Research map / Architectures and efficiency / Pruning and efficient tuning
Performance-Aware Approximation of Global Channel Pruning for Multitask CNNs
Hancheng Ye, Bo Zhang, Tao Chen, Jiayuan Fan, Bin Wang
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 |
|---|---|---|---|
| use sequentially greedy pruning instead of layer-wise pruning (ablation) | PAGCP SSD300 · PASCAL VOC | mAP (%): 73.2 -> 75.6 | Improved |
| use sequentially greedy pruning instead of one-shot global pruning (ablation) | PAGCP SSD300 · PASCAL VOC | mAP (%): 74.3 -> 75.6 | Improved |
| use sequentially greedy pruning instead of group pruning (ablation) | PAGCP SSD300 · PASCAL VOC | mAP (%): 75.2 -> 75.6 | Improved |
| use max over tasks instead of min in oracle criterion (ablation) | PAGCP SSD300 · PASCAL VOC | mAP (%): 49.0 -> 75.9 | Improved |
| use max instead of l1 sum in oracle criterion (ablation) | PAGCP SSD300 · PASCAL VOC | mAP (%): 75.4 -> 75.9 | Improved |
| use max instead of l2 in oracle criterion (ablation) | PAGCP SSD300 · PASCAL VOC | mAP (%): 75.6 -> 75.9 | Improved |
| include all four subtasks in the performance-aware criterion instead of initial segmentation only (ablation) | PAGCP MTI-Net · NYUD-v2 | mIoU (%): 35.9 -> 37.4 | Improved |
| run more pruning iterations on SSD300 (ablation) | PAGCP · PASCAL VOC | mAP (%): 76.2 -> 75.6 | Hurt |
| prune with PAGCP instead of scaling down to YOLOv5s | YOLOv5m · PASCAL VOC | mAP (%): 79.5 -> 81.7 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ThiNet — ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression (ICCV 2017) · better on 1 of 1 shared result
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
- Dynamic Structure Pruning for Compressing CNNs (AAAI 2023 · similarity 0.88)
- Rethinking the Pruning Criteria for Convolutional Neural Network (NeurIPS 2021 · similarity 0.88)
- DiSparse: Disentangled Sparsification for Multitask Model Compression (CVPR 2022 · similarity 0.88)
- Discrimination-aware Network Pruning for Deep Model Compression (TPAMI 2021 · similarity 0.88)
- REPrune: Channel Pruning via Kernel Representative Selection (AAAI 2024 · similarity 0.87)
- MULTIFLOW: Shifting Towards Task-Agnostic Vision-Language Pruning (CVPR 2024 · similarity 0.87)
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.
