YOLOv8实战垃圾分类目标检测 (视频课程）

**YOLOv8实战垃圾分类目标检测**

**视频课程内容概述**

本课程将带领您一步步地完成一个YOLOv8的垃圾分类目标检测项目。我们将从安装所需的库开始，接着是数据集准备、模型训练和测试等环节。

### **一、环境准备**

首先，我们需要安装所需的库：

bashpip install torch torchvision opencv-python numpy pandas

bashgit clone  yolov8python -m pip install .

bashimport os# 数据集路径data_path = 'path/to/trash_data'

# 训练图片路径train_img_path = os.path.join(data_path, 'train')
# 测试图片路径test_img_path = os.path.join(data_path, 'test')

print(f"训练图片数量：{len(os.listdir(train_img_path))}")
print(f"测试图片数量：{len(os.listdir(test_img_path))}")

import yaml# 模型配置文件路径config_path = 'yolov8/configs/yolov8.yaml'

with open(config_path, 'r') as f:
 config = yaml.safe_load(f)

print(config)

import torch# 模型路径model_path = 'yolov8/models/yolov8.pt'

# 数据集路径data_path = 'path/to/trash_data'

# 训练图片路径train_img_path = os.path.join(data_path, 'train')

# 测试图片路径test_img_path = os.path.join(data_path, 'test')

# 模型配置文件路径config_path = 'yolov8/configs/yolov8.yaml'

with open(config_path, 'r') as f:
 config = yaml.safe_load(f)

# 加载预训练权重model = torch.load(model_path, map_location=torch.device('cpu'))

# 微调模型model.train()
for param in model.parameters():
 param.requires_grad = True# 定义数据加载器train_loader = torch.utils.data.DataLoader(
 dataset=torchvision.datasets.ImageFolder(train_img_path),
 batch_size=32,
 shuffle=True,
)

test_loader = torch.utils.data.DataLoader(
 dataset=torchvision.datasets.ImageFolder(test_img_path),
 batch_size=32,
 shuffle=False,
)

# 定义优化器optimizer = torch.optim.Adam(model.parameters(), lr=0.001)

# 训练模型for epoch in range(10):
 for images, labels in train_loader:
 optimizer.zero_grad()
 outputs = model(images)
 loss = F.cross_entropy(outputs, labels)
 loss.backward()
 optimizer.step()

 # 测试模型 model.eval()
 correct =0 with torch.no_grad():
 for images, labels in test_loader:
 outputs = model(images)
 _, predicted = torch.max(outputs,1)
 correct += (predicted == labels).sum().item()

 accuracy = correct / len(test_loader.dataset)
 print(f"Epoch {epoch+1}, Loss: {loss.item():.4f}, Accuracy: {accuracy:.4f}")

import torch# 模型路径model_path = 'yolov8/models/yolov8.pt'

# 数据集路径data_path = 'path/to/trash_data'

# 测试图片路径test_img_path = os.path.join(data_path, 'test')

# 加载预训练权重model = torch.load(model_path, map_location=torch.device('cpu'))

# 定义数据加载器test_loader = torch.utils.data.DataLoader(
 dataset=torchvision.datasets.ImageFolder(test_img_path),
 batch_size=32,
 shuffle=False,
)

# 测试模型model.eval()
correct =0with torch.no_grad():
 for images, labels in test_loader:
 outputs = model(images)
 _, predicted = torch.max(outputs,1)
 correct += (predicted == labels).sum().item()

accuracy = correct / len(test_loader.dataset)
print(f"Accuracy: {accuracy:.4f}")