OrbitQuant is a data-agnostic PTQ technique for DiTs that uses RPBH rotation in a normalized basis to enable a single codebook across all inputs, achieving SOTA low-bit performance on FLUX.1, CogVideoX and similar models.
hub
OstQuant: Refining Large Language Model Quantization with Orthogonal and Scaling Transformations for Better Distribution Fitting
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
abstract
Post-training quantization (PTQ) has emerged as a widely adopted technique for compressing and accelerating Large Language Models (LLMs). The major challenge in LLM quantization is that uneven and heavy-tailed data distributions can expand the quantization range, thereby reducing bit precision for most values. Recent methods attempt to eliminate outliers and balance inter-channel differences by employing linear transformations; however, they remain heuristic and are often overlook optimizing the data distribution across the entire quantization space.In this paper, we introduce Quantization Space Utilization Rate (QSUR), a novel metric that effectively assesses the quantizability of transformed data by measuring the space utilization of the data in the quantization space. We complement QSUR with mathematical derivations that examine the effects and limitations of various transformations, guiding our development of Orthogonal and Scaling Transformation-based Quantization (OSTQuant). OSQuant employs a learnable equivalent transformation, consisting of an orthogonal transformation and a scaling transformation, to optimize the distributions of weights and activations across the entire quantization space. Futhermore, we propose the KL-Top loss function, designed to mitigate noise during optimization while retaining richer semantic information within the limited calibration data imposed by PTQ. OSTQuant outperforms existing work on various LLMs and benchmarks. In the W4-only setting, it retains 99.5\% of the floating-point accuracy. In the more challenging W4A4KV4 configuration, OSTQuant reduces the performance gap by 32\% on the LLaMA-3-8B model compared to state-of-the-art methods. \href{https://github.com/BrotherHappy/OSTQuant}{https://github.com/BrotherHappy/OSTQuant}.
hub tools
representative citing papers
LoopQ provides a loop-aware PTQ framework for recursive Transformers that mitigates distribution shift, state reuse, and recursive error accumulation, yielding 68.8% higher average accuracy and 87.7% lower perplexity under W4A4 versus static baselines.
QuantVLA is the first post-training quantization framework for VLA models that quantizes the diffusion transformer action head and reports higher task success rates than full-precision baselines with roughly 70% memory savings on the quantized components.
SplitQ improves low-bit PTQ for VLMs by isolating modality-specific outlier channels via MOCD and applying dual-branch adaptive calibration via ACC, outperforming prior methods on six datasets across W4A8 to W3A2 settings.
The paper introduces the Flatness metric, derives a theory-optimal quantization solution, and presents BDQ that uses bidirectional diagonal transformations to reduce outlier impact, achieving under 1% drop at W4A4 on LLaMA-3-8B.
CoQuant selects optimal high-precision subspaces for mixed-precision LLM quantization via a closed-form weighted PCA that balances weight and activation covariances derived from expected output error.
TAH-Quant introduces tile-wise adaptive Hadamard quantization for activations in pipeline parallelism, achieving 3-4 bit compression with up to 4.3x throughput speedup and O(1/sqrt(T)) convergence matching SGD.
BTC-LLM uses a binary codebook for pattern clustering and a learnable transformation to achieve 0.7-1.11 bit LLM quantization while limiting accuracy loss to a few percent on LLaMA and Qwen models.
MixFP4 extends NVFP4 by adaptively selecting between two FP4 micro-formats per block using repurposed scale sign bits and a unified E2M2 compute path, claiming better accuracy than standard NVFP4 at 3.1% area and 1.5% power overhead.
MGVQ introduces sensitivity-aware structured mixed-precision VQ and gradient-aware second-order error compensation using Kronecker and Block-LDL decompositions, reporting up to 4.9 point gains over prior methods at 2-bit on models like InternVL2-26B.
GAMMA is a post-training framework that learns stable module sensitivity rankings for mixed-precision LLM quantization and projects them to exact bit budgets via integer programming, enabling reuse across arbitrary memory targets.
citing papers explorer
-
OrbitQuant: Data-Agnostic Quantization for Image and Video Diffusion Transformers
OrbitQuant is a data-agnostic PTQ technique for DiTs that uses RPBH rotation in a normalized basis to enable a single codebook across all inputs, achieving SOTA low-bit performance on FLUX.1, CogVideoX and similar models.
-
LoopQ: Quantization for Recursive Transformers
LoopQ provides a loop-aware PTQ framework for recursive Transformers that mitigates distribution shift, state reuse, and recursive error accumulation, yielding 68.8% higher average accuracy and 87.7% lower perplexity under W4A4 versus static baselines.
-
QuantVLA: Scale-Calibrated Post-Training Quantization for Vision-Language-Action Models
QuantVLA is the first post-training quantization framework for VLA models that quantizes the diffusion transformer action head and reports higher task success rates than full-precision baselines with roughly 70% memory savings on the quantized components.
-
Breaking Modality Heterogeneity in Low-Bit Quantization for Large Vision-Language Models
SplitQ improves low-bit PTQ for VLMs by isolating modality-specific outlier channels via MOCD and applying dual-branch adaptive calibration via ACC, outperforming prior methods on six datasets across W4A8 to W3A2 settings.
-
Theory-optimal Quantization Based on Flatness
The paper introduces the Flatness metric, derives a theory-optimal quantization solution, and presents BDQ that uses bidirectional diagonal transformations to reduce outlier impact, achieving under 1% drop at W4A4 on LLaMA-3-8B.
-
CoQuant: Joint Weight-Activation Subspace Projection for Mixed-Precision LLMs
CoQuant selects optimal high-precision subspaces for mixed-precision LLM quantization via a closed-form weighted PCA that balances weight and activation covariances derived from expected output error.
-
TAH-QUANT: Effective Activation Quantization in Pipeline Parallelism over Slow Network
TAH-Quant introduces tile-wise adaptive Hadamard quantization for activations in pipeline parallelism, achieving 3-4 bit compression with up to 4.3x throughput speedup and O(1/sqrt(T)) convergence matching SGD.
-
BTC-LLM: Efficient Sub-1-Bit LLM Quantization via Learnable Transformation and Binary Codebook
BTC-LLM uses a binary codebook for pattern clustering and a learnable transformation to achieve 0.7-1.11 bit LLM quantization while limiting accuracy loss to a few percent on LLaMA and Qwen models.
-
MixFP4: Enhancing NVFP4 with Adaptive FP4/INT4 Block Representations
MixFP4 extends NVFP4 by adaptively selecting between two FP4 micro-formats per block using repurposed scale sign bits and a unified E2M2 compute path, claiming better accuracy than standard NVFP4 at 3.1% area and 1.5% power overhead.
-
MGVQ: Synergizing Multi-dimensional Sensitivity-Aware and Gradient-Hessian Fusion for Vector Quantization
MGVQ introduces sensitivity-aware structured mixed-precision VQ and gradient-aware second-order error compensation using Kronecker and Block-LDL decompositions, reporting up to 4.9 point gains over prior methods at 2-bit on models like InternVL2-26B.
-
GAMMA: Global Bit Allocation for Mixed-Precision Models under Arbitrary Budgets
GAMMA is a post-training framework that learns stable module sensitivity rankings for mixed-precision LLM quantization and projects them to exact bit budgets via integer programming, enabling reuse across arbitrary memory targets.
- KronQ: LLM Quantization via Kronecker-Factored Hessian
- HoloQ-VLA: Uniform W4A4 Quantization of Vision-Language-Action Models