Menu

View

JournalDigital

Theme

LightDarkSystem

Text size

AAA

The paper in your inbox every morning

Straight from the source · Glossary

The Tech Gazette

Quantization

Storing a model's weights with fewer bits (8 or 4 instead of 16) so it needs less memory and runs faster, at some loss of quality.

Read more at en.wikipedia.org →

Learn it

1. AWS Machine LearningOct 13, 2026Technical Guides
   Mixture-of-Experts kernel engineering on Amazon SageMaker AI
2. BasetenOct 9, 2026Technical Guides
   What makes a high-quality NVFP4 quantization?
3. AMD ROCmSep 25, 2026Technical Guides
   Local Quantization and Multi-Backend Deployment with AMD Quark on Strix Halo

In the news

1. ACM TACO (Architecture and Code Optimization)Oct 15, 2026Research
   LOAD: Enabling Low-Bit Group Quantization for Vision Transformers via Adaptive Data Formats
2. arXiv cs.LGOct 14, 2026Research
   Beyond Expert Rankings: Contrast Calibration for Mixed-Precision MoE Quantization
3. ACM TACO (Architecture and Code Optimization)Oct 14, 2026Research
   Q-GMRES: Quantized Data Transmission Optimization for Distributed GMRES Algorithm
4. arXiv (systems & security) cs.CROct 14, 2026Research
   What You Sign Is Not What You Run: The Verifiability Gap in Quantized LLM Supply Chains
5. arXiv cs.LGOct 13, 2026Research
   MultiAlloc: Budget-Aware Precision Allocation for KV Cache Quantization
6. arXiv cs.LGOct 9, 2026Research
   DualQuant: Dual-scale Calibration-free Weight-and-Activation Quantization for LLMs
7. QdrantOct 9, 2026Releases
   Qdrant 1.19.0: #9669, #9684, #9950 - Unify definition of memory usage strategy for collection components. Use "memory"…
8. arXiv (systems & security) cs.CROct 9, 2026Research
   Quantization-Tolerant Gradient Auditing for Active Inference-Attack Detection in Federated Learning
9. ONNX RuntimeOct 9, 2026Releases
   ONNX Runtime v1.31.0
10. arXiv cs.LGOct 8, 2026Research
    Rounding in Preconditioner Space: Redesigning 4-bit AdamW Optimizer-State Quantization
11. arXiv cs.LGOct 7, 2026Research
    ResidualQuant: KV Cache Quantization for Looped Transformers with 2-Bit Residuals
12. arXiv cs.LGOct 7, 2026Research
    Layerwise Error Attribution for Fast and Robust Mixed-Precision Post-Training Quantization
13. arXiv cs.LGOct 7, 2026Research
    Dual-QK: Sharp Queries and Flat Keys for Prunable 2-bit KV Caches
14. WeaviateOct 7, 2026Releases
    v1.40.0 - Alter Schema Drop Vector Index, HFresh MUVERA, RQ-4 quantization
15. llama.cppOct 6, 2026Releases
    b11446
16. llama.cppOct 2, 2026Releases
    b11331
17. arXiv (systems & security) cs.CROct 1, 2026Research
    MOMAT: Mixture of Multiple Atlases for Low-Power Jailbreak Defense of Quantized LLMs
18. arXiv (vision, retrieval & HCI) cs.CVSep 30, 2026Research
    Multimodal Flow: Unified Flow Modeling of Language and Vision in Embedding Spaces
19. llama.cppSep 30, 2026Releases
    b11284
20. arXiv (systems & security) cs.CRSep 29, 2026Research
    Security-Enhanced Seed-Based Weight Quantization for Large Language Models