资源算法MobileNet-v2-pruning

MobileNet-v2-pruning

2020-02-28 | |  157 |   0 |   0

MobileV2 Pruning

This repository aim to try out different pruning-approaches on lightweight Backbones.

Usage

  1. Training

    python main.py --arch MobileNetV2 (for l1norm pruner )
    python main.py --sr --arch MobileNetV2 (for slimming pruner) 
    python main.py --arch USMobileNetV2 (for Autoslim pruner )
  2. Pruning (prune+finetune)

    python prune.py --arch MobileNetV2 --pruner l1normpruner --pruneratio 0.6
    python prune.py --arch MobileNetV2 --pruner SlimmingPruner --sr --pruneratio 0.6
    python prune.py --arch USMobileNetV2 --pruner AutoSlimPruner

Results on Cifar10

BackBonePrunerPrune RatioOriginal/Pruned/Finetuned AccuracyFLOPs(M)Params(M)
MobileV2L1-Norm0.60.937/0.100/0.844313.5->225.52.24->1.15
MobileV2Slimming0.60.922/0.485/0.915313.5->214.52.24->0.98
MobileV2AutoSlim<200 flops0.920/0.561/0.916313.5->199.672.24->0.81

TODO

Pruning Methodsd

Backbones

  •  MobileV2

  •  ShuffleNet
    ....

Reference

rethinking-network-pruning

Pruned-MobileNet_v2


上一篇:tensorflow-MobilenetV1-SSD

下一篇:CoreML_MobileNet

用户评价
全部评价

热门资源

  • Keras-ResNeXt

    Keras ResNeXt Implementation of ResNeXt models...

  • seetafaceJNI

    项目介绍 基于中科院seetaface2进行封装的JAVA...

  • spark-corenlp

    This package wraps Stanford CoreNLP annotators ...

  • capsnet-with-caps...

    CapsNet with capsule-wise convolution Project ...

  • inferno-boilerplate

    This is a very basic boilerplate example for pe...