Research map / Architectures and efficiency / Pruning and efficient tuning
AutoLR: Layer-wise Pruning and Auto-tuning of Learning Rates in Fine-tuning of Deep Networks
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 |
|---|---|---|---|
| Add layer-wise learning-rate auto-tuning to pruning Layers 15 and 16 (ablation) | ResNet-50 fine-tuning · CUB-200 | Recall@1: 67.00 -> 70.10 | Helped |
| Add layer-wise learning-rate auto-tuning to pruning Layers 15 and 16 (ablation) | ResNet-50 fine-tuning · Cars-196 | Recall@1: 87.85 -> 89.02 | Helped |
| Add layer-wise learning-rate auto-tuning to pruning Layers 15 and 16 (ablation) | ResNet-50 fine-tuning · SOP | Recall@1: 83.31 -> 84.24 | Helped |
| Add layer-wise learning-rate auto-tuning to pruning Layers 15 and 16 (ablation) | ResNet-50 fine-tuning · Inshop | Recall@1: 88.62 -> 91.72 | Helped |
| Prune Layers 15 and 16 instead of retaining all layers (ablation) | ResNet-50 with shared learning rate · CUB-200 | Recall@1: 63.88 -> 67.00 | Helped |
| Prune Layers 15 and 16 instead of retaining all layers (ablation) | ResNet-50 with shared learning rate · Cars-196 | Recall@1: 85.89 -> 87.85 | Helped |
| Prune Layers 15 and 16 instead of retaining all layers (ablation) | ResNet-50 with shared learning rate · SOP | Recall@1: 80.02 -> 83.31 | Helped |
| Prune Layers 15 and 16 instead of retaining all layers (ablation) | ResNet-50 with shared learning rate · Inshop | Recall@1: 87.64 -> 88.62 | Helped |
| Prune Layer 16 (ablation) | ResNet-50 fine-tuning · CUB-200 | Recall@1: 63.49 -> 66.95 | Helped |
| Prune Layers 15 and 16 (ablation) | ResNet-50 fine-tuning · CUB-200 | Recall@1: 63.49 -> 67.00 | Helped |
| Also prune Layer 14 (ablation) | ResNet-50 with Layers 15 and 16 pruned · CUB-200 | Recall@1: 67.00 -> 51.00 | Hurt |
| Reduce learning rates in Layers 1, 2, and 3 to 1/10 (ablation) | ResNet-50 with Layers 15 and 16 pruned · CUB-200 | Recall@1: 67.00 -> 67.15 | Helped |
Closest papers by meaning
- Growing a Brain: Fine-Tuning by Increasing Model Capacity (CVPR 2017 · similarity 0.87)
- Auto-Balanced Filter Pruning for Efficient Convolutional Neural Networks (AAAI 2018 · similarity 0.86)
- Pay Attention to Small Weights (NeurIPS 2025 · similarity 0.85)
- The Unreasonable Ineffectiveness of the Deeper Layers (ICLR 2025 · similarity 0.85)
- Reducing Fine-Tuning Memory Overhead by Approximate and Memory-Sharing Backpropagation (ICML 2024 · similarity 0.84)
- Tuning Stable Rank Shrinkage: Aiming at the Overlooked Structural Risk in Fine-tuning (CVPR 2024 · similarity 0.84)
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.