# TAH-QUANT: Effective Activation Quantization in Pipeline Parallelism over Slow Network

arXiv: https://arxiv.org/abs/2506.01352 · HTML: https://mlsanity.com/paper/2506.01352/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 44 of its 60 references that are in the map.
- First posted on arXiv 2 Jun 2025.

## Builds on (most cited first, 30 of 44)

- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [Program Synthesis with Large Language Models](https://mlsanity.com/paper/2108.07732/) (2021)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [Training Compute-Optimal Large Language Models](https://mlsanity.com/paper/2203.15556/) (2022)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding](https://mlsanity.com/paper/1510.00149/) (2015)
- [GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers](https://mlsanity.com/paper/2210.17323/) (2022)
- [SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models](https://mlsanity.com/paper/2211.10438/) (2022)
- [AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration](https://mlsanity.com/paper/2306.00978/) (2023)
- [KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache](https://mlsanity.com/paper/2402.02750/) (2024)
- [QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMs](https://mlsanity.com/paper/2404.00456/) (2024)
- [SpinQuant: LLM quantization with learned rotations](https://mlsanity.com/paper/2405.16406/) (2024)
- [Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations](https://mlsanity.com/paper/1609.07061/) (2016)
- [LLM-QAT: Data-Free Quantization Aware Training for Large Language Models](https://mlsanity.com/paper/2305.17888/) (2023)
- [DuQuant: Distributing Outliers via Dual Transformation Makes Stronger Quantized LLMs](https://mlsanity.com/paper/2406.01721/) (2024)
- [Memory-Efficient Pipeline-Parallel DNN Training](https://mlsanity.com/paper/2006.09503/) (2020)
- [The Lazy Neuron Phenomenon: On Emergence of Activation Sparsity in Transformers](https://mlsanity.com/paper/2210.06313/) (2022)
- [OstQuant: Refining Large Language Model Quantization with Orthogonal and Scaling Transformations for Better Distribution Fitting](https://mlsanity.com/paper/2501.13987/) (2025)
- [Decentralized Training of Foundation Models in Heterogeneous Environments](https://mlsanity.com/paper/2206.01288/) (2022)
- [ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training](https://mlsanity.com/paper/2104.14129/) (2021)
- [SWARM Parallelism: Training Large Models Can Be Surprisingly Communication-Efficient](https://mlsanity.com/paper/2301.11913/) (2023)
- [Distributed Deep Learning in Open Collaborations](https://mlsanity.com/paper/2106.10207/) (2021)
- [Helix: Serving Large Language Models over Heterogeneous GPUs and Network via Max-Flow](https://mlsanity.com/paper/2406.01566/) (2024)
- [Towards Crowdsourced Training of Large Neural Networks using Decentralized Mixture-of-Experts](https://mlsanity.com/paper/2002.04013/) (2020)
- [Distributed Inference and Fine-tuning of Large Language Models Over The Internet](https://mlsanity.com/paper/2312.08361/) (2023)
- [Compressing DMA Engine: Leveraging Activation Sparsity for Training Deep Neural Networks](https://mlsanity.com/paper/1705.01626/) (2017)
- [Accelerating Convolutional Neural Networks via Activation Map Compression](https://mlsanity.com/paper/1812.04056/) (2018)

## Built on by (most cited first, 4 of 4)

- [Evolution of Optimization Methods: Algorithms, Scenarios, and Evaluations](https://mlsanity.com/paper/2604.12968/) (2026)
- [TACO: Efficient Communication Compression of Intermediate Tensors for Scalable Tensor-Parallel LLM Training](https://mlsanity.com/paper/2604.24088/) (2026)
- [NCCLZ: Compression-Enabled GPU Collectives with Decoupled Quantization and Entropy Coding](https://mlsanity.com/paper/2605.12396/) (2026)
- [Communication-Efficient LLM Adaptation over Decentralized GPU Meshes](https://mlsanity.com/paper/2609.14339/) (2026)

## Abstract (arXiv)

Decentralized training of large language models offers the opportunity to pool computational resources across geographically distributed participants, but is often bottlenecked by network communication, particularly under pipeline parallel settings. While pipeline parallelism partitions model layers across devices to handle large-scale models, it necessitates frequent communication of intermediate activations, creating challenges when network bandwidth is limited. To address these issues, we propose TAH-Quant (Tile-wise Adaptive Hadamard Quantization), a novel activation quantization framework for pipeline parallelism. TAH-Quant integrates fine-grained tile-wise quantization, entropy-guided tile-wise adaptive bit allocation for optimal bit usage, and a Hadamard-based transformation with pivot swapping to effectively suppress outliers. Compared with token-level allocation, the tile-wise allocator assigns precision at the granularity of small channel windows within each token, reducing quantization error under the same bit budget. We prove that pipeline parallel training equipped with TAH-Quant maintains a convergence rate of O(1/sqrt(T)), matching that of vanilla stochastic gradient descent. Extensive experiments demonstrate that TAH-Quant achieves an aggressive activation quantization ratio of 3-4 bits, providing up to 4.3x throughput speedup over uncompressed FP32 and up to 1.33x wall-clock speedup over AQ-SGD, while preserving training convergence, avoiding AQ-SGD's activation-cache overhead, and generalizing well across various training scenarios.