optuna优化MLP神经网络pytorch实现

# -*- coding: utf-8 -*-
import torch
import optuna
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from torch.utils.data import DataLoader, TensorDataset
from torch import nn, optim
import pandas as pd

# -*- coding: utf-8 -*-
import torch
import optuna
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from torch.utils.data import DataLoader, TensorDataset
from torch import nn, optim
import pandas as pd

# 设置随机种子
torch.manual_seed(42)

#准备数据
def prepare_data():
   

    return train_loader, val_loader,X_train, X_val


# 定义模型类
class MLPModel(nn.Module):
    

# 定义训练函数
def train(model, optimizer, criterion, data_loader):
    model.train()
    total_loss = 0.0
    total_samples = 0

    for inputs, targets in data_loader:
        optimizer.zero_grad()
        outputs = model(inputs)
        loss = criterion(outputs, targets)
        loss.backward()
        optimizer.step()

        total_loss += loss.item() * inputs.size(0)
        total_samples += inputs.size(0)

    return total_loss / total_samples

# 定义评估函数
def evaluate(model, data_loader):
    model.eval()
    correct = 0
    total_samples = 0

    with torch.no_grad():
        for inputs, targets in data_loader:
            outputs = model(inputs)
            _, predicted_labels = torch.max(outputs.data, 1)
            correct += (predicted_labels == targets).sum().item()
            total_samples += targets.size(0)

    accuracy = correct / total_samples
    return accuracy

# 定义目标函数
def objective(trial):
    # 定义超参数范围
    hidden_dim = trial.suggest_int("hidden_dim", 2, 100)
    learning_rate = trial.suggest_float("learning_rate", 1e-5, 1e-1)

    # 创建MLP模型
    input_dim = X_train.shape[1]
    output_dim = 3
    model = MLPModel(input_dim, hidden_dim, output_dim)

   
    return val_accuracy

if __name__ == '__main__':

    # 训练模型num_epochs
    num_epochs = 1000
    # 调用函数来准备数据
    train_loader, val_loader,X_train, X_val = prepare_data()

    # 设置试验次数
    num_trials = 10

    # 使用Optuna搜索最佳超参数配置
    study = optuna.create_study(direction="maximize",storage='sqlite:///example.db')

    # 加载指定相同的存储器地址
    #study = optuna.load_study(study_name='my_study', storage='sqlite:///example.db')

    #开始优化
    study.optimize(objective, n_trials=num_trials)

    # 打印最佳超参数配置和对应的验证集准确率
    best_params = study.best_params
    best_val_accuracy = study.best_value
    print("Best Parameters:", best_params)
    print("Best Validation Accuracy:", best_val_accuracy)

以上为部分代码，更多请关注同名公众号