Methods › General › Large Batch Optimization › 1-bit Adam
1-bit Adam
Introduced by Hanlin Tang et al. in 1-bit Adam: Communication Efficient Large-Scale Training with Adam's Convergence Speed
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
1-bit Adam is a stochastic optimization technique that is a variant of ADAM with error-compensated 1-bit compression, based on finding that Adam's variance term becomes stable at an early stage. First vanilla Adam is used for a few epochs as a warm-up. After the warm-up stage, the compression stage starts and we stop updating the variance term 𝐯 and use it as a fixed precondition. At the compression stage, we communicate based on the momentum applied with error-compensated 1-bit compression. The momentums are quantized into 1-bit representation (the sign of each element). Accompanying the vector, a scaling factor is computed as (magnitude of compensated gradient)/(magnitude of quantized gradient). This scaling factor ensures that the compressed momentum has the same magnitude as the uncompressed momentum. This 1-bit compression could reduce the communication cost by 97 % and 94 % compared to the original float 32 and float 16 training, respectively.
Papers archive 2025-07-28
30 shown of 40, newest first. Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code; it is per sample and not a correctness claim.
-
ARWI: Arabic Write and Improve 16 Apr 2025 · 0 repositories · arXiv:2504.11814
-
DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning 22 Jan 2025 · 4 repositories · arXiv:2501.12948
-
ShowUI: One Vision-Language-Action Model for GUI Visual Agent 26 Nov 2024 · 1 repository · arXiv:2411.17465Syntology ran 1 of 16 samples · 15 unverified
-
MiniCPM-V: A GPT-4V Level MLLM on Your Phone 3 Aug 2024 · 2 repositories · arXiv:2408.01800Syntology ran 9 of 14 samples · 5 unverified
-
YOLOv10: Real-Time End-to-End Object Detection 23 May 2024 · 3 repositories · arXiv:2405.14458Syntology ran 0 of 3 samples · 3 unverified
-
Fast Fishing: Approximating BAIT for Efficient and Scalable Deep Active Image Classification 13 Apr 2024 · 1 repository · arXiv:2404.08981
-
Tracking Anything in High Quality 26 Jul 2023 · 1 repository · arXiv:2307.13974
-
DINOv2: Learning Robust Visual Features without Supervision 14 Apr 2023 · 26 repositories · arXiv:2304.07193Syntology ran 21 of 46 samples · 25 unverified · 12 pointer-only (licence)
-
ChatDoctor: A Medical Chat Model Fine-Tuned on a Large Language Model Meta-AI (LLaMA) Using Medical Domain Knowledge 24 Mar 2023 · 1 repository · arXiv:2303.14070
-
CCTV-Gun: Benchmarking Handgun Detection in CCTV Images 19 Mar 2023 · 1 repository · arXiv:2303.10703
-
MusicLM: Generating Music From Text 26 Jan 2023 · 5 repositories · arXiv:2301.11325Syntology ran 6 of 16 samples · 10 unverified
-
Dynamic Gradient Reactivation for Backward Compatible Person Re-identification 12 Jul 2022 · 0 repositories · arXiv:2207.05658
-
Solving Quantitative Reasoning Problems with Language Models 29 Jun 2022 · 1 repository · arXiv:2206.14858
-
Maximizing Communication Efficiency for Large-scale Training via 0/1 Adam 12 Feb 2022 · 1 repository · arXiv:2202.06009
-
Fer2013 Recognition - ResNet18 With Tricks 29 Dec 2021 · 1 repository
-
NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion 24 Nov 2021 · 1 repository · arXiv:2111.12417Syntology ran 3 of 3 samples · 0 unverified · 2 pointer-only (licence)
-
1-bit LAMB: Communication Efficient Large-Scale Large-Batch Training with LAMB's Convergence Speed 13 Apr 2021 · 1 repository · arXiv:2104.06069
-
LayoutParser: A Unified Toolkit for Deep Learning Based Document Image Analysis 29 Mar 2021 · 6 repositories · arXiv:2103.15348
-
What is it Like to Be a Bot: Simulated, Situated, Structurally Coherent Qualia (S3Q) Theory of Consciousness 13 Mar 2021 · 0 repositories · arXiv:2103.12638
-
Zero-Shot Text-to-Image Generation 24 Feb 2021 · 12 repositories · arXiv:2102.12092Syntology ran 7 of 7 samples · 0 unverified · 3 pointer-only (licence)
-
1-bit Adam: Communication Efficient Large-Scale Training with Adam's Convergence Speed 4 Feb 2021 · 2 repositories · arXiv:2102.02888
-
Infinite Nature: Perpetual View Generation of Natural Scenes from a Single Image 17 Dec 2020 · 1 repository · arXiv:2012.09855
-
In silico comparison of spike protein-ACE2 binding affinities across species; significance for the possible origin of the SARS-CoV-2 virus 13 May 2020 · 0 repositories · arXiv:2005.06199
-
MTI-Net: Multi-Scale Task Interaction Networks for Multi-Task Learning 19 Jan 2020 · 1 repository · arXiv:2001.06902
-
Advances and Open Problems in Federated Learning 10 Dec 2019 · 9 repositories · arXiv:1912.04977Syntology ran 0 of 9 samples · 9 unverified
-
RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds 25 Nov 2019 · 9 repositories · arXiv:1911.11236Syntology ran 1 of 4 samples · 3 unverified
-
Green AI 22 Jul 2019 · 2 repositories · arXiv:1907.10597Syntology ran 0 of 1 samples · 1 unverified
-
Self-Supervised Representation Learning by Rotation Feature Decoupling 1 Jun 2019 · 1 repository
-
A Comprehensive Overhaul of Feature Distillation 3 Apr 2019 · 2 repositories · arXiv:1904.01866Syntology ran 1 of 6 samples · 5 unverified
-
Analysis of the Generalization Error: Empirical Risk Minimization over Deep Artificial Neural Networks Overcomes the Curse of Dimensionality in the Numerical Approximation of Black-Scholes Partial Differential Equations 9 Sep 2018 · 0 repositories · arXiv:1809.03062
Tasks archive 2025-07-28
20 shown of 96 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