Research map / Architectures and efficiency / Quantization
LSGQuant: Layer-Sensitivity Guided Quantization for One-Step Diffusion Real-World Video Super-Resolution
Tianxing Wu, Zheng Chen, Cirou Xu, Bowen Chai, Yong Guo, Yutong Liu, Linghe Kong, Yulun Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add DRAQ instead of no scaling (ablation) | LSGQuant (W4A4) · UDM10 | PSNR: 23.72 -> 23.80 | Improved |
| add DRAQ instead of no scaling (ablation) | LSGQuant (W4A4) · UDM10 | DOVER: 0.3996 -> 0.4333 | Improved |
| use static calibrated channel scaling instead of DRAQ (ablation) | LSGQuant (W4A4) · UDM10 | PSNR: 23.80 -> 20.04 | Hurt |
| use static calibrated channel scaling instead of DRAQ (ablation) | LSGQuant (W4A4) · UDM10 | DOVER: 0.4333 -> 0.0231 | Hurt |
| add QAO (ablation) | LSGQuant (W4A4) · UDM10 | PSNR: 23.23 -> 23.80 | Improved |
| add QAO (ablation) | LSGQuant (W4A4) · UDM10 | DISTS: 0.2582 -> 0.2348 | Improved |
| add QAO (ablation) | LSGQuant (W4A4) · UDM10 | DOVER: 0.4162 -> 0.4333 | Improved |
| three-tier VOLTS instead of uniform fully optimized training (ablation) | LSGQuant (W4A4) · UDM10 | PSNR: 23.43 -> 23.80 | Improved |
| three-tier VOLTS instead of uniform lightly adapted training (ablation) | LSGQuant (W4A4) · UDM10 | PSNR: 23.74 -> 23.80 | Improved |
| three-tier VOLTS instead of frozen + fully optimized only (ablation) | LSGQuant (W4A4) · UDM10 | DOVER: 0.4571 -> 0.4333 | Mixed |
| three-tier VOLTS instead of frozen + lightly adapted only (ablation) | LSGQuant (W4A4) · UDM10 | PSNR: 23.75 -> 23.80 | Improved |
| enlarge calibration set from 50 to 200 samples (ablation) | LSGQuant (W4A4) · layer sensitivity thresholds | delta_1,low (x1e-5): 3.03 -> 2.47 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- QuaRot — QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMs (NeurIPS 2024) · better on 6 of 6 shared results
- SmoothQuant — SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models (ICML 2023) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- PassionSR: Post-Training Quantization with Adaptive Scale in One-Step Diffusion based Image Super-Resolution (CVPR 2025 · similarity 0.90)
- 2DQuant: Low-bit Post-Training Quantization for Image Super-Resolution (NeurIPS 2024 · similarity 0.89)
- $\text{S}^2$Q-VDiT: Accurate Quantized Video Diffusion Transformer with Salient Data and Sparse Token Distillation (NeurIPS 2025 · similarity 0.89)
- EfficientDM: Efficient Quantization-Aware Fine-Tuning of Low-Bit Diffusion Models (ICLR 2024 · similarity 0.89)
- Q-DiT: Accurate Post-Training Quantization for Diffusion Transformers (CVPR 2025 · similarity 0.88)
- DOVE: Efficient One-Step Diffusion Model for Real-World Video Super-Resolution (NeurIPS 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
