Select a compute environment

MOSTLY AI assigns the most appropriate compute for each new synthetic dataset. For example, if the generator includes fine-tuned LLMs, MOSTLY AI assigns a GPU compute, if one is available.

The assigned compute is used to perform the tasks listed below:

• generate the synthetic data for each table
• generate a Data report for each table (if the Data report is enabled).

You always have the option to select a different compute for a synthetic dataset.

Steps

1. On the Synthetic dataset configuration page, click Compute in the upper right.
2. From the Compute pop-up window, select a compute.
3. Click Proceed.

What’s next

You can now generate the synthetic dataset with the selected compute.

In addition, you can configure any of the remaining options on the Synthetic dataset configuration page.

• Set sample size and temperature
• Rebalance columns
• Data imputation
• Configure fairness
• Seeded generation
• Deliver to a database or a cloud storage bucket