资源分类
》 资源 》算法》MobileNetV3-MXNET-insightface

MobileNetV3-MXNET-insightface

2020-02-28 | |  40 |   0 |   0

MobileNetV3-MXNET-insightface

How to use

pip insatll mxnet-cu90 (updated to version 1.5.0)

cp fmobilenetv3.py ../insightface/src/symbols/

cp m3_train_softmax.py ../insight/src/

train:

cd ../insightface/src/

export MXNET_CPU_WORKER_NTHREADS=24

export MXNET_ENGINE_TYPE=ThreadedEnginePerDevice

CUDA_VISIBLE_DEVICES='0,1,2,3' python -u m3_train_softmax.py --data-dir your_data_dir --network l3 --loss-type 0 --prefix your_savemodel_dir --per-batch-size 128

Solve the problem of model oversize

cd ../insightface/deploy/

python -u model-slim.py --model your_savemodel_dir/,epoch

Reference

https://github.com/deepinsight/insightface

https://github.com/AmigoCDT/MXNet-MobileNetV3

https://github.com/wsqshiqing/MobileNetV3


上一篇：TensorFlow_MobileNetV2_PortraitMatting

下一篇：Network-Analysis-Made-Simple

用户评价
全部评价

热门资源

  • 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 ...

  • Keras-ResNeXt

    Keras ResNeXt Implementation of ResNeXt models...

  • capsnet-with-caps...

    CapsNet with capsule-wise convolution Project ...

  • inferno-boilerplate

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