Score: 0

LRQ-DiT: Log-Rotation Post-Training Quantization of Diffusion Transformers for Text-to-Image Generation

Published: August 5, 2025 | arXiv ID: 2508.03485v1

By: Lianwei Yang , Haokun Lin , Tianchen Zhao and more

Potential Business Impact:

Makes AI art generators smaller and faster.

Diffusion Transformers (DiTs) have achieved impressive performance in text-to-image generation. However, their high computational cost and large parameter sizes pose significant challenges for usage in resource-constrained scenarios. Post-training quantization (PTQ) is a promising solution to reduce memory usage and accelerate inference, but existing PTQ methods suffer from severe performance degradation under extreme low-bit settings. We identify two key obstacles to low-bit post-training quantization for DiT models: (1) model weights follow a Gaussian-like distribution with long tails, causing uniform quantization to poorly allocate intervals and leading to significant errors; (2) two types of activation outliers: (i) Mild Outliers with slightly elevated values, and (ii) Salient Outliers with large magnitudes concentrated in specific channels, which disrupt activation quantization. To address these issues, we propose LRQ-DiT, an efficient and accurate PTQ framework. We introduce Twin-Log Quantization (TLQ), a log-based method that aligns well with the weight distribution and reduces quantization errors. We also propose an Adaptive Rotation Scheme (ARS) that dynamically applies Hadamard or outlier-aware rotations based on activation fluctuation, effectively mitigating the impact of both types of outliers. We evaluate LRQ-DiT on PixArt and FLUX under various bit-width settings, and validate the performance on COCO, MJHQ, and sDCI datasets. LRQ-DiT achieves low-bit quantization of DiT models while preserving image quality, outperforming existing PTQ baselines.

Page Count
17 pages

Category
Computer Science:
CV and Pattern Recognition