Huawei uses machine translation combined with human proofreading to translate this document to different languages in order to help you better understand the content of this document.
Note: Even the most advanced machine translation cannot match the quality of professional translators.
Huawei shall not bear any responsibility for translation accuracy and it is recommended that you refer to the English document (a link for which has been provided).
Preparations
Obtaining the Dataset
This sample uses the ImageNet dataset as an example. Download the dataset from http://www.image-net.org/.
About ResNet-50
ResNet-50 is a deep residual network that can be used to classify 1000 classes of the CIFAR-10 and ImageNet datasets.
Obtaining the Original Model
The original ResNet network script is available at https://github.com/tensorflow/models/tree/r2.1_model_reference/official.
Directory Structure
The directory is organized as follows. (Only some involved files are listed. For more files, see the original ResNet script.)
├── r1 // Original model directory. │ ├── resnet // ResNet main directory. │ ├── __init__.py │ ├── imagenet_main.py // Script for training the network based on the ImageNet dataset. │ ├── imagenet_preprocessing.py // ImageNet preprocessing module. │ ├── resnet_model.py // ResNet model file. │ ├── resnet_run_loop.py // Data input processing and run loop (training, validation, and test). │ ├── README.md // Project description file. │ ├── utils │ │ ├── export.py // Data receive functions, which define the parameter format that the exported model can respond to. ├── utils │ ├── flags │ │ ├── core.py // Public APIs including the parameter definition. │ ├── logs │ │ ├── hooks_helper.py // Tool used for custom model tests and training, such as the function for calculating the number of steps per second, and the function of capturing CPU/GPU analysis information. │ │ ├── logger.py // Log tool. │ ├── misc │ │ ├── distribution_utils.py // Auxiliary functions used for running models in distributed mode. │ │ ├── model_helpers.py // Functions that can be called by models, such as functions that control whether models stop.