{"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/adafactor-adaptive-learning-rates-with","title":"Adafactor: Adaptive Learning Rates with Sublinear Memory Cost","arxiv_id":"1804.04235","date":"2018-04-11","proceeding":"ICML 2018 7","authors":["Noam Shazeer","Mitchell Stern"],"abstract":"In several recently proposed stochastic optimization methods (e.g. RMSProp,\nAdam, Adadelta), parameter updates are scaled by the inverse square roots of\nexponential moving averages of squared past gradients. Maintaining these\nper-parameter second-moment estimators requires memory equal to the number of\nparameters. For the case of neural network weight matrices, we propose\nmaintaining only the per-row and per-column sums of these moving averages, and\nestimating the per-parameter second moments based on these sums. We demonstrate\nempirically that this method produces similar results to the baseline.\nSecondly, we show that adaptive methods can produce larger-than-desired updates\nwhen the decay rate of the second moment accumulator is too slow. We propose\nupdate clipping and a gradually increasing decay rate scheme as remedies.\nCombining these methods and dropping momentum, we achieve comparable results to\nthe published Adam regime in training the Transformer model on the WMT 2014\nEnglish-German machine translation task, while using very little auxiliary\nstorage in the optimizer. Finally, we propose scaling the parameter updates\nbased on the scale of the parameters themselves.","url_abs":"http://arxiv.org/abs/1804.04235v1","url_pdf":"http://arxiv.org/pdf/1804.04235v1.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":"adafactor-adaptive-learning-rates-with","repo_url":"https://github.com/DeadAt0m/adafactor-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"adafactor-adaptive-learning-rates-with","repo_url":"https://github.com/arampacha/clip-rsicd","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"jax","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"adafactor-adaptive-learning-rates-with","repo_url":"https://github.com/booydar/t5-experiments","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"adafactor-adaptive-learning-rates-with","repo_url":"https://github.com/huggingface/pytorch-image-models","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"adafactor-adaptive-learning-rates-with","repo_url":"https://github.com/yurakuratov/t5-experiments","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"stochastic-optimization","task_name":"Stochastic Optimization"},{"task_slug":"translation","task_name":"Translation"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adafactor","method_name":"Adafactor"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"rmsprop","method_name":"RMSProp"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[{"slug":"adafactor","name":"Adafactor","full_name":"Adafactor"}],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.04235","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1804.04235"}},"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":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/huggingface/pytorch-image-models","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/DeadAt0m/adafactor-pytorch","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/booydar/t5-experiments","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/yurakuratov/t5-experiments","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/arampacha/clip-rsicd","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"unverified":4},"by_repo_kind":{"listed":{"samples":4,"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":"f9a40716dffb3689","entry":"create_learning_rate_fn","repo":"arampacha/clip-rsicd","repo_kind":"listed","path":"run_clip_flax.py","file_url":"https://github.com/arampacha/clip-rsicd/blob/HEAD/run_clip_flax.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"f9a40716dffb3689"}},{"code_sha256_prefix":"49ef7ff26b7ca326","entry":"get_model_basename","repo":"arampacha/clip-rsicd","repo_kind":"listed","path":"run_clip_eval.py","file_url":"https://github.com/arampacha/clip-rsicd/blob/HEAD/run_clip_eval.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"49ef7ff26b7ca326"}},{"code_sha256_prefix":"0219f41ee0de195f","entry":"make_batch","repo":"arampacha/clip-rsicd","repo_kind":"listed","path":"run_clip_flax_tv.py","file_url":"https://github.com/arampacha/clip-rsicd/blob/HEAD/run_clip_flax_tv.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"0219f41ee0de195f"}},{"code_sha256_prefix":"0e30fafffd7ed8cd","entry":"mb_item","repo":"arampacha/clip-rsicd","repo_kind":"listed","path":"run_clip_flax.py","file_url":"https://github.com/arampacha/clip-rsicd/blob/HEAD/run_clip_flax.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"0e30fafffd7ed8cd"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}