AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

Back Razor: Memory-Efficient Transfer Learning by Self-Sparsified Backpropagation

Ziyu Jiang, Xuxi Chen, Xueqin Huang, Xianzhi Du, Denny Zhou, Zhangyang Wang

NeurIPS 2022 · Architectures and efficiency › Pruning and efficient tuning · approach: Joint pruning and quantization · 1 citations (OpenAlex)

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 changedWhereMeasuredEffect
prune activations only in backward instead of both forward and backward (ablation)Back Razor · Aircraft / Pets (CNN)top-1 accuracy: accuracy below 5% at 70% pruning for symmetric pruning vs high accuracy at 99% for backward-onlyImproved
raise sparsity from 90% to 99% (ablation)Back Razor · Petstop-1 accuracy: 92.9 -> 92.5No clear effect
raise sparsity from 90% to 99% (ablation)Back Razor · Aircrafttop-1 accuracy: 87.6 -> 80.7Hurt
raise sparsity from 90% to 97% (ablation)Back Razor · Aircrafttop-1 accuracy: 87.6 -> 85.2Hurt
raise sparsity from 90% to 99% (memory) (ablation)Back Razor · CIFAR100 CNNtraining memory: 42MB -> 39MBNo clear effect
replace TinyTL@320 with Back Razor@90%ProxylessNAS-Mobile · Aircrafttop-1 accuracy: 82.3 -> 87.6Improved
replace FT-Norm+Last with Back Razor@90%ProxylessNAS-Mobile · Aircrafttop-1 accuracy: 68.1 -> 87.6Improved
replace SWAT@80% with Back Razor@90%ResNet-18 · CIFAR100top-1 accuracy: 79.6 -> 82.2Improved
replace SWAT@80% with Back Razor@90%ResNet-18 · CIFAR10top-1 accuracy: 95.8 -> 96.4Improved
raise ViT sparsity from 80% to 95% (ablation)Back Razor · Aircraft (ViT-B/16)top-1 accuracy: 77.3 -> 74.3Hurt
raise ViT sparsity from 80% to 95% (ablation)Back Razor · CIFAR10 (ViT-B/16)top-1 accuracy: 98.9 -> 98.8No clear effect
combine Back Razor@80% with Mesa quantization (ablation)Back Razor · CIFAR100 (ViT-B/16)training memory: 4565MB -> 2937MBImproved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy