Create Datasets

This article explains how to create and manage datasets in Hyperstack AI Studio using the API or UI. Datasets help you organize logs for fine-tuning, training analysis, and efficient model iteration.

In this article

• Create Datasets Using the API
• Create Datasets Using the UI

---

Create Datasets Using the API

To create a dataset via API, use the command below and replace the following variables:

• API_KEY: Your AI Studio API key.
• name: A unique name to identify your dataset.
• At least one of tags or models is required to define the log selection criteria.
• Optional fields such as description, tags, models, from_date, and to_date can be included to filter logs or add metadata. See the full list in the Optional Parameters section.

curl -X POST "https://api.genai.hyperstack.cloud/tailor/v1/datasets" \
  -H "X-API-KEY: API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "YOUR_DATASET_NAME",
    "tags": ["tag1", "tag2"],
    "models": ["model1", "model2"],
    "description": "A brief description of your dataset",
    "from_date": "2024-01-01T00:00:00Z",
    "to_date": "2024-12-31T23:59:59Z"
  }'

Required Parameters

• name (string) – The name of the dataset. Must be unique for the user.
• tags or models – You must specify at least one of these fields to filter logs for inclusion.

Optional Parameters

• description (string) – A brief description of the dataset.
• tags (array of strings) – Array of tags to filter the dataset by.
• models (array of strings) – Array of model names to filter the dataset by.
• from_date (string) – Start date for filtering data (ISO 8601 format).
• to_date (string) – End date for filtering data (ISO 8601 format).

Response

If successful, the API will return the following confirmation response:

{
  "status": "success",
  "message": "Successfully created dataset {name}"
}

Error Responses

• 400 Bad Request: If the name parameter is missing
• 409 Conflict: If a dataset with the same name already exists for the user

---

Create Datasets Using the UI

You can create datasets directly through the AI Studio UI using the following steps:

1. Open the Logs & Datasets Page

   Navigate to the Logs & Datasets page.

2. Filter Logs by Tags or Models

   Use the filtering options to select logs associated with specific tags or models.

   Filter requirements

   At least one tag or model must be selected to create a dataset.

3. Review and Select Logs

   Once filters are applied, review the matching logs displayed in the list.

4. Create Dataset

   Click the Create Dataset button. Provide a unique name for your dataset and optionally add a description.

5. Save Dataset

   Click Save to finalize dataset creation. The new dataset will appear under the Datasets tab.

Once created, you can:

• View all logs associated with the dataset.
• Add or remove metadata tags.
• Use the dataset for fine-tuning or evaluation tasks within Hyperstack AI Studio.