Skip to content

POST /v1/lamini/completions

This endpoint allows you to make a POST request to obtain a model completion. Input and Output data types are required for the model to properly ingest input data and produce properly formatted output data.

Request

HTTP Method: POST

URL: https://api.powerml.co/v1/lamini/completions

Headers:

  • Authorization: Bearer <LAMINI_API_KEY>
  • Content-Type: application/json

Example Body (JSON):

{
  "id": "APIExample",
  "model_name": "openaccess-ai-collective/wizard-mega-13b",
  "input_type": {
    "title": "Question",
    "properties": {
      "question": {
        "description": "A question",
        "type": "string"
      }
    }
  },
  "output_type": {
    "title": "Answer",
    "properties": {
      "answer": {
        "description": "An answer to the question",
        "type": "string"
      }
    }
  },
  "input_value": {
    "question": "What is the hottest day of the year?"
  }
}

Parameters

  • id: string: An id which will allow you to iterate on trained models
  • model_name: string: The name of your base or trained model. This can be any openai or huggingface model name.
  • input_type: dict: Type Schema of the input. Input type must be a dictionary with format 
    {
    "title": <TYPE_NAME>,
    "properties": {
        <FIELD_NAME>: {
            "description": <FIELD_DESCRIPTION>,
            "type": "string" | "integer" | "number" | "boolean"
        }, ...
    }
}
  • output_type: dict: Type Schema of the output. Output type has the same format as input type.
  • input_value: dict: An Input Datapoint. Input value must be a dictionary with format 
    {
    <FIELD_NAME>: "Brainstorm 20 compelling headlines for a Facebook ad promoting the Best Business Financing Options for [Business Owners]. Format the output as a table.",
    ...
}

Response

Body (JSON):

  • Output will be formatted as specified by the output_type argument passed in above.
{
  "answer": "The hottest day of the year varies depending on the location, but generally, it occurs during the summer months when the sun is closest to the Earth. In many regions, July or August tend to be the hottest months."
}

Example

Request

curl --location 'https://api.powerml.co/v1/lamini/completions' \
--header 'Authorization: Bearer <LAMINI_API_KEY>' \
--header 'Content-Type: application/json' \
--data '{
    "id": "LaminiTest",
    "model_name": "352b52d9f194debbf8258bf7763176d3812579bc59ecc80f0f24d8a7c7bf7658",
    "input_type": {
        "title": "Question",
        "properties": {
            "question": {
                "description": "A question",
                "type": "string"
            }
        }
    },
    "output_type": {
        "title": "Answer",
        "properties": {
            "answer": {
                "description": "An answer to the question",
                "type": "string"
            }
        }
    },
    "input_value": {
        "question": "What is Lamini?"
    }
}'

Response

{
  "answer": "Lamini is a Python package that provides a simple interface for training and using language models. It uses the Large Language Model (LLM) engine from the llama library to interface with the model."
}