@aws/ml-container-creator 0.2.2 → 0.2.4

package/README.md

@@ -1,106 +1,342 @@
-# ML Container Creator
+# sharp-transformer-deployment
 
-A CLI tool that creates SageMaker-compatible Docker containers for deploying ML models using the Bring Your Own Container (BYOC) paradigm.
+SageMaker-compatible ML container for deploying transformers models using vllm.
 
-> **Note:** This is a pre-release (`0.x`). APIs may change between minor versions. Weekly releases are planned until v1.
+Generated on 2026-05-08T09-52-06 using [ML Container Creator](https://github.com/yourusername/ml-container-creator).
 
-## Supported Configurations
+## Quick Start
 
-| Architecture | Model Servers | Use Case |
-|---|---|---|
-| HTTP (traditional ML) | Flask, FastAPI | sklearn, XGBoost, TensorFlow |
-| Transformers (LLMs) | vLLM, SGLang, TensorRT-LLM, DJL/LMI | HuggingFace models, JumpStart, S3 |
-| Triton | FIL, ONNX, Python, TensorRT-LLM, vLLM | Multi-framework serving |
-| Diffusors | vLLM | Image generation models |
+### 1. Build the Container
 
-| Deployment Target | Description |
-|---|---|
-| Managed Inference | SageMaker real-time endpoints |
-| Async Inference | SageMaker async endpoints with S3 output |
-| Batch Transform | SageMaker batch processing |
-| HyperPod EKS | Kubernetes-based deployment |
+```bash
+./do/build
+```
 
-## Quick Start
+Builds a Docker image tagged as `sharp-transformer-deployment:latest`.
+
+### 2. Test Locally
+
+```bash
+# Start the container
+./do/run
+
+# In another terminal, test the endpoints
+./do/test
+```
+
+### 3. Push to ECR
+
+```bash
+./do/push
+```
+
+Pushes the image to Amazon ECR in the `us-west-2` region.
+
+### 4. Deploy to SageMaker
+
+```bash
+./do/deploy <your-sagemaker-execution-role-arn>
+```
+
+Creates a SageMaker endpoint named `sharp-transformer-deployment-endpoint`.
+
+### 5. Test the Endpoint
+
+```bash
+./do/test sharp-transformer-deployment-endpoint
+```
+
+## Project Structure
+
+```
+sharp-transformer-deployment/
+├── do/                      # do-framework lifecycle scripts
+│   ├── build                # Build Docker image
+│   ├── push                 # Push to Amazon ECR
+│   ├── deploy               # Deploy to SageMaker
+│   ├── run                  # Run container locally
+│   ├── test                 # Test container or endpoint
+│   ├── clean                # Clean up resources
+│   ├── submit               # Submit build to CodeBuild
+│   ├── config               # Configuration variables
+│   └── README.md            # Detailed do-framework documentation
+├── code/                    # Model serving code
+│   └── serve               # vllm entrypoint script
+├── deploy/                 # Legacy scripts (deprecated)
+│   ├── build_and_push.sh   # Use ./do/build && ./do/push instead
+│   └── deploy.sh           # Use ./do/deploy instead
+
+├── test/                  # Test suite
+│   ├── test_endpoint.sh    # Test SageMaker endpoint
+│   └── test_local_image.sh # Test local container
+
+├── Dockerfile              # Container definition
+├── requirements.txt        # Python dependencies
+└── README.md               # This file
+```
+
+## Configuration
+
+All deployment configuration is centralized in `do/config`:
+
+```bash
+# Project identification
+PROJECT_NAME="sharp-transformer-deployment"
+DEPLOYMENT_CONFIG="transformers-vllm"
+
+# AWS configuration
+AWS_REGION="us-west-2"
+INSTANCE_TYPE="ml.g5.xlarge"
+
+# Framework configuration
+FRAMEWORK="transformers"
+MODEL_SERVER="vllm"
+
+# Model configuration
+MODEL_NAME="openai/gpt-oss-20b"
+
+```
+
+You can override these values by setting environment variables before running do scripts.
+
+## Deployment Workflows
+
+### Local Development Workflow
+
+```bash
+# Build and test locally
+./do/build
+./do/run &
+./do/test
+
+# When satisfied, push to ECR
+./do/push
+```
+
+### CodeBuild Workflow
+
+```bash
+# Submit build to CodeBuild (builds and pushes to ECR)
+./do/submit
+
+# Deploy to SageMaker
+./do/deploy <role-arn>
+
+# Test the endpoint
+./do/test sharp-transformer-deployment-endpoint
+```
+
+### Cleanup
+
+```bash
+# Remove local images
+./do/clean local
+
+# Remove ECR images
+./do/clean ecr
+
+# Delete SageMaker endpoint
+./do/clean endpoint
+
+# Clean everything
+./do/clean all
+```
+
+## do-framework Commands
+
+This project uses the [do-framework](https://github.com/iankoulski/do-framework) for standardized container lifecycle management.
+
+### Available Commands
+
+| Command | Description |
+|---------|-------------|
+| `./do/build` | Build Docker image locally |
+| `./do/push` | Push image to Amazon ECR |
+| `./do/deploy <role-arn>` | Deploy to SageMaker endpoint |
+| `./do/run` | Run container locally on port 8080 |
+| `./do/test [endpoint]` | Test local container or SageMaker endpoint |
+| `./do/clean <target>` | Clean up resources (local/ecr/endpoint/all) |
+| `./do/submit` | Submit build to AWS CodeBuild |
+
+For detailed documentation on each command, see `do/README.md`.
+
+## Framework-Specific Information
+
+### Transformers (vllm)
+
+This container serves transformer models using vllm.
+
+**Model**: openai/gpt-oss-20b
 
-### Install from npm
+**Server**: vLLM - High-throughput LLM serving with PagedAttention
+
+**Features**:
+- Continuous batching
+- Optimized CUDA kernels
+- OpenAI-compatible API
+
+
+**Inference**: Send requests to `/invocations` endpoint with:
+```json
+{
+  "inputs": "Your prompt here",
+  "parameters": {
+    "max_new_tokens": 100,
+    "temperature": 0.7
+  }
+}
+```
+
+
+
+## SageMaker Endpoints
+
+### Health Check
+
+SageMaker calls the `/ping` endpoint to verify container health:
 
 ```bash
-npm install -g @aws/ml-container-creator
+curl http://localhost:8080/ping
 ```
 
-### Or use without installing (npx)
+Expected response: `200 OK`
+
+### Inference
+
+Send prediction requests to the `/invocations` endpoint:
 
 ```bash
-npx @aws/ml-container-creator --help
+curl -X POST http://localhost:8080/invocations \
+  -H "Content-Type: application/json" \
+  -d '{
+    "inputs": "What is machine learning?",
+    "parameters": {
+      "max_new_tokens": 100,
+      "temperature": 0.7
+    }
+  }'
 ```
 
-### Or install from source
 
+## AWS Requirements
+
+### IAM Permissions
+
+The SageMaker execution role needs these permissions:
+
+- `ecr:GetAuthorizationToken`
+- `ecr:BatchCheckLayerAvailability`
+- `ecr:GetDownloadUrlForLayer`
+- `ecr:BatchGetImage`
+- `s3:GetObject` (if using S3 for model artifacts)
+- `logs:CreateLogGroup`
+- `logs:CreateLogStream`
+- `logs:PutLogEvents`
+
+See `IAM_PERMISSIONS.md` for detailed permission requirements.
+
+### AWS CLI Configuration
+
+Ensure AWS CLI is configured with appropriate credentials:
+
+```bash
+aws configure
+```
+
+Or use environment variables:
+
+```bash
+export AWS_ACCESS_KEY_ID=your-access-key
+export AWS_SECRET_ACCESS_KEY=your-secret-key
+export AWS_DEFAULT_REGION=us-west-2
+```
+
+## Troubleshooting
+
+### Build Issues
+
+**Docker Not Found**
+
+Install Docker: https://docs.docker.com/get-docker/
+
+**Permission Denied**
+
+Add your user to the docker group:
 ```bash
-git clone https://github.com/awslabs/ml-container-creator.git
-cd ml-container-creator
-npm install && npm link
+sudo usermod -aG docker $USER
 ```
 
-### Bootstrap AWS infrastructure (one-time)
+### Deployment Issues
 
+**ECR Push Failed**
+
+Check AWS credentials and IAM permissions:
 ```bash
-ml-container-creator bootstrap
+aws sts get-caller-identity
 ```
 
-Sets up an IAM execution role, ECR repository, optional S3 buckets, and optional CI Integration Harness for automated testing. Configuration is saved to `~/.ml-container-creator/config.json`.
+**Endpoint Creation Failed**
+
+- Verify the execution role ARN is correct
+- Check IAM permissions
+- Ensure the instance type is available in your region
 
-### Generate a project
+**Endpoint Stuck in Creating**
 
+Check CloudWatch logs:
 ```bash
-# Interactive
-ml-container-creator
+aws logs tail /aws/sagemaker/Endpoints/sharp-transformer-deployment-endpoint --follow
+```
 
-# Non-interactive
-ml-container-creator my-model \
-  --deployment-config=transformers-vllm \
-  --model-name=openai/gpt-oss-20b \
-  --instance-type=ml.g6.12xlarge \
-  --region=us-east-1 \
-  --skip-prompts
+### Runtime Issues
+
+**Container Exits Immediately**
+
+Check container logs:
+```bash
+docker logs $(docker ps -a | grep sharp-transformer-deployment | awk '{print $1}')
 ```
 
-### Build, push, deploy
+**Out of Memory**
 
+Increase instance size or optimize model:
 ```bash
-./do/build        # Build Docker image
-./do/push         # Push to Amazon ECR
-./do/deploy       # Deploy to SageMaker
-./do/test         # Test the endpoint
+# Edit do/config
+INSTANCE_TYPE="ml.m5.2xlarge"  # Larger instance
 ```
 
-## Documentation
+## Migration from Legacy Scripts
+
+If you're familiar with the old `deploy/` scripts, see `MIGRATION.md` for a command mapping guide.
+
+**Quick Reference**:
+
+| Legacy Command | do-framework Command |
+|----------------|---------------------|
+| `./deploy/build_and_push.sh` | `./do/build && ./do/push` |
+| `./deploy/deploy.sh <role>` | `./do/deploy <role>` |
+| `./deploy/submit_build.sh` | `./do/submit` |
 
-Full documentation is available at [awslabs.github.io/ml-container-creator](https://awslabs.github.io/ml-container-creator/).
+The legacy scripts are still available but deprecated. They will display warnings and forward to do-framework commands.
 
-- [Getting Started](https://awslabs.github.io/ml-container-creator/getting-started/) — Installation and walkthroughs
-- [Configuration](https://awslabs.github.io/ml-container-creator/configuration/) — CLI flags, env vars, config files, MCP servers
-- [Deployment Guide](https://awslabs.github.io/ml-container-creator/deployments/) — All deployment targets and lifecycle scripts
-- [CI Integration](https://awslabs.github.io/ml-container-creator/ci-integration/) — Automated lifecycle testing for all deployment configurations
-- [Examples](https://awslabs.github.io/ml-container-creator/EXAMPLES/) — Framework-specific walkthroughs
-- [Troubleshooting](https://awslabs.github.io/ml-container-creator/TROUBLESHOOTING/) — Common issues and solutions
+## Additional Resources
 
-## Prerequisites
+- [do-framework Documentation](https://github.com/iankoulski/do-framework)
+- [AWS SageMaker Documentation](https://docs.aws.amazon.com/sagemaker/)
+- [SageMaker BYOC Guide](https://docs.aws.amazon.com/sagemaker/latest/dg/your-algorithms.html)
 
-| Tool | Version | Purpose |
-|---|---|---|
-| [Node.js](https://nodejs.org/) | 24+ | Runs the CLI |
-| [Docker](https://docs.docker.com/get-docker/) | 20+ | Container builds |
-| [AWS CLI](https://aws.amazon.com/cli/) | 2+ | AWS resource management |
+- [vLLM Documentation](https://docs.vllm.ai/)
 
-## Contributing
 
-See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
+## Support
 
-## Security
+For issues or questions:
 
-See [CONTRIBUTING.md](CONTRIBUTING.md#security-issue-notifications) for reporting security issues.
+1. Check `do/README.md` for detailed command documentation
+2. Review CloudWatch logs for deployment issues
+3. See `MIGRATION.md` if migrating from legacy scripts
+4. Open an issue on the [ML Container Creator repository](https://github.com/yourusername/ml-container-creator)
 
 ## License
 
-Apache-2.0. See [LICENSE](LICENSE).
+This generated project is provided as starter code. Modify as needed for your use case.

package/bin/cli.js

@@ -27,7 +27,8 @@ program
 
     // --- General ---
     .addOption(new Option('--skip-prompts', 'Skip interactive prompts and use configuration from other sources'))
-    .addOption(new Option('--config <path>', 'Path to configuration file'))
+    .addOption(new Option('--auto-prompt', 'Fill defaults, prompt only for missing required values'))
+    .addOption(new Option('--config <path>', 'Path to JSON configuration file'))
     .addOption(new Option('--project-name <name>', 'Project name'))
     .addOption(new Option('--project-dir <dir>', 'Output directory path'))
     .addOption(new Option('--force', 'Overwrite existing output directory without prompting'))
@@ -41,7 +42,7 @@ program
     .addOption(new Option('--base-image <image>', 'Base container image for Dockerfile'))
 
     // --- Build & Infrastructure ---
-    .addOption(new Option('--deployment-target <target>', 'Deployment target (managed-inference, async-inference, batch-transform, hyperpod-eks)'))
+    .addOption(new Option('--deployment-target <target>', 'Deployment target (realtime-inference, async-inference, batch-transform, hyperpod-eks)'))
     .addOption(new Option('--instance-type <type>', 'SageMaker instance type (e.g. ml.g5.xlarge, ml.m5.large)'))
     .addOption(new Option('--region <region>', 'AWS region'))
     .addOption(new Option('--role-arn <arn>', 'IAM role ARN for SageMaker execution'))
@@ -154,7 +155,7 @@ program.configureHelp({
 
         for (const opt of allOptions) {
             const long = opt.long || '';
-            if (['--skip-prompts', '--config', '--project-name', '--project-dir', '--force', '--version', '--help'].includes(long)) {
+            if (['--skip-prompts', '--auto-prompt', '--config', '--project-name', '--project-dir', '--force', '--version', '--help'].includes(long)) {
                 groups.general.push(opt);
             } else if (['--deployment-config', '--framework', '--model-format', '--model-name', '--model-server', '--base-image'].includes(long)) {
                 groups.model.push(opt);

package/config/parameter-schema.json

@@ -1,7 +1,7 @@
 {
     "schemaVersion": "1.0.0",
     "deploymentTargets": {
-        "managed-inference": {
+        "realtime-inference": {
             "endpoint": {
                 "initialInstanceCount": {
                     "type": "integer",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aws/ml-container-creator",
-  "version": "0.2.2",
+  "version": "0.2.4",
   "description": "Generator for SageMaker AI BYOC paradigm for predictive inference use-cases.",
   "type": "module",
   "main": "src/app.js",

package/src/app.js

@@ -156,6 +156,22 @@ export async function run(projectName, options) {
             console.log('   If your model package lacks an InferenceSpecification, use the S3 path');
             console.log('   directly instead: --model-name="s3://bucket/path/model.tar.gz"\n');
         }
+    } else if (configManager.isAutoPrompt()) {
+        // Auto-prompt mode: run the wizard with all resolved values pre-filled.
+        // The wizard skips prompts for values already in explicitConfig and
+        // uses phase-level gates to skip irrelevant sections entirely.
+        // This gives context-aware prompting (correct MCP queries, filtered choices)
+        // while only asking for what's truly missing.
+        console.log('\n🔄 Auto-prompt mode — prompting only for missing values with full context');
+
+        const promptRunner = new PromptRunner({
+            configManager,
+            options: kebabOptions,
+            registryConfigManager,
+            baseConfig
+        });
+        const promptAnswers = await promptRunner.run();
+        answers = configManager.getFinalConfiguration(promptAnswers);
     } else {
         const promptRunner = new PromptRunner({
             configManager,
@@ -482,7 +498,7 @@ async function _ensureTemplateVariables(answers, registryConfigManager = null) {
         testTypes: [],
         buildTimestamp: new Date().toISOString(),
         buildTarget: 'codebuild',
-        deploymentTarget: 'managed-inference',
+        deploymentTarget: 'realtime-inference',
         hyperPodCluster: null,
         hyperPodNamespace: 'default',
         hyperPodReplicas: 1,

package/src/lib/auto-prompt-builder.js

@@ -0,0 +1,172 @@
+// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+// SPDX-License-Identifier: Apache-2.0
+
+/**
+ * Auto-Prompt Builder — generates targeted prompts for missing required parameters.
+ * 
+ * Used by --auto-prompt mode to ask only for values that cannot be inferred
+ * or defaulted from the provided CLI flags.
+ */
+
+/**
+ * Builds a minimal set of prompts for the given missing parameters.
+ * Each prompt is self-contained and doesn't depend on multi-phase wizard state.
+ *
+ * @param {string[]} missingParams - Parameter names that need values
+ * @param {object} currentConfig - Current configuration (with defaults filled)
+ * @returns {Array} Array of prompt objects compatible with runPrompts()
+ */
+export function buildAutoPrompts(missingParams, currentConfig) {
+    const prompts = [];
+
+    for (const param of missingParams) {
+        const builder = PROMPT_BUILDERS[param];
+        if (builder) {
+            const prompt = builder(currentConfig);
+            if (prompt) {
+                prompts.push(prompt);
+            }
+        } else {
+            // Fallback: generic text input for unknown parameters
+            prompts.push({
+                type: 'input',
+                name: param,
+                message: `Enter value for ${param}:`
+            });
+        }
+    }
+
+    return prompts;
+}
+
+/**
+ * Map of parameter names to prompt builder functions.
+ * Each builder receives the current config and returns a prompt object.
+ */
+const PROMPT_BUILDERS = {
+    deploymentConfig: (_config) => ({
+        type: 'list',
+        name: 'deploymentConfig',
+        message: 'Select deployment configuration:',
+        choices: [
+            { type: 'separator', separator: '── Large Language Models ──' },
+            { name: 'Transformers with vLLM', value: 'transformers-vllm' },
+            { name: 'Transformers with SGLang', value: 'transformers-sglang' },
+            { name: 'Transformers with TensorRT-LLM', value: 'transformers-tensorrt-llm' },
+            { name: 'Transformers with LMI', value: 'transformers-lmi' },
+            { name: 'Transformers with DJL', value: 'transformers-djl' },
+            { type: 'separator', separator: '── HTTP Serving ──' },
+            { name: 'HTTP with Flask', value: 'http-flask' },
+            { name: 'HTTP with FastAPI', value: 'http-fastapi' },
+            { type: 'separator', separator: '── NVIDIA Triton ──' },
+            { name: 'Triton FIL (XGBoost, LightGBM)', value: 'triton-fil' },
+            { name: 'Triton ONNX Runtime', value: 'triton-onnxruntime' },
+            { name: 'Triton TensorFlow', value: 'triton-tensorflow' },
+            { name: 'Triton PyTorch', value: 'triton-pytorch' },
+            { name: 'Triton vLLM', value: 'triton-vllm' },
+            { name: 'Triton TensorRT-LLM', value: 'triton-tensorrtllm' },
+            { name: 'Triton Python Backend', value: 'triton-python' },
+            { type: 'separator', separator: '── Diffusion Models ──' },
+            { name: 'Diffusors with vLLM Omni', value: 'diffusors-vllm-omni' }
+        ]
+    }),
+
+    instanceType: (config) => {
+        const architecture = config.architecture || 'http';
+        const isGpu = architecture === 'transformers' || architecture === 'triton' || architecture === 'diffusors';
+
+        const gpuChoices = [
+            { name: 'ml.g5.xlarge  (1× A10G 24GB — small LLMs)', value: 'ml.g5.xlarge' },
+            { name: 'ml.g5.2xlarge (1× A10G 24GB — medium LLMs)', value: 'ml.g5.2xlarge' },
+            { name: 'ml.g5.4xlarge (1× A10G 24GB — larger models)', value: 'ml.g5.4xlarge' },
+            { name: 'ml.g5.12xlarge (4× A10G 96GB — large LLMs)', value: 'ml.g5.12xlarge' },
+            { name: 'ml.g5.48xlarge (8× A10G 192GB — very large)', value: 'ml.g5.48xlarge' },
+            { name: 'ml.g6.xlarge  (1× L4 24GB)', value: 'ml.g6.xlarge' },
+            { name: 'ml.g6.2xlarge (1× L4 24GB)', value: 'ml.g6.2xlarge' },
+            { name: 'ml.p4d.24xlarge (8× A100 320GB)', value: 'ml.p4d.24xlarge' },
+            { name: 'ml.p5.48xlarge (8× H100 640GB)', value: 'ml.p5.48xlarge' },
+            { name: 'Custom (enter manually)', value: '_custom' }
+        ];
+
+        const cpuChoices = [
+            { name: 'ml.m5.large   (2 vCPU, 8GB — lightweight)', value: 'ml.m5.large' },
+            { name: 'ml.m5.xlarge  (4 vCPU, 16GB — small models)', value: 'ml.m5.xlarge' },
+            { name: 'ml.m5.2xlarge (8 vCPU, 32GB — medium models)', value: 'ml.m5.2xlarge' },
+            { name: 'ml.m5.4xlarge (16 vCPU, 64GB — large models)', value: 'ml.m5.4xlarge' },
+            { name: 'ml.c5.xlarge  (4 vCPU, 8GB — compute-heavy)', value: 'ml.c5.xlarge' },
+            { name: 'ml.c5.2xlarge (8 vCPU, 16GB — compute-heavy)', value: 'ml.c5.2xlarge' },
+            { name: 'Custom (enter manually)', value: '_custom' }
+        ];
+
+        return {
+            type: 'list',
+            name: 'instanceType',
+            message: `Select instance type${isGpu ? ' (GPU recommended for this architecture)' : ''}:`,
+            choices: isGpu ? gpuChoices : cpuChoices
+        };
+    },
+
+    deploymentTarget: (_config) => ({
+        type: 'list',
+        name: 'deploymentTarget',
+        message: 'Select deployment target:',
+        choices: [
+            { name: 'Real-Time Inference', value: 'realtime-inference' },
+            { name: 'Async Inference', value: 'async-inference' },
+            { name: 'Batch Transform', value: 'batch-transform' },
+            { name: 'HyperPod EKS', value: 'hyperpod-eks' }
+        ]
+    }),
+
+    modelFormat: (config) => {
+        const engine = config.engine || 'sklearn';
+        const formatMap = {
+            sklearn: [
+                { name: 'pkl (pickle)', value: 'pkl' },
+                { name: 'joblib', value: 'joblib' }
+            ],
+            xgboost: [
+                { name: 'json', value: 'json' },
+                { name: 'model (binary)', value: 'model' },
+                { name: 'ubj (universal binary JSON)', value: 'ubj' }
+            ],
+            tensorflow: [
+                { name: 'keras', value: 'keras' },
+                { name: 'h5', value: 'h5' },
+                { name: 'SavedModel', value: 'SavedModel' }
+            ]
+        };
+
+        const choices = formatMap[engine] || formatMap.sklearn;
+
+        return {
+            type: 'list',
+            name: 'modelFormat',
+            message: `Select model format for ${engine}:`,
+            choices
+        };
+    },
+
+    awsRegion: (_config) => ({
+        type: 'list',
+        name: 'awsRegion',
+        message: 'Select AWS region:',
+        choices: [
+            { name: 'us-east-1 (N. Virginia)', value: 'us-east-1' },
+            { name: 'us-west-2 (Oregon)', value: 'us-west-2' },
+            { name: 'eu-west-1 (Ireland)', value: 'eu-west-1' },
+            { name: 'ap-northeast-1 (Tokyo)', value: 'ap-northeast-1' },
+            { name: 'ap-southeast-1 (Singapore)', value: 'ap-southeast-1' },
+            { name: 'Custom (enter manually)', value: '_custom' }
+        ]
+    }),
+
+    buildTarget: (_config) => ({
+        type: 'list',
+        name: 'buildTarget',
+        message: 'Select build target:',
+        choices: [
+            { name: 'CodeBuild (recommended)', value: 'codebuild' }
+        ]
+    })
+};

package/src/lib/ci-register-helpers.js

@@ -25,7 +25,7 @@ import { createHash } from 'node:crypto';
  * @param {string} modelName - e.g. "meta-llama/Llama-2-7b-chat-hf", defaults to "none"
  * @param {string} instanceType - e.g. "ml.g5.xlarge"
  * @param {string} region - e.g. "us-east-1"
- * @param {string} deploymentTarget - e.g. "managed-inference"
+ * @param {string} deploymentTarget - e.g. "realtime-inference"
  * @returns {string} 16-character lowercase hex string
  */
 export function computeConfigId(deploymentConfig, modelName, instanceType, region, deploymentTarget) {

package/src/lib/cli-handler.js

@@ -190,7 +190,7 @@ CLI OPTIONS:
   --instance-type=<type>      SageMaker instance type (e.g., ml.m5.large, ml.g5.xlarge)
   --region=<region>           AWS region
   --role-arn=<arn>            AWS IAM role ARN for SageMaker execution
-  --deployment-target=<target> Deployment target (managed-inference|hyperpod-eks)
+  --deployment-target=<target> Deployment target (realtime-inference|async-inference|batch-transform|hyperpod-eks)
   --hyperpod-cluster=<name> HyperPod EKS cluster name
   --hyperpod-namespace=<ns> Kubernetes namespace for HyperPod (default: default)
   --hyperpod-replicas=<n>   Number of replicas for HyperPod (default: 1)