{"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/single-headed-attention-rnn-stop-thinking","title":"Single Headed Attention RNN: Stop Thinking With Your Head","arxiv_id":"1911.11423","date":"2019-11-26","proceeding":null,"authors":["Stephen Merity"],"abstract":"The leading approaches in language modeling are all obsessed with TV shows of my youth - namely Transformers and Sesame Street. Transformers this, Transformers that, and over here a bonfire worth of GPU-TPU-neuromorphic wafer scale silicon. We opt for the lazy path of old and proven techniques with a fancy crypto inspired acronym: the Single Headed Attention RNN (SHA-RNN). The author's lone goal is to show that the entire field might have evolved a different direction if we had instead been obsessed with a slightly different acronym and slightly different result. We take a previously strong language model based only on boring LSTMs and get it to within a stone's throw of a stone's throw of state-of-the-art byte level language model results on enwik8. This work has undergone no intensive hyperparameter optimization and lived entirely on a commodity desktop machine that made the author's small studio apartment far too warm in the midst of a San Franciscan summer. The final results are achievable in plus or minus 24 hours on a single GPU as the author is impatient. The attention mechanism is also readily extended to large contexts with minimal computation. Take that Sesame Street.","url_abs":"https://arxiv.org/abs/1911.11423v2","url_pdf":"https://arxiv.org/pdf/1911.11423v2.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":"single-headed-attention-rnn-stop-thinking","repo_url":"https://github.com/Smerity/sha-rnn","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"single-headed-attention-rnn-stop-thinking","repo_url":"https://github.com/Tobias-K93/media-bias-prediction","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"single-headed-attention-rnn-stop-thinking","repo_url":"https://github.com/alisafaya/SHA-RNN.jl","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"single-headed-attention-rnn-stop-thinking","repo_url":"https://github.com/floleuerer/fastai_ulmfit","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"single-headed-attention-rnn-stop-thinking","repo_url":"https://github.com/saattrupdan/scholarly","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":null,"task_name":"GPU"},{"task_slug":"hyperparameter-optimization","task_name":"Hyperparameter Optimization"},{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"boom-layer","method_name":"Boom Layer"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"embedding-dropout","method_name":"Embedding Dropout"},{"method_slug":"lamb","method_name":"LAMB"},{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"sha-rnn","method_name":"SHA-RNN"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"single-headed-attention","method_name":"Single-Headed Attention"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"},{"method_slug":"wordpiece","method_name":"WordPiece"}],"datasets_introduced":[],"methods_introduced":[{"slug":"sha-rnn","name":"SHA-RNN","full_name":"Single Headed Attention RNN"},{"slug":"single-headed-attention","name":"Single-Headed Attention","full_name":"Single-Headed Attention"}],"results":[{"leaderboard":"/sota/language-modelling-on-enwiki8","task":"Language Modelling","dataset":"enwik8","model":"SHA-RNN (4 layers, h=1024, attention head per layer)","rank_in_archive_order":27,"of":42,"metrics":{"Bit per Character (BPC)":"1.068","Number of params":"54M"},"uses_additional_data":false},{"leaderboard":"/sota/language-modelling-on-enwiki8","task":"Language Modelling","dataset":"enwik8","model":"SHA-RNN (4 layers, h=1024, single attention head)","rank_in_archive_order":28,"of":42,"metrics":{"Bit per Character (BPC)":"1.076","Number of params":"52M"},"uses_additional_data":false},{"leaderboard":"/sota/language-modelling-on-enwiki8","task":"Language Modelling","dataset":"enwik8","model":"SHA-LSTM (4 layers, h=1024, no attention head)","rank_in_archive_order":39,"of":42,"metrics":{"Bit per Character (BPC)":"1.33","Number of params":"51M"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1911.11423","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}