代码如下:
import paddle
from paddle.nn import Linear
import paddle.nn.functional as F
import os
import numpy as np
import matplotlib.pyplot as plt
from PIL import Image
# train_dataset = paddle.vision.datasets.MNIST(mode='train')
# train_data0 = np.array(train_dataset[0][0])
# train_label_0 = np.array(train_dataset[0][1])
# plt.figure("Image")
# plt.figure(figsize=(2,2))
# plt.imshow(train_data0, cmap = plt.cm.binary)
# plt.axis('on')
# plt.title('image')
# plt.show()
#
# print("图像数据形状和对应的数据为:",train_data0.shape)
# print("图像的标签形状和对应的数据为:", train_label_0.shape, train_data0)
# print("\n打印第一个batch的第一个图像,对应标签数字为{}".format(train_label_0))
class MNIST(paddle.nn.Layer):
def __init__(self):
super(MNIST, self).__init__()
self.fc = paddle.nn.Linear(in_features=784, out_features=1)
def forward(self, inputs):
outputs = self.fc(inputs)
return outputs
model = MNIST()
def train(model):
model.train()
train_loader = paddle.io.DataLoader(paddle.vision.datasets.MNIST(mode='train'),
batch_size=16,
shuffle=True)
opt = paddle.optimizer.SGD(learning_rate=0.001, parameters=model.parameters())
def norm_img(img):
assert len(img.shape) == 3
batch_size, img_h, img_w = img.shape[0], img.shape[1], img.shape[2]
img = img / 255
img = paddle.reshape(img, [batch_size, img_h * img_w])
return img
paddle.vision.set_image_backend('cv2')
model = MNIST()
def train(model):
model.train()
train_loader = paddle.io.DataLoader(paddle.vision.datasets.MNIST(mode='train'),
batch_size=16,
shuffle=True)
opt = paddle.optimizer.SGD(learning_rate=0.001, parameters=model.parameters())
EPOCH_NUM = 10
for epoch in range(EPOCH_NUM):
for batch_id, data in enumerate(train_loader()):
images = norm_img(data[0]).astype('float32')
labels = data[1].astype('float32')
predicts = model(images)
loss = F.square_error_cost(predicts, labels)
avg_loss = paddle.mean(loss)
if batch_id % 1000 == 0:
print("epoch_id: {}, batch_id: {}, loss is: {}".format(epoch, batch_id, avg_loss.numpy()))
avg_loss.backward()
opt.step()
opt.clear_grad()
train(model)
paddle.save(model.state_dict(), './mnist.pdparams')
img_path = './work/example_0.jpg'
im = Image.open('./work/example_0.jpg')
plt.imshow(im)
plt.show()
im = im.convert('L')
print('原始图像shape: ', np.array(im).shape)
im = im.resize((28, 28), Image.ANTIALIAS)
plt.imshow(im)
plt.show()
print("采样后图片shape: ", np.array(im).shape)
def load_image(img_path):
im = Image.open(img_path).convert('L')
im = im.resize((28, 28), Image.ANTIALIAS)
im = np.array(im).reshape(1, -1).astype(np.float32)
im = 1 - im / 255
return im
model = MNIST()
params_file_path = 'mnist.pdparams'
img_path = './work/example_0.jpg'
param_dict = paddle.load(params_file_path)
model.load_dict(param_dict)
model.eval()
tensor_img = load_image(img_path)
result = model(paddle.to_tensor(tensor_img))
print('result',result)
print("本次预测的数字是", result.numpy().astype('int32'))
输出结果:
import imp
epoch_id: 0, batch_id: 0, loss is: [19.691336]
epoch_id: 0, batch_id: 1000, loss is: [4.8562613]
epoch_id: 0, batch_id: 2000, loss is: [3.5893354]
epoch_id: 0, batch_id: 3000, loss is: [3.6005962]
epoch_id: 1, batch_id: 0, loss is: [7.029731]
epoch_id: 1, batch_id: 1000, loss is: [6.2418976]
epoch_id: 1, batch_id: 2000, loss is: [2.5513744]
epoch_id: 1, batch_id: 3000, loss is: [1.754514]
epoch_id: 2, batch_id: 0, loss is: [4.9103928]
epoch_id: 2, batch_id: 1000, loss is: [5.4251204]
epoch_id: 2, batch_id: 2000, loss is: [4.564935]
epoch_id: 2, batch_id: 3000, loss is: [2.9991786]
epoch_id: 3, batch_id: 0, loss is: [2.1028972]
epoch_id: 3, batch_id: 1000, loss is: [2.3331792]
epoch_id: 3, batch_id: 2000, loss is: [4.2814217]
epoch_id: 3, batch_id: 3000, loss is: [3.112694]
epoch_id: 4, batch_id: 0, loss is: [3.1716564]
epoch_id: 4, batch_id: 1000, loss is: [3.0454333]
epoch_id: 4, batch_id: 2000, loss is: [2.6289668]
epoch_id: 4, batch_id: 3000, loss is: [1.4390972]
epoch_id: 5, batch_id: 0, loss is: [6.6636868]
epoch_id: 5, batch_id: 1000, loss is: [3.5302186]
epoch_id: 5, batch_id: 2000, loss is: [3.886701]
epoch_id: 5, batch_id: 3000, loss is: [2.966082]
epoch_id: 6, batch_id: 0, loss is: [6.234954]
epoch_id: 6, batch_id: 1000, loss is: [5.239113]
epoch_id: 6, batch_id: 2000, loss is: [4.291875]
epoch_id: 6, batch_id: 3000, loss is: [2.9170785]
epoch_id: 7, batch_id: 0, loss is: [2.8740573]
epoch_id: 7, batch_id: 1000, loss is: [1.9314649]
epoch_id: 7, batch_id: 2000, loss is: [7.5295153]
epoch_id: 7, batch_id: 3000, loss is: [1.8619927]
epoch_id: 8, batch_id: 0, loss is: [2.6857028]
epoch_id: 8, batch_id: 1000, loss is: [3.0323493]
epoch_id: 8, batch_id: 2000, loss is: [3.6132627]
epoch_id: 8, batch_id: 3000, loss is: [3.0137868]
epoch_id: 9, batch_id: 0, loss is: [4.07297]
epoch_id: 9, batch_id: 1000, loss is: [4.682923]
epoch_id: 9, batch_id: 2000, loss is: [4.087107]
epoch_id: 9, batch_id: 3000, loss is: [3.667269]
result Tensor(shape=[1, 1], dtype=float32, place=CPUPlace, stop_gradient=False,
[[1.11405170]])
本次预测的数字是 [[1]]