight","method_name":"DeLighT"},{"method_slug":"delight-block","method_name":"DeLighT Block"},{"method_slug":"feedforward-network","method_name":"Feedforward Network"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"}],"datasets_introduced":[],"methods_introduced":[{"slug":"dextra","name":"DExTra","full_name":"DExTra"},{"slug":"delight","name":"DeLighT","full_name":"DeLighT"},{"slug":"delight-block","name":"DeLighT Block","full_name":"DeLighT Block"}],"results":[{"leaderboard":"/sota/language-modelling-on-wikitext-103","task":"Language Modelling","dataset":"WikiText-103","model":"DeLighT","rank_in_archive_order":55,"of":89,"metrics":{"Number of params":"99M","Test perplexity":"24.14"},"uses_additional_data":false},{"leaderboard":"/sota/machine-translation-on-iwslt2014-german","task":"Machine Translation","dataset":"IWSLT2014 German-English","model":"DeLighT","rank_in_archive_order":23,"of":34,"metrics":{"BLEU score":"35.3"},"uses_additional_data":false},{"leaderboard":"/sota/machine-translation-on-wmt2016-english-french","task":"Machine Translation","dataset":"WMT2016 English-French","model":"DeLighT","rank_in_archive_order":1,"of":1,"metrics":{"BLEU score":"40.5"},"uses_additional_data":false},{"leaderboard":"/sota/machine-translation-on-wmt2016-english-german","task":"Machine Translation","dataset":"WMT2016 English-German","model":"DeLighT","rank_in_archive_order":4,"of":12,"metrics":{"BLEU score":"28.0"},"uses_additional_data":false},{"leaderboard":"/sota/machine-translation-on-wmt2016-english-1","task":"Machine Translation","dataset":"WMT2016 English-Romanian","model":"DeLighT","rank_in_archive_order":1,"of":21,"metrics":{"BLEU score":"34.7"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2008.00623","atlas_url":"https://app.syntology.ai/?focus=2008.00623","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2008.00623"}},"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/sacmehta/delight","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/pranay185417/delight","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":3},"by_repo_kind":{"official":{"samples":3,"ran":0,"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":"f24e9760137749a2","entry":"get_activation_layer","repo":"sacmehta/delight","repo_kind":"official","path":"fairseq/delight_modules/activation_layers.py","file_url":"https://github.com/sacmehta/delight/blob/HEAD/fairseq/delight_modules/activation_layers.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":"f24e9760137749a2"}},{"code_sha256_prefix":"2e619b36f4ddb228","entry":"get_embedding_layer","repo":"sacmehta/delight","repo_kind":"official","path":"fairseq/delight_modules/nn_functions.py","file_url":"https://github.com/sacmehta/delight/blob/HEAD/fairseq/delight_modules/nn_functions.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":"2e619b36f4ddb228"}},{"code_sha256_prefix":"15706edf9d0484d9","entry":"get_norm_layer","repo":"sacmehta/delight","repo_kind":"official","path":"fairseq/delight_modules/normalization_layers.py","file_url":"https://github.com/sacmehta/delight/blob/HEAD/fairseq/delight_modules/normalization_layers.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":"15706edf9d0484d9"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}