# Post-Training 4-bit Quantization on Embedding Tables

arXiv: https://arxiv.org/abs/1911.02079 · HTML: https://mlsanity.com/paper/1911.02079/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Nov 2019.

## Built on by (most cited first, 22 of 22)

- [TT-Rec: Tensor Train Compression for Deep Learning Recommendation Models](https://mlsanity.com/paper/2101.11714/) (2021)
- [Experimental Analysis of Large-scale Learnable Vector Storage Compression](https://mlsanity.com/paper/2311.15578/) (2023)
- [Adaptive Low-Precision Training for Embeddings in Click-Through Rate Prediction](https://mlsanity.com/paper/2212.05735/) (2022)
- [LightLM: A Lightweight Deep and Narrow Language Model for Generative Recommendation](https://mlsanity.com/paper/2310.17488/) (2023)
- [Low-Precision Hardware Architectures Meet Recommendation Model Inference at Scale](https://mlsanity.com/paper/2105.12676/) (2021)
- [Supporting Massive DLRM Inference Through Software Defined Memory](https://mlsanity.com/paper/2110.11489/) (2021)
- [Accelerating Communication in Deep Learning Recommendation Model Training with Dual-Level Adaptive Lossy Compression](https://mlsanity.com/paper/2407.04272/) (2024)
- [Efficient Soft-Error Detection for Low-precision Deep Learning Recommendation Models](https://mlsanity.com/paper/2103.00130/) (2021)
- [The Future is Sparse: Embedding Compression for Scalable Retrieval in Recommender Systems](https://mlsanity.com/paper/2505.11388/) (2025)
- [CoVE: Compressed Vocabulary Expansion Makes Better LLM-based Recommender Systems](https://mlsanity.com/paper/2506.19993/) (2025)
- [Inshrinkerator: Compressing Deep Learning Training Checkpoints via Dynamic Quantization](https://mlsanity.com/paper/2306.11800/) (2023)
- [DQRM: Deep Quantized Recommendation Models](https://mlsanity.com/paper/2410.20046/) (2024)
- [On-device Content-based Recommendation with Single-shot Embedding Pruning: A Cooperative Game Perspective](https://mlsanity.com/paper/2411.13052/) (2024)
- [Fine-Grained Embedding Dimension Optimization During Training for Recommender Systems](https://mlsanity.com/paper/2401.04408/) (2024)
- [Rec-AD: An Efficient Computation Framework for FDIA Detection Based on Tensor Train Decomposition and Deep Learning Recommendation Model](https://mlsanity.com/paper/2507.14668/) (2025)
- [Toward Compact Parameter Representations for Architecture-Agnostic Neural Network Compression](https://mlsanity.com/paper/2111.10320/) (2021)
- [Review of compressed embedding layers and their applications for recommender systems](https://mlsanity.com/paper/2306.13724/) (2023)
- [FedShift: Robust Federated Learning Aggregation Scheme in Resource Constrained Environment via Weight Shifting](https://mlsanity.com/paper/2402.01070/) (2024)
- [ProactivePIM: Accelerating Weight-Sharing Embedding Layer with PIM for Scalable Recommendation System](https://mlsanity.com/paper/2402.04032/) (2024)
- [Stochastic Communication Avoidance for Recommendation Systems](https://mlsanity.com/paper/2411.01611/) (2024)
- [ShuffleGate: Scalable Feature Optimization for Recommender Systems via Batch-wise Sensitivity Learning](https://mlsanity.com/paper/2503.09315/) (2025)
- [DimGrow: Memory-Efficient Field-level Embedding Dimension Search](https://mlsanity.com/paper/2505.12683/) (2025)

## Abstract (arXiv)

Continuous representations have been widely adopted in recommender systems where a large number of entities are represented using embedding vectors. As the cardinality of the entities increases, the embedding components can easily contain millions of parameters and become the bottleneck in both storage and inference due to large memory consumption. This work focuses on post-training 4-bit quantization on the continuous embeddings. We propose row-wise uniform quantization with greedy search and codebook-based quantization that consistently outperforms state-of-the-art quantization approaches on reducing accuracy degradation. We deploy our uniform quantization technique on a production model in Facebook and demonstrate that it can reduce the model size to only 13.89% of the single-precision version while the model quality stays neutral.