Research map / Image generation / Fast diffusion sampling
Beyond Uniformity: Sample and Frequency Meta Weighting for Post-Training Quantization of Diffusion Models
Cuong Pham, Dung Anh Hoang, Cuong C. Nguyen, Trung Le, Dinh Phung, Gustavo Carneiro, Thanh-Toan Do
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 sample weighting to TFMQ-DM (ablation) | TFMQ-DM · LSUN-Bedrooms W4A8 | FID: 3.68 -> 3.47 | Improved |
| add frequency weighting to TFMQ-DM (ablation) | TFMQ-DM · LSUN-Bedrooms W4A8 | FID: 3.68 -> 3.38 | Improved |
| combine sample and frequency weighting (ablation) | Ours · LSUN-Bedrooms W4A8 | FID: 3.47 -> 3.28 | Improved |
| add regularization LReg (ablation) | Ours · LSUN-Bedrooms W4A8 | FID: 3.41 -> 3.28 | Improved |
| use DWT instead of FFT frequency transform (ablation) | Ours · LSUN-Bedrooms W4A8 | FID: 3.45 -> 3.28 | Improved |
| gamma 0.1 instead of 1.0 (ablation) | Ours · CIFAR-10 W4A32 | FID: 4.71 -> 4.21 | Improved |
| beta 0.05 instead of 0.1 (ablation) | Ours · CIFAR-10 W4A32 | FID: 4.73 -> 4.21 | Improved |
| real noised images as validation set instead of generated (ablation) | Ours · LSUN-Bedrooms W4A8 | FID: 3.28 -> 3.26 | No clear effect |
| alternative frequency loss on estimated final samples (ablation) | Ours · CIFAR-10 W4A32 | FID: 4.21 -> 4.29 | No clear effect |
| meta weighting vs TFMQ-DM | Ours · FFHQ W4A32 | FID: 9.89 -> 9.20 | Improved |
| meta weighting vs TFMQ-DM | Ours · CIFAR-10 W4A32 | FID: 4.73 -> 4.21 | Improved |
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.
- TFMQ-DM — TFMQ-DM: Temporal Feature Maintenance Quantization for Diffusion Models (CVPR 2024) · better on 8 of 8 shared results
- TCAQ-DM — TCAQ-DM: Timestep-Channel Adaptive Quantization for Diffusion Models (AAAI 2025) · better on 2 of 2 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
- AccuQuant: Simulating Multiple Denoising Steps for Quantizing Diffusion Models (NeurIPS 2025 · similarity 0.92)
- TFMQ-DM: Temporal Feature Maintenance Quantization for Diffusion Models (CVPR 2024 · similarity 0.90)
- Towards Accurate Post-training Quantization for Diffusion Models (CVPR 2024 · similarity 0.90)
- EfficientDM: Efficient Quantization-Aware Fine-Tuning of Low-Bit Diffusion Models (ICLR 2024 · similarity 0.90)
- Qua2SeDiMo: Quantifiable Quantization Sensitivity of Diffusion Models (AAAI 2025 · similarity 0.89)
- Q-DiT: Accurate Post-Training Quantization for Diffusion Transformers (CVPR 2025 · similarity 0.89)
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.
