mindcv_fixed/configs/mobilenetv2

• MobileNetV2
• • Introduction
  • Results
  • • • Notes
  • Quick Start
  • • Preparation
    • • Installation
      • Dataset Preparation
    • Training
    • Validation
    • Deployment (optional)

MobileNetV2

---

MobileNetV2: Inverted Residuals and Linear Bottlenecks

Introduction

---

The model is a new neural network architecture that is specifically tailored for mobile and resource-constrained environments.
This network pushes the state of the art for mobile custom computer vision models, significantly reducing the amount of operations and memory required while maintaining the same accuracy.

The main innovation of the model is the proposal of a new layer module: The Inverted Residual with Linear Bottleneck. The module takes as input a low-dimensional compressed representation that is first extended to high-dimensionality and then filtered with lightweight depth convolution.
Linear convolution is then used to project the features back to the low-dimensional representation.

Results

---

Model	Context	Top-1 (%)	Top-5 (%)	Params (M)	Train T.	Infer T.	Download	Config	Log
MobileNet_v2_075	D910x8-G	69.76	89.28	2.66	106s/epoch		model	cfg	log
MobileNet_v2_100	D910x8-G	72.02	90.92	3.54	98s/epoch		model	cfg	log
MobileNet_v2_140	D910x8-G	74.97	92.32	6.15	157s/epoch		model	cfg	log

Notes

• All models are trained on ImageNet-1K training set and the top-1 accuracy is reported on the validatoin set.
• Context: GPU_TYPE x pieces - G/F, G - graph mode, F - pynative mode with ms function.

Quick Start

---

Preparation

Installation

Please refer to the installation instruction in MindCV.

Dataset Preparation

Please download the ImageNet-1K dataset for model training and validation.

Training

• Hyper-parameters. The hyper-parameter configurations for producing the reported results are stored in the yaml files in mindcv/configs/mobilenetv1 folder. For example, to train with one of these configurations, you can run:

  # train mobilenetv2 on 8 GPUs
  export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
  mpirun -n 8 python train.py -c configs/mobilenetv2/mobilenetv2_100_gpu.yaml --data_dir /path/to/imagenet
  

  Note that the number of GPUs/Ascends and batch size will influence the training results. To reproduce the training result at most, it is recommended to use the same number of GPUs/Ascneds with the same batch size.

• Finetuning. Here is an example for finetuning a pretrained mobilenetv2_100 on CIFAR10 dataset using Momentum optimizer.

  python train.py -c configs/mobilenetv2/mobilenetv2_100_gpu.yaml --data_dir /path/to/imagenet
  

Detailed adjustable parameters and their default value can be seen in config.py.

Validation

• To validate the trained model, you can use validate.py. Here is an example for mobilenet_100 to verify the accuracy of
  pretrained weights.

  python validate.py -c /path/to/val.yaml --data_dir /path/to/imagenet
  

• To validate the model, you can use validate.py. Here is an example for mobilenetv1_100 to verify the accuracy of your
  training.

  python validate.py -c /path/to/val.yaml --data_dir /path/to/imagenet --dataset=imagenet --val_split=val --ckpt_path='./ckpt/mobilenet_v2_100_224-200_625.ckpt'
  

Deployment (optional)

Please refer to the deployment tutorial in MindCV.