BCJR-QAT makes trellis quantization differentiable via BCJR soft decoding at finite temperature, allowing QAT to improve 2-bit LLM perplexity over PTQ with a fused GPU kernel and a drift-budget escape condition.
Efficientqat: Efficient quantization-aware training for large language models
12 Pith papers cite this work, alongside 7 external citations. Polarity classification is still indexing.
12
Pith papers citing it
7
external citations · Pith
abstract
Large language models (LLMs) are crucial in modern natural language processing and artificial intelligence. However, they face challenges in managing their significant memory requirements. Although quantization-aware training (QAT) offers a solution by reducing memory consumption through low-bit representations with minimal accuracy loss, it is impractical due to substantial training resources. To address this, we propose Efficient Quantization-Aware Training (EfficientQAT), a more feasible QAT algorithm. EfficientQAT involves two consecutive phases: Block-wise training of all parameters (Block-AP) and end-to-end training of quantization parameters (E2E-QP). To the best of our knowledge, Block-AP is the first method to enable direct training of all parameters in a block-wise manner, reducing accuracy loss in low-bit scenarios by enhancing the solution space during optimization. E2E-QP then trains only the quantization parameters (step sizes) end-to-end, further improving the performance of quantized models by considering interactions among all sub-modules. Extensive experiments demonstrate that EfficientQAT outperforms previous quantization methods across a range of models, including base LLMs, instruction-tuned LLMs, and multimodal LLMs, with scales from 7B to 70B parameters at various quantization bits. For instance, EfficientQAT obtains a 2-bit Llama-2-70B model on a single A100-80GB GPU in 41 hours, with less than 3 points accuracy degradation compared to the full precision (69.48 vs. 72.41). Code is available at https://github.com/OpenGVLab/EfficientQAT.
citation-role summary
background 1
baseline 1
citation-polarity summary
representative citing papers
DPQuant uses epoch-wise probabilistic layer rotation and DP loss sensitivity to quantize only a changing subset of layers, reducing accuracy degradation from quantization noise in DP-SGD and delivering up to 2.21x throughput gains with under 2% accuracy drop.
LC-QAT achieves data-efficient 2-bit weight-only QAT for LLMs by representing quantized weights as a learned affine transform over discrete vectors, supporting end-to-end optimization from a high-quality PTQ start.
LiftQuant enables continuous bit-width LLM quantization via dimensional lifting and projection from a 1-bit lattice, allowing 2.4-bit compression of 70B models that outperforms fixed 2-bit baselines on identical hardware.
Nonlinear Bipolar Compensation with Bipolar Logarithmic Transformation reduces outlier effects in post-training quantization by performing compensation in a compressed transformed space.
SURGE proposes a dual-path gradient compensator and adaptive gradient scaler to mitigate gradient mismatch in binary neural network training via auxiliary backpropagation.
LAQuant improves long-decoding accuracy on quantized reasoning models like Qwen3-4B by 15pp on AIME25 via layer-wise lookahead loss, achieving 3.42x speedup over FP16.
BitRL enables on-device RL agents via 1-bit quantized language models, delivering 10-16x memory reduction and 3-5x energy efficiency gains with 85-98% retained performance.
Quantized LLMs diverge from their base models at the decision level even when accuracy is preserved, with query and key attention projections showing the greatest structural distortion under low-bit compression.
GNMR is a gradient-norm-based controller that maps local stability signals to budgeted recovery actions to stabilize low-precision LLM training while preserving quality.
Factorial experiments with over 1300 runs falsify the hypothesis that INT6 QAT needs a different LR schedule from higher precision and identify a 50M-parameter boundary for INT4 schedule sensitivity.
HCInfer recovers up to 5.2% accuracy over compressed LLMs and delivers 10.4x speedup versus full-precision models by offloading compensation parameters to CPU with async execution on resource-limited hardware.
citing papers explorer
-
BCJR-QAT: A Differentiable Relaxation of Trellis-Coded Weight Quantization
BCJR-QAT makes trellis quantization differentiable via BCJR soft decoding at finite temperature, allowing QAT to improve 2-bit LLM perplexity over PTQ with a fused GPU kernel and a drift-budget escape condition.
-
DPQuant: Efficient and Differentially-Private Model Training via Dynamic Quantization Scheduling
DPQuant uses epoch-wise probabilistic layer rotation and DP loss sensitivity to quantize only a changing subset of layers, reducing accuracy degradation from quantization noise in DP-SGD and delivering up to 2.21x throughput gains with under 2% accuracy drop.
-
LC-QAT: Data-Efficient 2-Bit QAT for LLMs via Linear-Constrained Vector Quantization
LC-QAT achieves data-efficient 2-bit weight-only QAT for LLMs by representing quantized weights as a learned affine transform over discrete vectors, supporting end-to-end optimization from a high-quality PTQ start.
-
LiftQuant: Continuous Bit-Width LLM via Dimensional Lifting and Projection
LiftQuant enables continuous bit-width LLM quantization via dimensional lifting and projection from a 1-bit lattice, allowing 2.4-bit compression of 70B models that outperforms fixed 2-bit baselines on identical hardware.
-
Nonlinear Bipolar Compensation: Handling Outliers in Post-Training Quantization
Nonlinear Bipolar Compensation with Bipolar Logarithmic Transformation reduces outlier effects in post-training quantization by performing compensation in a compressed transformed space.
-
SURGE: Surrogate Gradient Adaptation in Binary Neural Networks
SURGE proposes a dual-path gradient compensator and adaptive gradient scaler to mitigate gradient mismatch in binary neural network training via auxiliary backpropagation.
-
LAQuant: A Simple Overhead-free Large Reasoning Model Quantization by Layer-wise Lookahead Loss
LAQuant improves long-decoding accuracy on quantized reasoning models like Qwen3-4B by 15pp on AIME25 via layer-wise lookahead loss, achieving 3.42x speedup over FP16.
-
BitRL: Reinforcement Learning with 1-bit Quantized Language Models for Resource-Constrained Edge Deployment
BitRL enables on-device RL agents via 1-bit quantized language models, delivering 10-16x memory reduction and 3-5x energy efficiency gains with 85-98% retained performance.
-
The Illusion of Equivalency: Statistical Characterization of Quantization Effects in LLMs
Quantized LLMs diverge from their base models at the decision level even when accuracy is preserved, with query and key attention projections showing the greatest structural distortion under low-bit compression.
-
GNMR: Runtime Stability Control for Low-Precision Large Language Model Training
GNMR is a gradient-norm-based controller that maps local stability signals to budgeted recovery actions to stabilize low-precision LLM training while preserving quality.
-
Mapping the Schedule x Bit-Width Boundary in Sub-100M Quantisation-Aware Training
Factorial experiments with over 1300 runs falsify the hypothesis that INT6 QAT needs a different LR schedule from higher precision and identify a 50M-parameter boundary for INT4 schedule sensitivity.
-
HCInfer: An Efficient Inference System via Error Compensation for Resource-Constrained Devices
HCInfer recovers up to 5.2% accuracy over compressed LLMs and delivers 10.4x speedup versus full-precision models by offloading compensation parameters to CPU with async execution on resource-limited hardware.