# Scalable Methods for 8-bit Training of Neural Networks

arXiv: https://arxiv.org/abs/1805.11046 · HTML: https://mlsanity.com/paper/1805.11046/

## Key facts

- Cited by 142 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 May 2018.

## Built on by (most cited first, 30 of 142)

- [A Survey of Quantization Methods for Efficient Neural Network Inference](https://mlsanity.com/paper/2103.13630/) (2021)
- [Pruning and Quantization for Deep Neural Network Acceleration: A Survey](https://mlsanity.com/paper/2101.09671/) (2021)
- [NeRV: Neural Representations for Videos](https://mlsanity.com/paper/2110.13903/) (2021)
- [Accelerated Sparse Neural Training: A Provable and Efficient Method to Find N:M Transposable Masks](https://mlsanity.com/paper/2102.08124/) (2021)
- [Training Transformers with 4-bit Integers](https://mlsanity.com/paper/2306.11987/) (2023)
- [Optimizing Large Language Model Training Using FP4 Quantization](https://mlsanity.com/paper/2501.17116/) (2025)
- [Zero-shot Adversarial Quantization](https://mlsanity.com/paper/2103.15263/) (2021)
- [ZeroQuant-FP: A Leap Forward in LLMs Post-Training W4A8 Quantization Using Floating-Point Formats](https://mlsanity.com/paper/2307.09782/) (2023)
- [DAdaQuant: Doubly-adaptive quantization for communication-efficient Federated Learning](https://mlsanity.com/paper/2111.00465/) (2021)
- [Edge-Cloud Polarization and Collaboration: A Comprehensive Survey for AI](https://mlsanity.com/paper/2111.06061/) (2021)
- [Quartet: Native FP4 Training Can Be Optimal for Large Language Models](https://mlsanity.com/paper/2505.14669/) (2025)
- [Lightweight Deep Learning for Resource-Constrained Environments: A Survey](https://mlsanity.com/paper/2404.07236/) (2024)
- [Diversifying the High-level Features for better Adversarial Transferability](https://mlsanity.com/paper/2304.10136/) (2023)
- [FAST: DNN Training Under Variable Precision Block Floating Point with Stochastic Rounding](https://mlsanity.com/paper/2110.15456/) (2021)
- [GACT: Activation Compressed Training for Generic Network Architectures](https://mlsanity.com/paper/2206.11357/) (2022)
- [Jetfire: Efficient and Accurate Transformer Pretraining with INT8 Data Flow and Per-Block Quantization](https://mlsanity.com/paper/2403.12422/) (2024)
- [Distribution Adaptive INT8 Quantization for Training CNNs](https://mlsanity.com/paper/2102.04782/) (2021)
- [Differentiable Weightless Neural Networks](https://mlsanity.com/paper/2410.11112/) (2024)
- [Accurate Neural Training with 4-bit Matrix Multiplications at Standard Formats](https://mlsanity.com/paper/2112.10769/) (2021)
- [An Information-Theoretic Justification for Model Pruning](https://mlsanity.com/paper/2102.08329/) (2021)
- [8-bit Numerical Formats for Deep Neural Networks](https://mlsanity.com/paper/2206.02915/) (2022)
- [L2ight: Enabling On-Chip Learning for Optical Neural Networks via Efficient in-situ Subspace Optimization](https://mlsanity.com/paper/2110.14807/) (2021)
- [GSQ-Tuning: Group-Shared Exponents Integer in Fully Quantized Training for LLMs On-Device Fine-tuning](https://mlsanity.com/paper/2502.12913/) (2025)
- [CPT: Efficient Deep Neural Network Training via Cyclic Precision](https://mlsanity.com/paper/2101.09868/) (2021)
- [LVAC: Learned Volumetric Attribute Compression for Point Clouds using Coordinate Based Networks](https://mlsanity.com/paper/2111.08988/) (2021)
- [Extreme Compression for Pre-trained Transformers Made Simple and Efficient](https://mlsanity.com/paper/2206.01859/) (2022)
- [Distributed Learning Systems with First-order Methods](https://mlsanity.com/paper/2104.05245/) (2021)
- ["BNN - BN = ?": Training Binary Neural Networks without Batch Normalization](https://mlsanity.com/paper/2104.08215/) (2021)
- [Lightweight compression of neural network feature tensors for collaborative intelligence](https://mlsanity.com/paper/2105.06002/) (2021)
- [Adaptive Sharpness-Aware Pruning for Robust Sparse Networks](https://mlsanity.com/paper/2306.14306/) (2023)

## Abstract (arXiv)

Quantized Neural Networks (QNNs) are often used to improve network efficiency during the inference phase, i.e. after the network has been trained. Extensive research in the field suggests many different quantization schemes. Still, the number of bits required, as well as the best quantization scheme, are yet unknown. Our theoretical analysis suggests that most of the training process is robust to substantial precision reduction, and points to only a few specific operations that require higher precision. Armed with this knowledge, we quantize the model parameters, activations and layer gradients to 8-bit, leaving at a higher precision only the final step in the computation of the weight gradients. Additionally, as QNNs require batch-normalization to be trained at high precision, we introduce Range Batch-Normalization (BN) which has significantly higher tolerance to quantization noise and improved computational complexity. Our simulations show that Range BN is equivalent to the traditional batch norm if a precise scale adjustment, which can be approximated analytically, is applied. To the best of the authors' knowledge, this work is the first to quantize the weights, activations, as well as a substantial volume of the gradients stream, in all layers (including batch normalization) to 8-bit while showing state-of-the-art results over the ImageNet-1K dataset.