Oracle Cloud Infrastructure Documentation

Creating a Custom Model in Generative AI

Create a custom model in OCI Generative AI by fine-tuning a base model with your own dataset.

Before you create a model, ensure that you have permission to create custom models and list buckets, and that your dataset meets the training data requirements.
  1. In the navigation bar of the Console, select a region with Generative AI, for example, US Midwest (Chicago) or UK South (London). See which models are offered in your region.
  2. Open the navigation menu and click Analytics & AI. Under AI Services, click Generative AI.
  3. Select a compartment to create the custom model in.
  4. Click Custom models.
  5. Click Create model.
  6. Select Create a new model.
  7. Select a compartment to create the model in. The default compartment is the one that you selected in step 3, but you can select any compartment that you have permission to work in.
  8. (Optional) Enter a name for the custom model. Start the name with a letter or underscore, followed by letters, numbers, hyphens, or underscores. The length can be 1–255 characters. If you don't enter a name, the system generates a name that you can change later.

    The generated name has the format generativeaimodel<timestamp>. For example: generativeaimodel20240531234930

  9. (Optional) Enter a version for the model. If you don't enter a version, the system generates a version that you can change later.

    The generated version has the format v<timestamp>. For example: v20240531234930

  10. (Optional) Enter a description for the model.
  11. Click Next.
  12. Select a Base Model for the custom model.
    For details about the base model, see the key features of the pretrained models.
  13. Select a Fine-tuning method.
    Important

    For guidelines, see Choosing a Fine-Tuning Method in Generative AI.
  14. Select a fine-tuning dedicated AI cluster by performing one of the following actions:
    • Select a Dedicated AI cluster from the drop-down list. If you just created a cluster, wait for that cluster to become active. Ensure that the base model that's associated with this cluster matches the base model in step 12.
    • In the Dedicated AI cluster drop-down list, click Create new dedicated AI cluster and perform the following steps:
      1. (Optional) Enter a name and description.
      2. Select a Base model that matches the base model in step 12.
      3. Read the commitment unit hours for the fine-tuning dedicated AI cluster and select the checkbox to agree to the commitment.
      4. Click Create and wait for the cluster to become active.
      5. From the Dedicated AI cluster drop-down list, click the dedicated AI cluster that you created.
  15. (Optional) Under Advanced options, click Show hyperparameters, review the tooltips and hint text and update the values as needed. Also see Hyperparameters for Fine-Tuning a Model in Generative AI. To reset the values, click Restore defaults.
  16. Click Next.
  17. Select the Object Storage bucket that contains the training dataset. If the bucket isn't listed, perform the following actions:
    • Ensure that the bucket is in the same region as the custom model.
    • Click Change compartment and select the compartment that hosts the bucket with the training data.
    • Ask an administrator to give you permission to buckets and objects in this compartment.
  18. Select the Training file to use for this model.
  19. Preview how the data in the training file is ingested.
    • If the data records parse correctly, click Submit.
    • Otherwise, fix the data.
After the model is created, you can get the custom model's details from the detail page.
Note

Custom models aren't available in all regions. See Pretrained Foundational Models in Generative AI for the regions that a custom model is available in. For example, a custom text generation and summarization model with the base model of cohere.command, is only available in US Midwest (Chicago).