On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,657 papers · 148 categories

Trend · papers per month

count % of corpus

6 results for “PTQ”

LPCD: Unified Framework from Layer-Wise to Submodule Quantizationstat.ML

Unified framework LPCD optimizes quantization of complex submodules.

problem Quantization of complex submodules in neural networks.

method Layer-Projected Coordinate Descent (LPCD) for quantizing arbitrary submodules.

result LPCD enhances both layer-wise PTQ methods and existing submodule approaches.

quantization submodule Layer-Projected Coordinate Descent PTQ neural networks

2025-12-01abs ↗pdf ↗

Understanding Quantization-Aware Training: Gradients at Quantized Weights Bias to the Low-Loss Basincs.LG

Quantization-aware training can recover accuracy lost by post-training quantization.

problem Post-training quantization (PTQ) can fail sharply at aggressive bitwidths.

method A unified geometric framework that explains PTQ failure and QAT recovery.

result QAT has a useful bias that steers iterates back into the basin.

quantization-aware training post-training quantization geometric framework valley basin

2026-06-08abs ↗pdf ↗

Quantization Error Propagation: Revisiting Layer-Wise Post-Training Quantizationcs.LG

Proposes QEP to mitigate quantization error propagation in layer-wise post-training quantization.

problem Growth of quantization errors across layers degrades performance, especially in low-bit regimes.

method Quantization Error Propagation (QEP) framework that explicitly propagates and compensates for quantization errors.

result QEP-enhanced layer-wise PTQ achieves substantially higher accuracy, especially in low-bit regimes.

quantization error post-training quantization layer-wise error propagation language models

2025-04-13abs ↗pdf ↗

QFT: Post-training quantization via fast joint finetuning of all degrees of freedomstat.ML

Unified finetuning of all quantization degrees of freedom achieves state-of-the-art 4-bit quantization.

problem Achieving high accuracy in quantized neural networks while maintaining speed and resource constraints.

method Quantization-aware finetuning (QFT) that jointly optimizes all quantization degrees of freedom.

result 4-bit weight quantization results on-par with state-of-the-art (SoTA) within PTQ constraints.

quantization finetuning degrees of freedom 4-bit state-of-the-art

2022-12-05abs ↗pdf ↗

S2D: Selective Spectral Decay for Quantization-Friendly Conditioning of Neural Activationscs.LG

S2D selectively decays large singular values to improve quantization of neural activations.

problem Large activation outliers in transformer models cause accuracy drops during quantization.

method Selective Spectral Decay (

S^2D

) that surgically regularizes only the largest singular values.

result Significantly reduces activation outliers and produces well-conditioned representations.

quantization neural activations transformer models singular values fine-tuning

2026-02-16abs ↗pdf ↗

QuantEase: Optimization-based Quantization for Language Modelsstat.ML

QuantEase optimizes LLMs with CD-based quantization, achieving state-of-the-art performance.

problem Efficiently quantize large language models for deployment.

method Layer-wise quantization using CD-based algorithms with matrix and vector operations.

result State-of-the-art performance in perplexity and zero-shot accuracy.

Large Language Models Post-Training Quantization Coordinate Descent Matrix Operations Perplexity

2023-09-05abs ↗pdf ↗