Training with Fewer Bits: Unlocking Edge LLMs Training with Stochastic Rounding (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Quantized training improves computational and memory efficiency but introduces quantization noise. |
| Approach: | They propose to use stochastic rounding to improve LLM training but introduce quantization noise. |
| Outcome: | The proposed method can compensate for reduced accuracy during backpropagation. |
Similar Papers
Low-Bit Quantization Favors Undertrained LLMs (2025.acl-long)
Copied to clipboard
| Challenge: | Larger models or those trained on fewer tokens exhibit less quantization-induced degradation (QiD), while smaller, well-trained models face significant performance losses. |
| Approach: | They propose to use QiD to measure an LLM’s training levels and determine the number of training tokens required for fully training LLMs of various sizes. |
| Outcome: | The proposed scaling laws can predict the quantization performance of different-sized LLMs trained with tokens. |
BWLA: Breaking the Barrier of W1AX Post-Training Quantization for LLMs (2026.acl-long)
Copied to clipboard
| Challenge: | Large language models have driven major progress in NLP, but memory and compute requirements hinder practical deployment. |
| Approach: | They propose a framework that preserves high accuracy while achieving 1-bit weight quantization . the orthogonal-kronecker transformation learns an orthogonale mapping via EM minimization - a new approach to quantization is proposed . |
| Outcome: | The proposed framework achieves 1-bit weight quantization with low activations with low-bit activations. |
Half-S: Halving the Scale for Near-Lossless 4-Bit LLM Training (2026.findings-acl)
Copied to clipboard
Jinyang Du, Ruihao Gong, Linghan Ai, Zining Wang, Yunke Peng, Yao Wang, Lei Yan, null Wxuefei, Yaoyuan Wang, Jinyang Guo, Dahua Lin, Xianglong Liu
| Challenge: | Existing 4-bit training pipelines rely on max-scaling, which causes representation collapse . despite this, there are limitations in the accuracy of 4-bit LLM training . |
| Approach: | They propose a scaling strategy that uses half-scaling as a hardware-friendly default . they propose fp4 support that allows for a faster scaling of large language models . |
| Outcome: | The proposed scaling strategy narrows the gap between theoretical optimum and BF16 while maintaining the efficiency benefits of 4-bit training. |
Optimize Weight Rounding via Signed Gradient Descent for the Quantization of LLMs (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have demonstrated exceptional proficiency in language-related tasks, but their deployment poses significant memory and storage requirements. |
| Approach: | They propose a method that optimizes rounding values and weight clipping within 200 steps. |
| Outcome: | The proposed method achieves exceptional results across 2 to 4 bits while maintaining low tuning costs and avoiding additional inference overhead. |
On the Impact of Calibration Data in Post-training Quantization and Pruning (2024.acl-long)
Copied to clipboard
| Challenge: | Quantization and pruning are the foundations of compression for large language models . however, no prior work has investigated how calibration data impacts performance of compression methods. |
| Approach: | They propose an empirical study on the effect of calibration data on LLM performance. |
| Outcome: | The proposed methods improve performance in a post-training setting. |
Revisiting Block-based Quantisation: What is Important for Sub-8-bit LLM Inference? (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing quantisation methods mainly focus on 8-bit LLMs . a lack of scaling offsets in the quantisation process limits the use of LLM inference. |
| Approach: | They propose to use block quantisations to reduce scaling offsets in Large language models . they find that the block quantizations reduce scaling only from an arithmetic perspective . |
| Outcome: | The proposed methods reduce scaling offsets solely from an arithmetic perspective without additional treatments in the computational path. |
Achieving binary weight and activation for LLMs using Post-Training Quantization (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for quantizing large language models suffer from performance degradation when weights are quantized to 1 bit. |
| Approach: | They propose a post-training quantization framework with W(1+1)A(14) configuration . they propose utilizing Hessian-aware fine-grained grouping along with an EM-based quantization scheme . |
| Outcome: | The proposed method surpasses state-of-the-art (SOTA) LLM quantization baselines on W2A4 across multiple tasks. |
LLM-QAT: Data-Free Quantization Aware Training for Large Language Models (2024.findings-acl)
Copied to clipboard
Zechun Liu, Barlas Oguz, Changsheng Zhao, Ernie Chang, Pierre Stock, Yashar Mehdad, Yangyang Shi, Raghuraman Krishnamoorthi, Vikas Chandra
| Challenge: | Several post-training quantization methods have been shown to perform well down to 8-bits. |
| Approach: | They propose a data-free distillation method that leverages generations produced by the pre-trained model to quantize any generative model independent of its training data. |
| Outcome: | The proposed method outperforms SoTA PTQ and LLaMA models at low bit precision. |
PE-QAT: Parameter-Efficient Quantization-Aware Training for Large Language Models (2026.acl-srw)
Copied to clipboard
| Challenge: | Quantization Aware Training (QAT) is expensive to train and unscalable to large models. |
| Approach: | They propose a parameter-efficient framework targeting per-channel 4-bit weight-activation quantization of large language models. |
| Outcome: | The proposed framework preserves accuracy within 0.11 percentage points of the full-precision baseline on Llama-2-7B zero-shot tasks while training only 1.26% of total parameters. |
CompAct: Compressed Activations for Memory-Efficient LLM Training (2025.naacl-long)
Copied to clipboard
| Challenge: | Recent studies have focused on reducing peak memory utilization on GPUs, but most work only target the computation graph during training. |
| Approach: | They propose a technique that reduces peak memory utilization on GPUs by 25-30% for pretraining and 50% for fine-tuning of LLMs. |
| Outcome: | The proposed technique reduces peak memory utilization on GPUs by 25-30% for pretraining and 50% for fine-tuning of LLMs. |