# Example: GPU-accelerated deep learning training
import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader, TensorDataset
from torchvision import datasets, transforms
import time

# Initialize GPU device
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
print(f"Using device: {device}")

# Data preprocessing and loading
transform = transforms.Compose([
    transforms.ToTensor(),
    transforms.Normalize((0.5,), (0.5,))
])

train_dataset = datasets.MNIST(
    root='./data', train=True, download=True, transform=transform
)
train_loader = DataLoader(train_dataset, batch_size=64, shuffle=True)

# Define CNN model
class CNN(nn.Module):
    def __init__(self):
        super(CNN, self).__init__()
        self.conv1 = nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1)
        self.conv2 = nn.Conv2d(32, 64, kernel_size=3, stride=1, padding=1)
        self.pool = nn.MaxPool2d(kernel_size=2, stride=2)
        self.fc1 = nn.Linear(64 * 7 * 7, 128)
        self.fc2 = nn.Linear(128, 10)
        self.relu = nn.ReLU()

    def forward(self, x):
        x = self.pool(self.relu(self.conv1(x)))
        x = self.pool(self.relu(self.conv2(x)))
        x = x.view(-1, 64 * 7 * 7)
        x = self.relu(self.fc1(x))
        x = self.fc2(x)
        return x

# Initialize model, loss function, and optimizer
model = CNN().to(device)
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.001)

# Training loop with GPU utilization
start_time = time.time()
epochs = 5

for epoch in range(epochs):
    running_loss = 0.0
    for i, data in enumerate(train_loader, 0):
        inputs, labels = data[0].to(device), data[1].to(device)

        optimizer.zero_grad()
        outputs = model(inputs)
        loss = criterion(outputs, labels)
        loss.backward()
        optimizer.step()

        running_loss += loss.item()
        if i % 100 == 99:
            print(f'[{epoch + 1}, {i + 1}] loss: {running_loss / 100:.3f}')
            running_loss = 0.0

print(f'Training completed in {time.time() - start_time:.2f} seconds')
# Save model to cloud storage
torch.save(model.state_dict(), 'model_gpu.pth')
print('Model saved to cloud storage')

# Cloud GPU resource management with CLI
# =======================================
# Scale GPU cluster based on workload
# =======================================
# Scale up to 4 nodes during peak hours
gpucloud scale --cluster ai-training --nodes 4

# Monitor GPU and memory utilization in real-time
gpucloud monitor --cluster ai-training --metrics gpu,memory --interval 5s

# Set up auto-scaling based on GPU utilization threshold
gpucloud autoscale --cluster ai-training   --min-nodes 2   --max-nodes 8   --threshold 70%   --cooldown-period 5m

# =======================================
# Deploy and manage GPU workloads
# =======================================
# Deploy container with specific GPU requirements
gpucloud deploy   --name ml-training-job   --image pytorch:latest   --gpus 2   --memory 32GB   --cpu 8   --command "python train.py --epochs 10"

# List all active GPU instances
gpucloud list --type gpu --status active

# Get detailed information about a specific instance
gpucloud inspect --instance gpu-instance-123

# =======================================
# Optimize costs and performance
# =======================================
# List available GPU instance types and pricing
gpucloud instance-types --filter gpu

# Reserve GPU instances for predictable workloads
gpucloud reserve --instance-type a100-80gb --count 2 --duration 1mo

# Schedule a job during off-peak hours for lower rates
gpucloud schedule   --job-id training-job-001   --start-time 2023-06-15T22:00:00Z   --end-time 2023-06-16T06:00:00Z   --priority low

# =======================================
# Manage GPU clusters
# =======================================
# Create a new GPU cluster with custom configuration
gpucloud create-cluster   --name research-gpus   --instance-type v100-32gb   --initial-nodes 2   --region us-west-2   --network private-vpc

# Update cluster configuration
gpucloud update-cluster   --name research-gpus   --instance-type a100-40gb   --max-nodes 10