Cloud Deployment

1# Deploy via CLI
2langtrain deploy create \
3  --model my-fine-tuned-model \
4  --name production-api \
5  --region us-east-1 \
6  --min-instances 1 \
7  --max-instances 10
8
9# Deploy via Python SDK
10deployment = client.deployments.create(
11    model_id="your-model-id",
12    name="production-api",
13    config={
14        "region": "us-east-1",
15        "instance_type": "gpu-medium",
16        "min_instances": 1,
17        "max_instances": 10,
18        "auto_scaling": True
19    }
20)

1# Generate Dockerfile
2langtrain deploy generate-dockerfile --model my-model
3
4# Build and deploy
5docker build -t my-model:latest .
6docker push your-registry/my-model:latest
7
8# Kubernetes deployment
9apiVersion: apps/v1
10kind: Deployment
11metadata:
12  name: langtrain-model
13spec:
14  replicas: 3
15  selector:
16    matchLabels:
17      app: langtrain-model
18  template:
19    metadata:
20      labels:
21        app: langtrain-model
22    spec:
23      containers:
24      - name: model
25        image: your-registry/my-model:latest
26        ports:
27        - containerPort: 8000

1# Configure load balancer
2deployment_config = {
3    "load_balancer": {
4        "algorithm": "round_robin",
5        "health_check": {
6            "path": "/health",
7            "interval": 30,
8            "timeout": 5,
9            "healthy_threshold": 2,
10            "unhealthy_threshold": 3
11        },
12        "sticky_sessions": False
13    },
14    "auto_scaling": {
15        "metric": "requests_per_second",
16        "target": 100,
17        "scale_up_cooldown": 300,
18        "scale_down_cooldown": 600
19    }
20}

1# Set up monitoring
2client.monitoring.create_alert(
3    deployment_id="your-deployment-id",
4    metric="response_time_p95",
5    threshold=2000,  # 2 seconds
6    comparison="greater_than",
7    notification_channels=["email", "slack"]
8)
9
10# Custom metrics
11client.monitoring.track_metric(
12    deployment_id="your-deployment-id",
13    metric_name="business_metric",
14    value=42,
15    tags={"version": "v1.2", "region": "us-east-1"}
16)

1# GitHub Actions workflow
2name: Deploy Model
3on:
4  push:
5    branches: [main]
6
7jobs:
8  deploy:
9    runs-on: ubuntu-latest
10    steps:
11    - uses: actions/checkout@v3
12    - name: Deploy to LangTrain
13      uses: langtrain/deploy-action@v1
14      with:
15        api-key: ${{ secrets.LANGTRAIN_API_KEY }}
16        model-id: ${{ vars.MODEL_ID }}
17        deployment-name: production-api