Papers › Few-Bit Backward: Quantized Gradients of Activation Functions for Memory Footprint Reduction

Few-Bit Backward: Quantized Gradients of Activation Functions for Memory Footprint Reduction

1 Feb 2022arXiv:2202.00441archive 2025-07-28

Georgii Novikov, Daniel Bershatsky, Julia Gusak, Alex Shonenkov, Denis Dimitrov, Ivan Oseledets

Memory footprint is one of the main limiting factors for large neural network training. In backpropagation, one needs to store the input to each operation in the computational graph. Every modern neural network model has quite a few pointwise nonlinearities in its architecture, and such operation induces additional memory costs which -- as we show -- can be significantly reduced by quantization of the gradients. We propose a systematic approach to compute optimal quantization of the retained gradients of the pointwise nonlinear functions with only a few bits per each element. We show that such approximation can be achieved by computing optimal piecewise-constant approximation of the derivative of the activation function, which can be done by dynamic programming. The drop-in replacements are implemented for all popular nonlinearities and can be used in any existing pipeline. We confirm the memory reduction and the same convergence on several open benchmarks.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 17 code samples harvested from 1 repository linked to this paper: 15 ran (of which 0 constructed an object rather than computing a result; 15 with no instrument failure: 0 honoured, 0 violated, 15 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 17 samples from 1 repository, 15 ran (15 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written on 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

SkoltechAI/fewbit officialmentioned in papermentioned on GitHubpytorchBSD-3-ClauseSyntology: 17 samples, 15 ran (0 instrument-failed) · 2 unverified report
daskol/fewbit mentioned on GitHubpytorchBSD-3-ClauseSyntology: not harvested report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

17 samples harvested: 15 ran (of which 0 constructed an object rather than computing a result; 15 with no instrument failure: 0 honoured, 0 violated, 15 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

15ran
2unverified

Licence: 0 of the 17 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from SkoltechAI/fewbit. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: 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. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

SkoltechAI/fewbit official repository 17 samples, 15 ran (0 instrument-failed) · 2 unverified
append_reference fewbit/modules/activations.py ran fingerprinted BSD-3-Clause (permissive) · ffdba334e16081d0 · report fixture: signature doc:str/0/str, name:str/0/str · returned str · level recorded 2026-09-24
approximate fewbit/approx.py ran BSD-3-Clause (permissive) · 1f6d08cef9a84d45 · report fixture: signature fn:callable/0/?, fn_prim:callable/0/?, cardinality:int/0/int, domain:tuple/1/float, parity:bool/0/bool, max_iters:int/0/int, beps:float/0/float, leps:float/0/float, random_state:int/0/int · returned tuple · level recorded 2026-09-24
clamp in 78 papers
census: 7 implementations from 8 papers, 6 distinct outputs across 4 buckets, one shared input per bucket
fewbit/functional/linear.py ran fingerprinted BSD-3-Clause (permissive) · c8f42186142a2a86 · report fixture: signature val:int/0/int, minval:int/0/int, maxval:int/0/int · returned int · level recorded 2026-09-24
dct in 24 papers
census: 6 implementations from 8 papers, 4 distinct outputs on one shared input
fewbit/fft.py ran BSD-3-Clause (permissive) · 0984394a69efa89b · report fixture: signature x:Tensor/3/float, type:int/0/int, n:NoneType/0/?, dim:int/0/int, norm:str/0/str · returned Tensor · level recorded 2026-09-24
dispatch fewbit/functional/activations.py ran BSD-3-Clause (permissive) · 0490b30beebe8a38 · report fixture: signature name:str/0/str, wrapper:callable/0/? · returned function · level recorded 2026-09-24
estimate_correlation fewbit/modules/variance.py ran fingerprinted BSD-3-Clause (permissive) · 3ae5fdf58952210a · report fixture: signature input:Tensor/2/float, output:Tensor/2/float · returned Tensor · level recorded 2026-09-24
estimate_memory_usage fewbit/util.py ran BSD-3-Clause (permissive) · 11e69fa774fc1892 · report fixture: signature variable:Tensor/2/float, saved_only:bool/0/bool · returned int · level recorded 2026-09-24
estimate_variance_rmm fewbit/modules/variance.py ran fingerprinted BSD-3-Clause (permissive) · 6fc8638c18f0ce42 · report fixture: signature input:Tensor/2/float, output:Tensor/2/float, bs_proj:int/0/int · returned Tensor · level recorded 2026-09-24
estimate_variance_sgd fewbit/modules/variance.py ran fingerprinted BSD-3-Clause (permissive) · 68f0decf99978fcb · report fixture: signature input:Tensor/2/float, output:Tensor/2/float, bs:NoneType/0/? · returned Tensor · level recorded 2026-09-24
idct in 8 papers
census: 5 implementations from 5 papers, 4 distinct outputs across 2 buckets, one shared input per bucket
fewbit/fft.py ran BSD-3-Clause (permissive) · 227e256653d49044 · report fixture: signature x:Tensor/2/float, type:int/0/int, n:NoneType/0/?, dim:int/0/int, norm:str/0/str · returned Tensor · level recorded 2026-09-24
insert_args fewbit/modules/activations.py ran fingerprinted BSD-3-Clause (permissive) · 920472b6219d42b5 · report fixture: signature doc:str/0/str · returned str · level recorded 2026-09-24
map_module fewbit/util.py ran BSD-3-Clause (permissive) · 2adbb1ba9627b56f · report fixture: signature root:callable/0/?, func:callable/0/?, patt:str/0/str · returned SimpleModel · level recorded 2026-09-24
stepwise fewbit/approx.py ran BSD-3-Clause (permissive) · b43ba240a309b25f · report fixture: signature xs:ndarray/1/float, ys:ndarray/1/float · returned StepWiseFunction · level recorded 2026-09-24
teniter fewbit/util.py ran BSD-3-Clause (permissive) · 9a1cb09a2c15016b · report fixture: signature variable:Tensor/0/float, include_ordinary:bool/0/bool, include_saved:bool/0/bool · returned generator · level recorded 2026-09-24
update_args in 7 papers fewbit/modules/activations.py ran fingerprinted BSD-3-Clause (permissive) · e093939e7df95a04 · report fixture: signature doc:str/0/str, beg:int/0/int, prefix:str/0/str · returned str · level recorded 2026-09-24
estimate_error fewbit/approx.py unverified BSD-3-Clause (permissive) · a5c843a81567e736 · report
load_func fewbit/functional/activations.py unverified BSD-3-Clause (permissive) · 5d477abefd44b2d0 · report

Tasks

Neural Network CompressionQuantization

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

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