|
-
- # ============================================================================
- """
- ######################## inference lenet example ########################
- inference lenet according to model file
- """
- """
- ######################## 推理环境使用说明 ########################
-
- """
- import os
- import argparse
- import moxing as mox
- import mindspore.nn as nn
- from mindspore import context
- from mindspore.train.serialization import load_checkpoint, load_param_into_net
- from mindspore.train import Model
- from mindspore.nn.metrics import Accuracy
- from mindspore import Tensor
- import matplotlib.pyplot as plt
- import numpy as np
- from glob import glob
- from dataset import create_dataset
- from config import mnist_cfg as cfg
- from lenet import LeNet5
-
- if __name__ == "__main__":
- parser = argparse.ArgumentParser(description='MindSpore Lenet Example')
- parser.add_argument('--device_target', type=str, default="Ascend", choices=['Ascend', 'GPU', 'CPU'],
- help='device where the code will be implemented (default: Ascend)')
- parser.add_argument('--data_path', type=str, default="./Data",
- help='path where the dataset is saved')
- parser.add_argument('--ckpt_path', type=str, default="", help='if mode is test, must provide\
- path where the trained ckpt file')
- parser.add_argument('--ckpt_name', type=str, default="", help='the ckpt file name')
- parser.add_argument('--data_url',
- type=str,
- default="./Data",
- help='path where the dataset is saved')
- parser.add_argument('--train_url',
- help='model folder to save/load',
- default='./model')
- parser.add_argument('--result_url',
- help='model folder to save/load',
- default='./result')
-
- args = parser.parse_args()
-
- # Copy obs_file to local
- obs_data_url = args.data_url
- args.data_url = '/home/work/user-job-dir/inputs/data/'
- # args.data_url = '/home/ma-user/work/data/'
- obs_train_url = args.train_url
- args.train_url = '/home/work/user-job-dir/outputs/model/'
- # args.train_url = '/home/ma-user/work/model/'
- args.train_url = '/home/work/user-job-dir/outputs/model/'
- # args.ckpt_path = '/home/ma-user/work/'
- obs_ckpt_path = args.ckpt_path
-
- obs_result_url = args.result_url
- args.result_url = '/home/work/user-job-dir/outputs/result/'
- try:
- # 对文件夹进行操作,请使用mox.file.copy_parallel。如果拷贝一个文件。请使用mox.file.copy对文件操作,
- mox.file.copy_parallel(obs_data_url, args.data_url)
- print("Successfully Download {} to {}".format(obs_data_url,
- args.data_url))
- except Exception as e:
- print('moxing download {} to {} failed: '.format(
- obs_data_url, args.data_url) + str(e))
- try:
- mox.file.copy_parallel(obs_ckpt_path, args.ckpt_path)
- print("Successfully Download {} to {}".format(obs_ckpt_path,
- args.ckpt_path))
- except Exception as e:
- print('moxing download {} to {} failed: '.format(
- obs_ckpt_path, args.ckpt_path) + str(e))
-
- args.dataset_path = args.data_url
- args.save_checkpoint_path = args.train_url
-
- context.set_context(mode=context.GRAPH_MODE, device_target=args.device_target)
-
- network = LeNet5(cfg.num_classes)
- net_loss = nn.SoftmaxCrossEntropyWithLogits(sparse=True, reduction="mean")
- repeat_size = cfg.epoch_size
- net_opt = nn.Momentum(network.trainable_params(), cfg.lr, cfg.momentum)
- model = Model(network, net_loss, net_opt, metrics={"Accuracy": Accuracy()})
-
- print("============== Starting Testing ==============")
- print("ckpt_path is:{}", glob(os.path.join(args.ckpt_path, "*.ckpt")))
- args.load_ckpt_path = os.path.join(args.ckpt_path, args.ckpt_name)
- print("args.load_ckpt_path is:{}", args.load_ckpt_path )
- param_dict = load_checkpoint(args.load_ckpt_path )
- load_param_into_net(network, param_dict)
- # 定义测试数据集,batch_size设置为1,则取出一张图片
- ds_test = create_dataset(os.path.join(args.dataset_path, "test"), batch_size=1).create_dict_iterator()
- data = next(ds_test)
-
- # images为测试图片,labels为测试图片的实际分类
- images = data["image"].asnumpy()
- labels = data["label"].asnumpy()
- print('Tensor:', Tensor(data['image']))
-
- # 使用函数model.predict预测image对应分类
- output = model.predict(Tensor(data['image']))
- predicted = np.argmax(output.asnumpy(), axis=1)
- pred = np.argmax(output.asnumpy(), axis=1)
-
- # 输出预测分类与实际分类
- print(f'Predicted: "{predicted[0]}", Actual: "{labels[0]}"')
- # err_num = []
- # index = 1
- # for i in range(len(labels)):
- # plt.subplot(4, 8, i+1)
- # color = 'blue' if pred[i] == labels[i] else 'red'
- # plt.title("pre:{}".format(pred[i]), color=color)
- # plt.imshow(np.squeeze(images[i]))
- # plt.axis("off")
- # if color == 'red':
- # index = 0
- # print("Row {}, column {} is incorrectly identified as {}, the correct value should be {}".format(int(i/8)+1, i%8+1, pred[i], labels[i]), '\n')
- # if index:
- # print("All the figures in this group are predicted correctly!")
- # print(pred, "<--Predicted figures")
- # print(labels, "<--The right number")
- # plt.savefig(os.path.join(args.result_url, "pre.png"))
- # plt.show()
-
-
- # Upload results to obs
- ######################## 将输出的结果拷贝到obs(固定写法) ########################
- # 把训练后的模型数据从本地的运行环境拷贝回obs,在启智平台相对应的训练任务中会提供下载
- try:
- mox.file.copy_parallel(args.result_url, obs_result_url)
- print("Successfully Upload {} to {}".format(args.result_url, obs_result_url))
- except Exception as e:
- print('moxing upload {} to {} failed: '.format(args.result_url, obs_result_url) + str(e))
- ######################## 将输出的模型拷贝到obs ########################
|