# Towards Unified INT8 Training for Convolutional Neural Network

arXiv: https://arxiv.org/abs/1912.12607 · HTML: https://mlsanity.com/paper/1912.12607/

## Key facts

- Cited by 79 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Dec 2019.

## Built on by (most cited first, 30 of 79)

- [Pruning and Quantization for Deep Neural Network Acceleration: A Survey](https://mlsanity.com/paper/2101.09671/) (2021)
- [Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Models](https://mlsanity.com/paper/2209.13325/) (2022)
- [Understanding and Overcoming the Challenges of Efficient Transformer Quantization](https://mlsanity.com/paper/2109.12948/) (2021)
- [Stable and low-precision training for large-scale vision-language models](https://mlsanity.com/paper/2304.13013/) (2023)
- [Diversifying Sample Generation for Accurate Data-Free Quantization](https://mlsanity.com/paper/2103.01049/) (2021)
- [Training Transformers with 4-bit Integers](https://mlsanity.com/paper/2306.11987/) (2023)
- [BiT: Robustly Binarized Multi-distilled Transformer](https://mlsanity.com/paper/2205.13016/) (2022)
- [Zero-shot Adversarial Quantization](https://mlsanity.com/paper/2103.15263/) (2021)
- [Q-GaLore: Quantized GaLore with INT4 Projection and Layer-Adaptive Low-Rank Gradients](https://mlsanity.com/paper/2407.08296/) (2024)
- [Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety](https://mlsanity.com/paper/2104.14235/) (2021)
- [COAT: Compressing Optimizer states and Activation for Memory-Efficient FP8 Training](https://mlsanity.com/paper/2410.19313/) (2024)
- [Jetfire: Efficient and Accurate Transformer Pretraining with INT8 Data Flow and Per-Block Quantization](https://mlsanity.com/paper/2403.12422/) (2024)
- [A Survey of Low-bit Large Language Models: Basics, Systems, and Algorithms](https://mlsanity.com/paper/2409.16694/) (2024)
- [Distribution Adaptive INT8 Quantization for Training CNNs](https://mlsanity.com/paper/2102.04782/) (2021)
- [F8Net: Fixed-Point 8-bit Only Multiplication for Network Quantization](https://mlsanity.com/paper/2202.05239/) (2022)
- [Distribution-sensitive Information Retention for Accurate Binary Neural Network](https://mlsanity.com/paper/2109.12338/) (2021)
- [GSQ-Tuning: Group-Shared Exponents Integer in Fully Quantized Training for LLMs On-Device Fine-tuning](https://mlsanity.com/paper/2502.12913/) (2025)
- [Oscillation-Reduced MXFP4 Training for Vision Transformers](https://mlsanity.com/paper/2502.20853/) (2025)
- [CPT: Efficient Deep Neural Network Training via Cyclic Precision](https://mlsanity.com/paper/2101.09868/) (2021)
- [AdaptiveNet: Post-deployment Neural Architecture Adaptation for Diverse Edge Environments](https://mlsanity.com/paper/2303.07129/) (2023)
- [Nearest is Not Dearest: Towards Practical Defense against Quantization-conditioned Backdoor Attacks](https://mlsanity.com/paper/2405.12725/) (2024)
- [Data Quality-aware Mixed-precision Quantization via Hybrid Reinforcement Learning](https://mlsanity.com/paper/2302.04453/) (2023)
- [Solving Oscillation Problem in Post-Training Quantization Through a Theoretical Perspective](https://mlsanity.com/paper/2303.11906/) (2023)
- [Accurate INT8 Training Through Dynamic Block-Level Fallback](https://mlsanity.com/paper/2503.08040/) (2025)
- [Quantized Distributed Training of Large Models with Convergence Guarantees](https://mlsanity.com/paper/2302.02390/) (2023)
- [Is Integer Arithmetic Enough for Deep Learning Training?](https://mlsanity.com/paper/2207.08822/) (2022)
- [Bitwidth Heterogeneous Federated Learning with Progressive Weight Dequantization](https://mlsanity.com/paper/2202.11453/) (2022)
- [Low-Precision Training of Large Language Models: Methods, Challenges, and Opportunities](https://mlsanity.com/paper/2505.01043/) (2025)
- [Adaptive Message Quantization and Parallelization for Distributed Full-graph GNN Training](https://mlsanity.com/paper/2306.01381/) (2023)
- [Resource-Efficient Deep Learning: A Survey on Model-, Arithmetic-, and Implementation-Level Techniques](https://mlsanity.com/paper/2112.15131/) (2021)

## Abstract (arXiv)

Recently low-bit (e.g., 8-bit) network quantization has been extensively studied to accelerate the inference. Besides inference, low-bit training with quantized gradients can further bring more considerable acceleration, since the backward process is often computation-intensive. Unfortunately, the inappropriate quantization of backward propagation usually makes the training unstable and even crash. There lacks a successful unified low-bit training framework that can support diverse networks on various tasks. In this paper, we give an attempt to build a unified 8-bit (INT8) training framework for common convolutional neural networks from the aspects of both accuracy and speed. First, we empirically find the four distinctive characteristics of gradients, which provide us insightful clues for gradient quantization. Then, we theoretically give an in-depth analysis of the convergence bound and derive two principles for stable INT8 training. Finally, we propose two universal techniques, including Direction Sensitive Gradient Clipping that reduces the direction deviation of gradients and Deviation Counteractive Learning Rate Scaling that avoids illegal gradient update along the wrong direction. The experiments show that our unified solution promises accurate and efficient INT8 training for a variety of networks and tasks, including MobileNetV2, InceptionV3 and object detection that prior studies have never succeeded. Moreover, it enjoys a strong flexibility to run on off-the-shelf hardware, and reduces the training time by 22% on Pascal GPU without too much optimization effort. We believe that this pioneering study will help lead the community towards a fully unified INT8 training for convolutional neural networks.