Welcome
Developer Guide
Model Inference (beta)
Inference
The inference endpoint is used for making predictions using a pretrained model on data provided by the user.

Release Notes
Installation Guide
Developer Guide
- Overview
  To access the Marketing AION REST API interface, use the Docker Desktop application to start the Docker container.
- Tracker
- API Status
  Use these APIs to get the status and set the status of the different types of API.
- Data Handler Overview
  Data Handler has four different APIs, and three of the APIs represent a different method uploading data.
- Training
  For Training, you can execute either of the generic training by using Train for pipeline API.
- Prediction
  The Prediction section lists APIs that are used for handling single prediction and batch prediction.
- Model Inference (beta)
  - Inference
    The inference endpoint is used for making predictions using a pretrained model on data provided by the user.
- Download-model
  These APIs are used to download Python package, Docker container package, and Mlac.
- Monitoring
  Monitoring ML models checks for changes in data distribution ensuring that data is maintaining an acceptable level of variation.
- Xplain
  Explainability gives an explanation about Prediction and Model.
- Usecases
  Use the following APIs to retrieve the Use Case list, to delete use cases, and to retrieve the logs.
- EDA
- AutoPilot
  This Autopilot Mode refers to a feature that allows users to build an automated model without direct human intervention.
- Data Generator
- Building Usecases with API
  To build a usecase or to train a model, complete the following steps:
MLAC User Guide
User Guide

Inference

The inference endpoint is used for making predictions using a pretrained model on data provided by the user.

Method

POST

Endpoint

/v1/aion/inference/

Request Parameters


data_file	Mandatory	The CSV file containing the data for prediction. This file should be uploaded by the user and is described as "Upload CSV file for prediction".
data_model	Mandatory	The pretrained model file. This file should be uploaded by the user and is described as "Upload pre trained model file for prediction".

Response Parameters

If the inference is successful, the API returns a JSON response containing the predictions.

predictions (array): An array of prediction results. Each prediction contains the original data columns and the predicted value.

In case of an error, the response will contain an error message and a status code indicating the nature of the issue.