Using Future AGI Platform

This guide covers all the methods available for adding datasets through the Future AGI platform interface.

​

Prerequisites

Before adding a dataset, ensure you:

• Have access to the Future AGI platform
• Are logged into your account
• Have the necessary data or files ready

​

Methods to Add Datasets

The Future AGI platform offers multiple ways to add datasets:

​

Common First Step

For all methods, begin by:

1. Select Build from the top left corner under the Main section
2. Click on Add Dataset

​

Method 1: Uploading Files (JSON, CSV)

​

1. Select Upload Option

Choose the Upload a file option to upload files from your local computer.

​

2. Configure and Upload

1. Enter a name for your dataset
2. Choose your model type from the drop-down menu
3. Browse for your file and select it
4. Click Done and wait for upload completion

​

Method 2: Manual Dataset Creation

​

1. Select Manual Creation

Choose Manually Create Dataset from the options.

​

2. Configure Settings

1. Enter a descriptive name
2. Select appropriate model type
3. Click Create

​

Method 3: Importing from Hugging Face

​

1. Select Hugging Face Option

Choose Import from Hugging Face option.

​

2. Provide Dataset Details

1. Copy dataset name from Hugging Face Datasets
2. Paste the dataset name
3. Click Load Dataset

​

3. Configure Import

1. Give your dataset a name
2. Choose model type
3. Select dataset subset and split
4. Click Start Experimenting

​

Method 4: Adding from Existing Dataset

​

1. Select Existing Dataset Option

Choose Add from existing dataset option.

​

2. Configure New Dataset

1. Assign a name
2. Select source dataset
3. Choose model type

​

Verifying Dataset Creation

After using any method above:

1. Your new dataset should appear on the dashboard
2. If not visible, refresh the page
3. You can now begin experimenting with your dataset

​

Next Steps

Once your dataset is added, you can:

• Run prompts on your dataset
• Create experiments to test different configurations
• Evaluate results using various metrics