MobileNetV1¶
Introduction¶
Compared with the traditional convolutional neural network, MobileNetV1’s parameters and the amount of computation are greatly reduced on the premise that the accuracy rate is slightly reduced. (Compared to VGG16, the accuracy rate is reduced by 0.9%, but the model parameters are only 1/32 of VGG). The model is based on a streamlined architecture that uses depthwise separable convolutions to build lightweight deep neural networks. At the same time, two simple global hyperparameters are introduced, which can effectively trade off latency and accuracy.[1]
Figure 1. Architecture of MobileNetV1 [1]
Results¶
Our reproduced model performance on ImageNet-1K is reported as follows.
| Model | Context | Top-1 (%) | Top-5 (%) | Params (M) | Recipe | Download |
|---|---|---|---|---|---|---|
| MobileNet_v1_025 | D910x8-G | 54.64 | 78.29 | 0.47 | yaml | weights |
| MobileNet_v1_050 | D910x8-G | 66.39 | 86.71 | 1.34 | yaml | weights |
| MobileNet_v1_075 | D910x8-G | 70.66 | 89.49 | 2.60 | yaml | weights |
| MobileNet_v1_100 | D910x8-G | 71.83 | 90.26 | 4.25 | yaml | weights |
Notes¶
Context: Training context denoted as {device}x{pieces}-{MS mode}, where mindspore mode can be G - graph mode or F - pynative mode with ms function. For example, D910x8-G is for training on 8 pieces of Ascend 910 NPU using graph mode.
Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K.
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¶
Distributed Training
It is easy to reproduce the reported results with the pre-defined training recipe. For distributed training on multiple Ascend 910 devices, please run
# distrubted training on multiple GPU/Ascend devices
mpirun -n 8 python train.py --config configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml --data_dir /path/to/imagenet
If the script is executed by the root user, the
--allow-run-as-rootparameter must be added tompirun.
Similarly, you can train the model on multiple GPU devices with the above mpirun command.
For detailed illustration of all hyper-parameters, please refer to config.py.
Note: As the global batch size (batch_size x num_devices) is an important hyper-parameter, it is recommended to keep the global batch size unchanged for reproduction or adjust the learning rate linearly to a new global batch size.
Standalone Training
If you want to train or finetune the model on a smaller dataset without distributed training, please run:
# standalone training on a CPU/GPU/Ascend device
python train.py --config configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml --data_dir /path/to/dataset --distribute False
Validation¶
To validate the accuracy of the trained model, you can use validate.py and parse the checkpoint path with --ckpt_path.
python validate.py -c configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt
Deployment¶
Please refer to the deployment tutorial in MindCV.
References¶
[1] Howard A G, Zhu M, Chen B, et al. Mobilenets: Efficient convolutional neural networks for mobile vision applications[J]. arXiv preprint arXiv:1704.04861, 2017.