Methods › General › Learning Rate Schedules › Cosine Annealing

Cosine Annealing

3,965 papers tagged archive 2025-07-28

Introduced by Ilya Loshchilov et al. in SGDR: Stochastic Gradient Descent with Warm Restarts

archive 2025-07-28 Description, source and code snippet are the archive's method entry.

Cosine Annealing is a type of learning rate schedule that has the effect of starting with a large learning rate that is relatively rapidly decreased to a minimum value before being increased rapidly again. The resetting of the learning rate acts like a simulated restart of the learning process and the re-use of good weights as the starting point of the restart is referred to as a "warm restart" in contrast to a "cold restart" where a new set of small random numbers may be used as a starting point.

ηₜ = ηₘᵢₙⁱ + 1/2(ηₘₐₓⁱ-ηₘᵢₙⁱ)(1+cos((T_(cur))/Tᵢπ))

Where where ηₘᵢₙⁱ and ηₘₐₓⁱ are ranges for the learning rate, and T_(cur) account for how many epochs have been performed since the last restart.

Text Source: Jason Brownlee

Image Source: Gao Huang

PaperSource

In Syntology name not resolved (unknown_name). The lookup matched no method, paper or venue by this name; that asserts nothing about whether code exists.

Near matches the lookup lists, each a different name: COSINE; COSINE; p-annealing; CWGD-Cosine; Annealing Flow; data annealing; annealing approach; Logit Annealing Loss.

Syntology's have check for this name (free, it answers coverage only), read 2026-10-05; ask it again live. A level is the served verification level of that one answer; nothing here is summed across methods.

Papers archive 2025-07-28

archive papers tagged: 3,965 · with a code link: 1,734 · where Syntology ran a sample: 627 (495 with a run with no instrument failure, 132 where every run was a failure of Syntology's instrument) Syntology

Show: all tagged papersonly where code ran (627 of 3,965 tagged: 495 with a run with no instrument failure, 132 where every run was a failure of Syntology's instrument)

Page 1 of 40: papers 1 to 100 of 3,965, newest first by the archive's date (ties by slug), in archive order.

Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code, as “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the instrument figure counts failures of Syntology's instrument, not of the code. It is per sample and not a correctness claim. When the archive marks a repository official for the paper, the line starts with that repository's state (the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from.

Tasks archive 2025-07-28

20 shown of 1,053 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.

TaskPapers
Language Modelling800
Language Modeling591
Question Answering303
Large Language Model298
Text Generation285
Retrieval203
Sentence168
In-Context Learning162
Prompt Engineering133
Few-Shot Learning111
Decoder109
Code Generation106
Decision Making102
Text Classification91
RAG90
Translation89
GPU88
Retrieval-augmented Generation88
Transfer Learning88
Math87

Usage over time archive 2025-07-28

Papers per year tagged with Cosine Annealing: 2016 to 2025, peak 1,385 1,385 0 2016: 2 papers 2016 2017: 0 papers 2017 2018: 9 papers 2018 2019: 95 papers 2019 2020: 188 papers 2020 2021: 308 papers 2021 2022: 412 papers 2022 2023: 1196 papers 2023 2024: 1385 papers 2024 2025: 370 papers 2025
Papers per year the archive tags with this method, by the paper's archive date (3,965 dated). Bars are counts, not a trend claim.

Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).

Categories archive 2025-07-28

Learning Rate Schedules

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections