Imagen de Portada

Float8@2bits: Entropy Coding Enables Data-Free Model Compression

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Putzky, Patrick, Genzel, Martin, Mollenhauer, Mattes, Schulze, Sebastian, Wollmann, Thomas, Dietzel, Stefan
Formato:	Preprint
Publicado:	2026
Materias:	Machine Learning
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Choose Your Model Size: Any Compression of Large Language Models Without Re-Computation
por: Genzel, Martin, et al.
Publicado: (2025)

Improving Block-Wise LLM Quantization by 4-bit Block-Wise Optimal Float (BOF4): Analysis and Variations
por: Blumenberg, Patrick, et al.
Publicado: (2025)

Towards Federated Learning with On-device Training and Communication in 8-bit Floating Point
por: Wang, Bokun, et al.
Publicado: (2024)

Towards Optimal Sobolev Norm Rates for the Vector-Valued Regularized Least-Squares Algorithm
por: Li, Zhu, et al.
Publicado: (2023)

Memorization With Neural Nets: Going Beyond the Worst Case
por: Dirksen, Sjoerd, et al.
Publicado: (2023)

The Hidden Power of Pure 16-bit Floating-Point Neural Networks
por: Yun, Juyoung, et al.
Publicado: (2023)

Golomb–Rice Coding for Efficient Compression of 8-bit Floating-Point Data
por: Vucic, Nikola, et al.
Publicado: (2025)

Optimal Rates for Vector-Valued Spectral Regularization Learning Algorithms
por: Meunier, Dimitri, et al.
Publicado: (2024)

One-bit Compressed Sensing using Generative Models
por: Kafle, Swatantra, et al.
Publicado: (2025)

decoupleQ: Towards 2-bit Post-Training Uniform Quantization via decoupling Parameters into Integer and Floating Points
por: Guo, Yi, et al.
Publicado: (2024)

Accuracy Booster: Enabling 4-bit Fixed-point Arithmetic for DNN Training
por: Harma, Simla Burcu, et al.
Publicado: (2022)

Asymptotic e-processes
por: Massiani, Pierre-François, et al.
Publicado: (2026)

CCQ: Convolutional Code for Extreme Low-bit Quantization in LLMs
por: Zhou, Zhaojing, et al.
Publicado: (2025)

Learning linear operators: Infinite-dimensional regression as a well-behaved non-compact inverse problem
por: Mollenhauer, Mattes, et al.
Publicado: (2022)

BitNet a4.8: 4-bit Activations for 1-bit LLMs
por: Wang, Hongyu, et al.
Publicado: (2024)

The AetherFloat Family: Block-Scale-Free Quad-Radix Floating-Point Architectures for AI Accelerators
por: Morisaki, Keita
Publicado: (2026)

Entropy Law: The Story Behind Data Compression and LLM Performance
por: Yin, Mingjia, et al.
Publicado: (2024)

ICQuant: Index Coding enables Low-bit LLM Quantization
por: Li, Xinlin, et al.
Publicado: (2025)

Regularized least squares learning with heavy-tailed noise is minimax optimal
por: Mollenhauer, Mattes, et al.
Publicado: (2025)

KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache
por: Liu, Zirui, et al.
Publicado: (2024)

Generative adversarial imitation learning for robot swarms: Learning from human demonstrations and trained policies
por: Kraus, Mattes, et al.
Publicado: (2026)

Entropy Coding of Unordered Data Structures
por: Kunze, Julius, et al.
Publicado: (2024)

Robust Entropy Search for Safe Efficient Bayesian Optimization
por: Weichert, Dorina, et al.
Publicado: (2024)

Curve Your Enthusiasm: Concurvity Regularization in Differentiable Generalized Additive Models
por: Siems, Julien, et al.
Publicado: (2023)

Revisiting Block-based Quantisation: What is Important for Sub-8-bit LLM Inference?
por: Zhang, Cheng, et al.
Publicado: (2023)

When Less is More: 8-bit Quantization Improves Continual Learning in Large Language Models
por: Zhang, Michael S., et al.
Publicado: (2025)

Ascend HiFloat8 Format for Deep Learning
por: Luo, Yuanyong, et al.
Publicado: (2024)

Research on Personalized Compression Algorithm for Pre-trained Models Based on Homomorphic Entropy Increase
por: Li, Yicong, et al.
Publicado: (2024)

Haiku to Opus in Just 10 bits: LLMs Unlock Massive Compression Gains
por: Rinberg, Roy, et al.
Publicado: (2026)

Qrazor: Reliable and Effortless 4-bit LLM Quantization by Significant Data Razoring
por: Lee, Dongyoung, et al.
Publicado: (2025)

A Free Lunch in LLM Compression: Revisiting Retraining after Pruning
por: Wagner, Moritz, et al.
Publicado: (2025)

Collusion Detection with Graph Neural Networks
por: Gomes, Lucas, et al.
Publicado: (2024)

Floating Anchor Diffusion Model for Multi-motif Scaffolding
por: Liu, Ke, et al.
Publicado: (2024)

Floating-Point Networks with Automatic Differentiation Can Represent Almost All Floating-Point Functions and Their Gradients
por: Park, Sejun, et al.
Publicado: (2026)

Learning Sparse Codes with Entropy-Based ELBOs
por: Velychko, Dmytro, et al.
Publicado: (2023)

BitNet v2: Native 4-bit Activations with Hadamard Transformation for 1-bit LLMs
por: Wang, Hongyu, et al.
Publicado: (2025)

On the Expressive Power of Floating-Point Transformers
por: Park, Sejun, et al.
Publicado: (2026)

Floating-Body Hydrodynamic Neural Networks
por: Zhang, Tianshuo, et al.
Publicado: (2025)

Model-Free Inference of Investor Preferences: A Relative Entropy IRL Approach
por: Xu, Chen
Publicado: (2026)

Data-Free Dynamic Compression of CNNs for Tractable Efficiency
por: Meiner, Lukas, et al.
Publicado: (2023)