# Create Dataset

Create a new dataset.

Upload types:
* : Supports up to 200 images, uploading image files
* : Supports up to 5000 images, uploading a zip file

Constraints:
* Dataset must have at least 1 image to be completed
* In basic datasets, a Maximum of 200 images per dataset


When creating a dataset, a default caption prefix is created in all cases.

 When creating a dataset with , 'defined_character' or 'object_variants' IP types, it is  to generate an advanced caption prefix before uploading images.  

To do this, use the  endpoint, send up to 6 images, and update the dataset with the received prefix using the  endpoint. Once the prefix is updated, proceed with uploading images.

 Uploaded images will be automatically resized so that the shortest side is 1024 pixels while maintaining the aspect ratio.  
Then, a centered 1024x1024 crop will be applied. The final cropped image will be saved for training.


Endpoint: POST /tailored-gen/datasets

## Header parameters:

  - `api_token` (string, required)

## Request fields (application/json):

  - `project_id` (integer)
    Associated project ID (required)

  - `name` (string)
    Dataset name (required)

  - `upload_type` (string)
    The method used to upload images to the dataset. 'basic' is the default
    Enum: "basic", "advanced"

## Response 201 fields (application/json):

  - `id` (integer)
    Unique identifier for the dataset

  - `project_id` (integer)
    Associated project ID

  - `name` (string)
    Dataset name

  - `caption_prefix` (string)
    Text automatically prepended to all image captions in the dataset. Each image caption should naturally continues this prefix. A default prefix is automatically created but can be modified, and this same prefix is later used as the default generation prefix during image generation.

  - `status` (string)
    Status of the dataset
    Enum: "draft"

  - `captions_update_status` (string)
    Status of captions update process
    Enum: "empty", "in_progress", "completed"

  - `upload_type` (string)
    Upload type
    Enum: "basic", "advanced"

  - `captions_source` (string)
    The source of the captions. For 'basic' datasets, this is a null. For 'advanced' datasets, this indicates if captions were generated 'automatic' or provided 'manual'. this value is determined during bulk-upload
    Enum: "automatic", "manual"

  - `created_at` (string)
    Timestamp when the dataset was created

  - `updated_at` (string)
    Timestamp when the dataset was last updated

## Response 400 fields (application/json):

  - `errors` (array)

  - `errors.msg` (string)
    A human-readable description of the error.

  - `errors.field_name` (string)
    The name of the field that caused the error.

  - `errors.field_input` (string)
    The input value of the field that caused the error.