flask与深度学习

Question

使用flask将LeNet5模型的运行结果和测试结果上传web

Answer 1

我这里有一个示例，你可以参考借鉴思路

from flask import Flask, request, jsonify
import pandas as pd
import torch
from torch import nn
from torch.utils.data import DataLoader
from torchvision import transforms, datasets
import torch.optim as optim

app = Flask(__name__)

# 定义LeNet5模型
class LeNet5(nn.Module):
    def __init__(self):
        super(LeNet5, self).__init__()
        self.conv1 = nn.Conv2d(1, 6, 5)
        self.conv2 = nn.Conv2d(6, 16, 5)
        self.fc1 = nn.Linear(256, 120)
        self.fc2 = nn.Linear(120, 84) 
        self.fc3 = nn.Linear(84, 10)

    def forward(self, x):
        x = F.max_pool2d(F.relu(self.conv1(x)), (2, 2)) 
        x = F.max_pool2d(F.relu(self.conv2(x)), 2) 
        x = x.view(x.size(0), -1)
        x = F.relu(self.fc1(x))
        x = F.relu(self.fc2(x))
        x = self.fc3(x)
        return x

# 训练好的模型参数    
model_params = torch.load('model.pth') 
model = LeNet5()
model.load_state_dict(model_params)

# 测试数据集
test_set = datasets.MNIST('mnist_data', train=False, download=True, 
                           transform=transforms.Compose([
                               transforms.ToTensor(), 
                               transforms.Normalize((0.1307,), (0.3081,))
                           ]))
test_loader = DataLoader(test_set, batch_size=128, shuffle=True)

# 模型预测结果
preds = []
with torch.no_grad():
    for batch in test_loader:
        images, labels = batch 
        preds.append(model(images).max(1)[1].numpy())
preds = np.concatenate(preds)

# API
@app.route('/predict', methods=['POST'])
def predict():
    img = request.files['img']
    img = torch.from_numpy(preprocessing(img)).unsqueeze(0)
    with torch.no_grad():
        pred = model(img).max(1)[1].item()
    return jsonify(pred)

@app.route('/accuracy', methods=['GET'])
def accuracy():
    acc = (preds == test_set.targets.numpy()).mean()
    return jsonify(acc)

if __name__ == '__main__':
    app.run(debug=True)