Upload Custom ML Model in UiPath AI Center | Python ML Model for Future Prediction

Table of Contents

1. Introduction
2. Uploading a Custom Package on UiPath AI Fabric
   • 2.1 Checking the Guidelines for Building Machine Learning Package
   • 2.2 Uploading a Ready-to-Serve ML Model
   • 2.3 Use Case: Predicting Customer Future Visits to a Showroom
3. Creating a Project in UiPath AI Fabric
4. Uploading the Custom Package
   • 4.1 Creating the Folder Structure for the Package
   • 4.2 Creating a Zip File and Uploading the Package
5. Creating an ML Skill and Full Pipeline
   • 5.1 Creating an ML Skill
   • 5.2 Creating a Full Pipeline for Retraining the Model
6. testing the Custom Package in UiPath Studio

Uploading a Custom Package on UiPath AI Fabric

UiPath AI Fabric allows users to upload and deploy custom machine learning packages for use in automation projects. In this Tutorial, we will explore the process of uploading a custom package on UiPath AI Fabric, step by step.

1. Introduction

As automation becomes more prevalent, there is a growing need for integrating machine learning models into UiPath workflows. UiPath AI Fabric provides the capability to upload custom machine learning packages and easily deploy them as skills in automation projects.

2. Uploading a Custom Package on UiPath AI Fabric

Before we dive into the process, let's take a moment to understand the different types of machine learning packages that can be uploaded on UiPath AI Fabric. There are two main types: ready-to-serve ML models and retrainable ML models.

2.1 Checking the Guidelines for Building Machine Learning Package

When building a machine learning package for UiPath AI Fabric, it is important to follow the guidelines provided by UiPath. These guidelines ensure that the package is correctly structured and compatible with the platform. The guidelines cover topics such as folder structure, file requirements, and model formats.

2.2 Uploading a Ready-to-Serve ML Model

A ready-to-serve ML model is a pre-trained model that can be directly uploaded to UiPath AI Fabric. It does not require any additional training and is ready for inference. To upload a ready-to-serve model, you need to have the following files: a pre-trained model file (such as a .sav or .bin file), a main.py file for the prediction class, and a requirements.txt file listing the necessary dependencies.

2.3 Use Case: Predicting Customer Future Visits to a Showroom

In this tutorial, we will focus on a specific use case: predicting whether a customer will visit a showroom again in the future. This use case is Relevant for businesses that want to analyze customer feedback and determine the likelihood of repeat visits. We will train a machine learning model based on historical data and create a package for uploading on UiPath AI Fabric.

3. Creating a Project in UiPath AI Fabric

To start the process, we first need to navigate to UiPath AI Fabric and create a project. The project will serve as the container for our custom package. We will give the project a name and provide a description that accurately reflects its purpose.

4. Uploading the Custom Package

Once the project is created, we can proceed with uploading the custom package. Before uploading, we need to ensure that we have the correct folder structure for the package. This includes having a main.py file for the prediction class, a requirements.txt file listing the dependencies, and any additional files required by the model.

4.1 Creating the Folder Structure for the Package

The folder structure for the package should follow the guidelines provided by UiPath. It should include the necessary files, such as the pre-trained model file, the main.py file, and the requirements.txt file. We will organize these files in a specific structure to ensure compatibility with UiPath AI Fabric.

4.2 Creating a Zip File and Uploading the Package

Once the folder structure is in place, we can create a zip file containing all the files. This zip file will be uploaded to UiPath AI Fabric. The upload process may take some time, especially if the model file is large. However, once uploaded, the package will be available for use in automation projects.

5. Creating an ML Skill and Full Pipeline

In UiPath AI Fabric, an ML skill represents a specific machine learning model that can be used in automation projects. We will create an ML skill for our uploaded custom package. Additionally, we will create a full pipeline for retraining the model over time.

5.1 Creating an ML Skill

To create an ML skill, we need to navigate to the ML Skill section in UiPath AI Fabric and click on "Create New." We will give the skill a name and select the relevant package and version. We can also provide a description if needed. Once created, the ML skill will be available for use in automation projects.

5.2 Creating a Full Pipeline for Retraining the Model

To enable continuous improvement of the machine learning model, we can create a full pipeline for retraining the model over time. This pipeline will execute at a specified time interval and update the model with new data. We will configure the pipeline to retrain the model using a train.py file and evaluate its performance using a test dataset.

6. Testing the Custom Package in UiPath Studio

Once the custom package and ML skill are set up in UiPath AI Fabric, we can test it in UiPath Studio. We will open a workflow in UiPath Studio and import the necessary ML activities. These activities will allow us to use the ML skill and make predictions based on the input data. We will test the package with sample inputs and verify the accuracy of the predictions.

In conclusion, uploading a custom package on UiPath AI Fabric opens up a world of possibilities for integrating machine learning models into automation projects. By following the provided guidelines and using the tools available in UiPath Studio, users can create powerful automation solutions that leverage the capabilities of machine learning.

Pros:

• Easy integration of machine learning models into UiPath workflows
• Flexibility to upload both ready-to-serve and retrainable ML models
• Ability to create ML skills and full pipelines for continuous improvement

Cons:

• Requires proper understanding of machine learning concepts and model building process
• Initial setup and configuration may require some time and effort

Highlights

• Uploading a custom machine learning package on UiPath AI Fabric
• Check guidelines for building machine learning packages
• Upload ready-to-serve or retrainable machine learning models
• Use case: predicting customer future visits to a showroom
• Create a project and upload the custom package
• Organize the folder structure and create a zip file
• Create an ML skill and full pipeline for retraining the model
• Test the custom package in UiPath Studio
• Integrate machine learning into automation projects
• Improve accuracy and efficiency of automation tasks

FAQ:

Q: What is UiPath AI Fabric? A: UiPath AI Fabric is a component of the UiPath platform that allows users to upload and deploy custom machine learning models for use in automation projects.

Q: What types of machine learning packages can be uploaded on UiPath AI Fabric? A: UiPath AI Fabric supports two main types of machine learning packages: ready-to-serve ML models and retrainable ML models.

Q: Can I upload my own pre-trained machine learning model on UiPath AI Fabric? A: Yes, you can upload a pre-trained machine learning model as a ready-to-serve ML model on UiPath AI Fabric. Just make sure to follow the guidelines for building the package.

Q: Can I retrain the machine learning model after uploading it on UiPath AI Fabric? A: Yes, UiPath AI Fabric allows you to create a full pipeline for retraining the machine learning model over time. This enables continuous improvement of the model's performance.

Q: How can I test the accuracy of the custom package in UiPath Studio? A: In UiPath Studio, you can use the ML activities provided by UiPath AI Fabric to test the accuracy of the custom package. You can input sample data and verify the predictions made by the model.