Methods › General › Sharded Data Parallel Methods › ZeRO
ZeRO
Introduced by Samyam Rajbhandari et al. in ZeRO: Memory Optimizations Toward Training Trillion Parameter Models
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
Zero Redundancy Optimizer (ZeRO) is a sharded data parallel method for distributed training. ZeRODP removes the memory state redundancies across data-parallel processes by partitioning the model states instead of replicating them, and it retains the compute/communication efficiency by retaining the computational granularity and communication volume of DP using a dynamic communication schedule during training.
In Syntology about another paper, arXiv:2405.18330: repository known, no verified implementation (declared_code; repositories: 1 listed of 1 in the graph, strongest link: official). The lookup matched 'ZERO' ignoring case. About another paper: this answer is about arXiv:2405.18330, while the archive credits arXiv:1910.02054.
From the lookup's note on this name: The same name also belongs to the paper “ZeRO: Memory Optimizations Toward Training Trillion Parameter Models” (arXiv:1910.02054, 2,002 citations), which the lookup finds cited far more than the paper this answer is about (41 citations), so this answer may be about a different sense of the name. For arXiv:1910.02054: its page on this site · ask the lookup about it.
Code Syntology holds under other names, as the lookup lists them (code_nearby); not a complete list, and none of them is this method.
- Zero-Shot Tokenizer Transfer (ZeTT) · shares_name_tokens (shares name tokens with this name; a different method) · level V2 · retrieve with
get_reference_implementation(name='Zero-Shot Tokenizer Transfer (ZeTT)') - Simple Zero-shot Prompt Weighting Technique · shares_name_tokens (shares name tokens with this name; a different method) · level V3 · retrieve with
get_reference_implementation(name='Simple Zero-shot Prompt Weighting Technique')
Origin paper of this answer, arXiv:2405.18330 (its page on this site · its page on Syntology): Syntology's run record for its harvested code, 4 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 3 where Syntology's instrument failed) · 4 unverified (of 8 harvested samples) · 4 pointer-only (licence). The instrument figure counts failures of Syntology's instrument, not of the code. Ran means executed on a synthesized input, not that the code is correct.
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.
What each level means, quoted from Syntology's server card: V0 “imports”; V1 “executes against a synthesized fixture and returns a value of the recorded type”; V2 “two implementations GENERATED from the paper's signature and spec agree numerically (1e-6) on fixture inputs” “neither of the two ever read the authors' own code, so V2 measures how unambiguously the paper specifies the method, not fidelity to the authors' implementation (measured 2026-09-11, commit 75eaa77: cross_checked_by is a second LLM on 2,628 of 2,831 served samples)”; V3 “paper-derived property tests pass on both of those same two generations.” The card adds: “No rung on this ladder compares the served code against the authors' implementation, so treat every level as evidence about runnability and internal agreement, not about matching the paper's results.” A level quoted in two parts is one definition the card writes as two clauses joined by a double hyphen, each clause quoted on its own (how these levels are quoted).
Papers archive 2025-07-28
archive papers tagged: 9 · with a code link: 3 · where Syntology ran a sample: 2 (2 with a run with no instrument failure, 0 where every run was a failure of Syntology's instrument) Syntology
Show: all tagged papersonly where code ran (2 of 9 tagged: 2 with a run with no instrument failure, 0 where every run was a failure of Syntology's instrument)
Page 1 of 1: papers 1 to 9 of 9, 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.
-
Memory Analysis on the Training Course of DeepSeek Models 11 Feb 2025 · 0 repositories · arXiv:2502.07846
-
Accelerating Large Language Model Training with Hybrid GPU-based Compression 4 Sep 2024 · 0 repositories · arXiv:2409.02423
-
A Study of Optimizations for Fine-tuning Large Language Models 4 Jun 2024 · 0 repositories · arXiv:2406.02290
-
Zero redundancy distributed learning with differential privacy 20 Nov 2023 · 0 repositories · arXiv:2311.11822
-
Dissecting the Runtime Performance of the Training, Fine-tuning, and Inference of Large Language Models 7 Nov 2023 · 0 repositories · arXiv:2311.03687
-
ReMax: A Simple, Effective, and Efficient Reinforcement Learning Method for Aligning Large Language Models 16 Oct 2023 · 3 repositories · arXiv:2310.10505Syntology official (archive's flag): 11 ran · 11 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 1 honoured, 0 violated, 8 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified (of 13 harvested samples) · 13 pointer-only (licence)
-
Rethinking Memory and Communication Cost for Efficient Large Language Model Training 9 Oct 2023 · 0 repositories · arXiv:2310.06003
-
ZeRO++: Extremely Efficient Collective Communication for Giant Model Training 16 Jun 2023 · 1 repository · arXiv:2306.10209
-
ZeRO: Memory Optimizations Toward Training Trillion Parameter Models 4 Oct 2019 · 10 repositories · arXiv:1910.02054Syntology community repositories only · 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified (of 1 harvested sample)
Tasks archive 2025-07-28
11 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.
Usage over time archive 2025-07-28
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
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