Research map / Architectures and efficiency / Quantization
FracBits: Mixed Precision Quantization via Fractional Bit-Widths
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use fractional-bit differentiable mixed-precision search instead of uniform precision | SAT on MobileNet V1 · ImageNet, 3-bit computation-constrained | top-1 accuracy: 67.1 -> 68.7 | Helped |
| Use fractional-bit differentiable mixed-precision search instead of uniform precision | SAT on MobileNet V2 · ImageNet, 3-bit computation-constrained | top-1 accuracy: 67.2 -> 67.8 | Helped |
| Use fractional-bit differentiable mixed-precision search instead of uniform precision | SAT on ResNet18 · ImageNet, 4-bit computation-constrained | top-1 accuracy: 70.3 -> 70.6 | Helped |
| Use fractional-bit differentiable mixed-precision search instead of uniform precision | SAT on MobileNet V1 · ImageNet, 2-bit model-size-constrained | top-1 accuracy: 66.3 -> 69.7 | Helped |
| Use fractional-bit differentiable mixed-precision search instead of uniform precision | SAT on MobileNet V2 · ImageNet, 2-bit model-size-constrained | top-1 accuracy: 66.8 -> 69.9 | Helped |
| Use kernel-wise rather than layer-wise precision assignment | FracBits-SAT on MobileNet V2 · ImageNet, 3-bit computation-constrained | top-1 accuracy: 67.8 -> 68.2 | Helped |
Closest papers by meaning
- One-Shot Model for Mixed-Precision Quantization (CVPR 2023 · similarity 0.89)
- MetaMix: Meta-State Precision Searcher for Mixed-Precision Activation Quantization (AAAI 2024 · similarity 0.88)
- Efficient Multi-bit Quantization Network Training via Weight Bias Correction and Bit-wise Coreset Sampling (NeurIPS 2025 · similarity 0.88)
- Retraining-Free Model Quantization via One-Shot Weight-Coupling Learning (CVPR 2024 · similarity 0.87)
- Learning from Loss Landscape: Generalizable Mixed-Precision Quantization via Adaptive Sharpness-Aware Gradient Aligning (ICML 2025 · similarity 0.87)
- AMPA: Adaptive Mixed Precision Allocation for Low-Bit Integer Training (ICML 2024 · similarity 0.86)
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.