Research map / Image generation / Fast diffusion sampling
Q-DM: An Efficient Low-bit Quantized Diffusion Model
Yanjing Li, Sheng Xu, Xianbin Cao, Xiao Sun, Baochang Zhang
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 timestep-aware quantization (TaQ) (ablation) | Quantized DM baseline (LSQ) 4-bit · CIFAR-10 DDIM 50 steps | FID: 10.22 -> 9.25 | Improved |
| Add noise-estimating mimicking (NeM) (ablation) | Quantized DM baseline (LSQ) 4-bit · CIFAR-10 DDIM 50 steps | FID: 10.22 -> 8.98 | Improved |
| Add TaQ and NeM together (Q-DM) (ablation) | Quantized DM baseline (LSQ) 4-bit · CIFAR-10 DDIM 50 steps | FID: 10.22 -> 6.89 | Improved |
| Add TaQ (ablation) | Quantized DM baseline 3-bit · CIFAR-10 DDIM 50 steps | FID: 13.24 -> 11.19 | Improved |
| Add NeM (ablation) | Quantized DM baseline 3-bit · CIFAR-10 DDIM 50 steps | FID: 13.24 -> 11.02 | Improved |
| Add TaQ and NeM together (ablation) | Quantized DM baseline 3-bit · CIFAR-10 DDIM 50 steps | FID: 13.24 -> 9.07 | Improved |
| Add TaQ (ablation) | Quantized DM baseline 2-bit · CIFAR-10 DDIM 50 steps | FID: 18.74 -> 16.83 | Improved |
| Add NeM (ablation) | Quantized DM baseline 2-bit · CIFAR-10 DDIM 50 steps | FID: 18.74 -> 16.97 | Improved |
| Add TaQ and NeM together (ablation) | Quantized DM baseline 2-bit · CIFAR-10 DDIM 50 steps | FID: 18.74 -> 15.26 | Improved |
| Q-DM versus LSQ baseline at 4-bit with 1000-step DDPM | Quantized DM · CIFAR-10 DDPM 1000 steps | FID: 9.11 -> 5.17 | Improved |
| Q-DM versus LSQ baseline at 4-bit with 50-step DDIM | Quantized DM · ImageNet 64x64 | FID: 24.78 -> 20.02 | Improved |
| 4-bit Q-DM versus 8-bit PTQ4DM | Quantized DM · CIFAR-10 DDIM 50 steps | FID: 18.02 -> 6.89 | Improved |
Closest papers by meaning
- MPQ-DM: Mixed Precision Quantization for Extremely Low Bit Diffusion Models (AAAI 2025 · similarity 0.93)
- D^2-DPM: Dual Denoising for Quantized Diffusion Probabilistic Models (AAAI 2025 · similarity 0.93)
- Qua2SeDiMo: Quantifiable Quantization Sensitivity of Diffusion Models (AAAI 2025 · similarity 0.93)
- Optimizing Quantized Diffusion Models via Distillation with Cross-Timestep Error Correction (AAAI 2025 · similarity 0.92)
- EfficientDM: Efficient Quantization-Aware Fine-Tuning of Low-Bit Diffusion Models (ICLR 2024 · similarity 0.92)
- AccuQuant: Simulating Multiple Denoising Steps for Quantizing Diffusion Models (NeurIPS 2025 · similarity 0.92)
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.
