Advanced Micro Devices Models API

openapi: 3.0.3
info:
  title: AMD Developer Cloud Configuration Models API
  description: The AMD Developer Cloud API provides access to AMD Instinct GPU instances for AI inference, training, and HPC workloads. Supports managing compute instances, deploying AI models, monitoring GPU utilization, and integrating with ROCm-compatible frameworks including PyTorch, TensorFlow, and vLLM.
  version: '1'
  contact:
    name: AMD Developer Support
    url: https://developer.amd.com/support/
  termsOfService: https://www.amd.com/en/legal/terms-and-conditions.html
  license:
    name: AMD Terms and Conditions
    url: https://www.amd.com/en/legal/terms-and-conditions.html
servers:
- url: https://api.developer.amd.com/v1
  description: AMD Developer Cloud API Production
security:
- apiKey: []
tags:
- name: Models
  description: AI model deployment and management
paths:
  /models:
    get:
      operationId: listModels
      summary: AMD Developer Cloud List Deployed Models
      description: List AI models deployed on the AMD Developer Cloud for inference serving.
      tags:
      - Models
      responses:
        '200':
          description: List of deployed models.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ModelList'
              examples:
                listModels200Example:
                  summary: Default listModels 200 response
                  x-microcks-default: true
                  value:
                    models:
                    - id: model-001
                      name: llama-3.1-70b
                      framework: vLLM
                      status: serving
                      endpoint: https://inf.developer.amd.com/model-001/v1
        '401':
          description: Unauthorized.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
              examples:
                listModels401Example:
                  summary: Default listModels 401 response
                  x-microcks-default: true
                  value:
                    code: UNAUTHORIZED
                    message: Invalid API key
      x-microcks-operation:
        delay: 0
        dispatcher: FALLBACK
    post:
      operationId: deployModel
      summary: AMD Developer Cloud Deploy AI Model
      description: Deploy an AI model for inference serving on AMD Instinct GPUs using vLLM or another serving framework.
      tags:
      - Models
      requestBody:
        required: true
        description: Model deployment configuration.
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ModelInput'
            examples:
              deployModelRequestExample:
                summary: Default deployModel request
                x-microcks-default: true
                value:
                  name: llama-3.1-70b
                  modelId: meta-llama/Llama-3.1-70B-Instruct
                  framework: vLLM
                  gpuType: mi300x
                  gpuCount: 2
      responses:
        '201':
          description: Model deployed.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Model'
              examples:
                deployModel201Example:
                  summary: Default deployModel 201 response
                  x-microcks-default: true
                  value:
                    id: model-002
                    name: llama-3.1-70b
                    framework: vLLM
                    status: deploying
                    endpoint: https://inf.developer.amd.com/model-002/v1
        '401':
          description: Unauthorized.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
              examples:
                deployModel401Example:
                  summary: Default deployModel 401 response
                  x-microcks-default: true
                  value:
                    code: UNAUTHORIZED
                    message: Invalid API key
      x-microcks-operation:
        delay: 0
        dispatcher: FALLBACK
components:
  schemas:
    Model:
      type: object
      description: A deployed AI model for inference serving.
      properties:
        id:
          type: string
          description: Model deployment identifier.
          example: model-001
        name:
          type: string
          description: Model display name.
          example: llama-3.1-70b
        framework:
          type: string
          description: Inference serving framework.
          enum:
          - vLLM
          - TGI
          - TorchServe
          - Triton
          example: vLLM
        status:
          type: string
          description: Deployment status.
          enum:
          - deploying
          - serving
          - stopped
          - failed
          example: serving
        endpoint:
          type: string
          description: Inference endpoint URL.
          example: https://inf.developer.amd.com/model-001/v1
    ModelList:
      type: object
      description: List of deployed AI models.
      properties:
        models:
          type: array
          description: Array of model deployments.
          items:
            $ref: '#/components/schemas/Model'
    ModelInput:
      type: object
      description: Request to deploy an AI model.
      required:
      - modelId
      - framework
      - gpuType
      properties:
        name:
          type: string
          description: Display name for the deployment.
          example: llama-3.1-70b
        modelId:
          type: string
          description: Hugging Face model ID or path.
          example: meta-llama/Llama-3.1-70B-Instruct
        framework:
          type: string
          description: Inference framework to use.
          enum:
          - vLLM
          - TGI
          - TorchServe
          - Triton
          example: vLLM
        gpuType:
          type: string
          description: GPU hardware type.
          example: mi300x
        gpuCount:
          type: integer
          description: Number of GPUs to allocate.
          example: 2
    ErrorResponse:
      type: object
      description: API error response.
      properties:
        code:
          type: string
          description: Error code.
          example: NOT_FOUND
        message:
          type: string
          description: Error message.
          example: The requested resource was not found.
  securitySchemes:
    apiKey:
      type: apiKey
      in: header
      name: X-API-Key