optuna和 lightgbm

tony3652024-12-30 0:37

文章目录

optuna使用

1.导入相关包

python 复制代码 
import torch
import torch.nn as nn
import torch.nn.functional as F
import torchvision
from fvcore.nn import FlopCountAnalysis

import optuna


DEVICE = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
DIR = ".."
BATCHSIZE = 128
N_TRAIN_EXAMPLES = BATCHSIZE * 30   # 128 * 30个训练
N_VALID_EXAMPLES = BATCHSIZE * 10   # 128 * 10个预测

2.定义模型可选参数

optuna支持很多种搜索方式：

（1）trial.suggest_categorical('optimizer', ['MomentumSGD', 'Adam'])：表示从SGD和adam里选一个使用；

（2）trial.suggest_int('num_layers', 1, 3)：从1～3范围内的int里选；

（3）trial.suggest_uniform('dropout_rate', 0.0, 1.0)：从0～1内的uniform分布里选；

（4）trial.suggest_loguniform('learning_rate', 1e-5, 1e-2)：从1e-5～1e-2的log uniform分布里选；

（5）trial.suggest_discrete_uniform('drop_path_rate', 0.0, 1.0, 0.1)：从0～1且step为0.1的离散uniform分布里选；

python 复制代码 
def define_model(trial):
    n_layers = trial.suggest_int("n_layers", 1, 3) # 从[1，3]范围里面选一个
    layers = []

    in_features = 28 * 28
    for i in range(n_layers):
        out_features = trial.suggest_int("n_units_l{}".format(i), 4, 128)
        layers.append(nn.Linear(in_features, out_features))
        layers.append(nn.ReLU())
        p = trial.suggest_float("dropout_{}".format(i), 0.2, 0.5)
        layers.append(nn.Dropout(p))

        in_features = out_features

    layers.append(nn.Linear(in_features, 10))
    layers.append(nn.LogSoftmax(dim=1))

    return nn.Sequential(*layers)

3.定义训练代码和评估代码

python 复制代码 
# Defines training and evaluation.
def train_model(model, optimizer, train_loader):
    model.train()
    for batch_idx, (data, target) in enumerate(train_loader):
        data, target = data.view(-1, 28 * 28).to(DEVICE), target.to(DEVICE)
        optimizer.zero_grad()
        F.nll_loss(model(data), target).backward()
        optimizer.step()


def eval_model(model, valid_loader):
    model.eval()
    correct = 0
    with torch.no_grad():
        for batch_idx, (data, target) in enumerate(valid_loader):
            data, target = data.view(-1, 28 * 28).to(DEVICE), target.to(DEVICE)
            pred = model(data).argmax(dim=1, keepdim=True)
            correct += pred.eq(target.view_as(pred)).sum().item()

    accuracy = correct / N_VALID_EXAMPLES

    flops = FlopCountAnalysis(model, inputs=(torch.randn(1, 28 * 28).to(DEVICE),)).total()
    return flops, accuracy

4.定义目标函数

python 复制代码 
def objective(trial):
    train_dataset = torchvision.datasets.FashionMNIST(
        DIR, train=True, download=True, transform=torchvision.transforms.ToTensor()
    )
    train_loader = torch.utils.data.DataLoader(
        torch.utils.data.Subset(train_dataset, list(range(N_TRAIN_EXAMPLES))),
        batch_size=BATCHSIZE,
        shuffle=True,
    )

    val_dataset = torchvision.datasets.FashionMNIST(
        DIR, train=False, transform=torchvision.transforms.ToTensor()
    )
    val_loader = torch.utils.data.DataLoader(
        torch.utils.data.Subset(val_dataset, list(range(N_VALID_EXAMPLES))),
        batch_size=BATCHSIZE,
        shuffle=True,
    )
    model = define_model(trial).to(DEVICE)

    optimizer = torch.optim.Adam(
        model.parameters(), trial.suggest_float("lr", 1e-5, 1e-1, log=True)
    )

    for epoch in range(10):
        train_model(model, optimizer, train_loader)
    flops, accuracy = eval_model(model, val_loader)
    return flops, accuracy

5.运行程序

运行30次实验，每次实验返回 flops,accuracy

python 复制代码 
study = optuna.create_study(directions=["minimize", "maximize"]) # flops 最小化， accuracy 最大化
study.optimize(objective, n_trials=30, timeout=300)

print("Number of finished trials: ", len(study.trials))

6.可视化

flops, accuracy 二维图

optuna.visualization.plot_pareto_front(study, target_names=["FLOPS", "accuracy"])

7.超参数的重要性

对于flops

optuna.visualization.plot_param_importances(

study, target=lambda t: t.values[0], target_name="flops"

)

对于accuracy

optuna.visualization.plot_param_importances(

study, target=lambda t: t.values[1], target_name="accuracy"

)

8.查看相关信息

python 复制代码 
# https://optuna.readthedocs.io/en/stable/tutorial/20_recipes/002_multi_objective.html
# 利用pytorch mnist 识别
# 设置了一些超参数，lr, layer number, feature_number等
# 然后目标是 flops 和 accurary

# 最后是可视化：
# 显示试验的一些结果：
# optuna.visualization.plot_pareto_front(study, target_names=["FLOPS", "accuracy"])
# 左上角是最好的

# 显示重要性：
# optuna.visualization.plot_param_importances(
#     study, target=lambda t: t.values[0], target_name="flops"
# )
# optuna.visualization.plot_param_importances(
#     study, target=lambda t: t.values[1], target_name="accuracy"
# )


# trials的属性：
print(f"Number of trials on the Pareto front: {len(study.best_trials)}")

trial_with_highest_accuracy = max(study.best_trials, key=lambda t: t.values[1])
print(f"Trial with highest accuracy: ")
print(f"\tnumber: {trial_with_highest_accuracy.number}")
print(f"\tparams: {trial_with_highest_accuracy.params}")
print(f"\tvalues: {trial_with_highest_accuracy.values}")

9.可视化的一个完整示例

python 复制代码 
# You can use Matplotlib instead of Plotly for visualization by simply replacing `optuna.visualization` with
# `optuna.visualization.matplotlib` in the following examples.
from optuna.visualization import plot_contour
from optuna.visualization import plot_edf
from optuna.visualization import plot_intermediate_values
from optuna.visualization import plot_optimization_history
from optuna.visualization import plot_parallel_coordinate
from optuna.visualization import plot_param_importances
from optuna.visualization import plot_rank
from optuna.visualization import plot_slice
from optuna.visualization import plot_timeline

def objective(trial):
    train_dataset = torchvision.datasets.FashionMNIST(
        DIR, train=True, download=True, transform=torchvision.transforms.ToTensor()
    )
    train_loader = torch.utils.data.DataLoader(
        torch.utils.data.Subset(train_dataset, list(range(N_TRAIN_EXAMPLES))),
        batch_size=BATCHSIZE,
        shuffle=True,
    )

    val_dataset = torchvision.datasets.FashionMNIST(
        DIR, train=False, transform=torchvision.transforms.ToTensor()
    )
    val_loader = torch.utils.data.DataLoader(
        torch.utils.data.Subset(val_dataset, list(range(N_VALID_EXAMPLES))),
        batch_size=BATCHSIZE,
        shuffle=True,
    )
    model = define_model(trial).to(DEVICE)

    optimizer = torch.optim.Adam(
        model.parameters(), trial.suggest_float("lr", 1e-5, 1e-1, log=True)
    )

    for epoch in range(10):
        train_model(model, optimizer, train_loader)

        val_accuracy = eval_model(model, val_loader)
        trial.report(val_accuracy, epoch)

        if trial.should_prune():
            raise optuna.exceptions.TrialPruned()

    return val_accuracy

study = optuna.create_study(
    direction="maximize",
    sampler=optuna.samplers.TPESampler(seed=SEED),
    pruner=optuna.pruners.MedianPruner(),
)
study.optimize(objective, n_trials=30, timeout=300)

运行之后可视化：


10.lightgbm实验

python 复制代码 
"""
Optuna example that optimizes a classifier configuration for cancer dataset using LightGBM.

In this example, we optimize the validation accuracy of cancer detection using LightGBM.
We optimize both the choice of booster model and their hyperparameters.

"""

import numpy as np
import optuna

import lightgbm as lgb
import sklearn.datasets
import sklearn.metrics
from sklearn.model_selection import train_test_split


# FYI: Objective functions can take additional arguments
# (https://optuna.readthedocs.io/en/stable/faq.html#objective-func-additional-args).
def objective(trial):
    data, target = sklearn.datasets.load_breast_cancer(return_X_y=True)
    train_x, valid_x, train_y, valid_y = train_test_split(data, target, test_size=0.25)
    dtrain = lgb.Dataset(train_x, label=train_y)

    param = {
        "objective": "binary",
        "metric": "binary_logloss",
        "verbosity": -1,
        "boosting_type": "gbdt",
        "lambda_l1": trial.suggest_float("lambda_l1", 1e-8, 10.0, log=True),
        "lambda_l2": trial.suggest_float("lambda_l2", 1e-8, 10.0, log=True),
        "num_leaves": trial.suggest_int("num_leaves", 2, 256),
        "feature_fraction": trial.suggest_float("feature_fraction", 0.4, 1.0),
        "bagging_fraction": trial.suggest_float("bagging_fraction", 0.4, 1.0),
        "bagging_freq": trial.suggest_int("bagging_freq", 1, 7),
        "min_child_samples": trial.suggest_int("min_child_samples", 5, 100),
    }

    gbm = lgb.train(param, dtrain)
    preds = gbm.predict(valid_x)
    pred_labels = np.rint(preds)
    accuracy = sklearn.metrics.accuracy_score(valid_y, pred_labels)
    return accuracy


if __name__ == "__main__":
    study = optuna.create_study(direction="maximize")
    study.optimize(objective, n_trials=100)

    print("Number of finished trials: {}".format(len(study.trials)))

    print("Best trial:")
    trial = study.best_trial

    print("  Value: {}".format(trial.value))

    print("  Params: ")
    for key, value in trial.params.items():
        print("    {}: {}".format(key, value))

运行结果：

https://github.com/microsoft/LightGBM/tree/master/examples

https://blog.csdn.net/yang1015661763/article/details/131364826

相关推荐
胖达不服输1 分钟前
「日拱一码」021 机器学习——特征工程
人工智能·python·机器学习·特征工程
小哥谈1 小时前
论文解析篇 | YOLOv12：以注意力机制为核心的实时目标检测算法
人工智能·深度学习·yolo·目标检测·机器学习·计算机视觉
screenCui1 小时前
macOS运行python程序遇libiomp5.dylib库冲突错误解决方案
开发语言·python·macos
水龙吟啸1 小时前
从零开始搭建深度学习大厦系列-2.卷积神经网络基础（5-9）
人工智能·pytorch·深度学习·cnn·mxnet
小眼睛羊羊1 小时前
pyinstaller打包paddleocr
python
java1234_小锋1 小时前
基于Python的旅游推荐协同过滤算法系统(去哪儿网数据分析及可视化(Django+echarts))
python·数据分析·旅游
蓝婷儿1 小时前
Python 机器学习核心入门与实战进阶 Day 4 - 支持向量机（SVM）原理与分类实战
python·机器学习·支持向量机
%d%d22 小时前
python 在运行时没有加载修改后的版本
java·服务器·python
HollowKnightZ2 小时前
论文阅读笔记：VI-Net: Boosting Category-level 6D Object Pose Estimation
人工智能·深度学习·计算机视觉
yzx9910132 小时前
AI大模型平台
大数据·人工智能·深度学习·机器学习
热门推荐
01使用Ruby接入实时行情API教程02KGG转MP3工具|非KGM文件|解密音频03Coze扣子平台完整体验和实践（附国内和国际版对比）04深度神经网络训练过程与常见概念05集群聊天服务器---MySQL数据库的建立06扣子（coze）实战|我用扣子搭建了一个自动分析小红薯笔记内容的AI应用|详细步骤拆解07【无标题】08Java学习第十五部分——MyBatis09DeepSeek各版本说明与优缺点分析10LOT: 通过逻辑增强大型语言模型的零样本Chain-of-Thought推理能力