Roboflow Dataset Upload¶

v2¶

Class: RoboflowDatasetUploadBlockV2 (there are multiple versions of this block)

Source: inference.core.workflows.core_steps.sinks.roboflow.dataset_upload.v2.RoboflowDatasetUploadBlockV2

Warning: This block has multiple versions. Please refer to the specific version for details. You can learn more about how versions work here: Versioning

Block let users save their images and predictions into Roboflow Dataset. Persisting data from production environments helps iteratively building more robust models.

Block provides configuration options to decide how data should be stored and what are the limits to be applied. We advice using this block in combination with rate limiter blocks to effectively collect data that the model struggle with.

Type identifier¶

Use the following identifier in step "type" field: roboflow_core/roboflow_dataset_upload@v2to add the block as as step in your workflow.

Properties¶

Name	Type	Description	Refs
`name`	`str`	Enter a unique identifier for this step..	❌
`target_project`	`str`	name of Roboflow dataset / project to be used as target for collected data.	✅
`usage_quota_name`	`str`	Unique name for Roboflow project pointed by `target_project` parameter, that identifies usage quota applied for this block..	❌
`data_percentage`	`float`	Percent of data that will be saved (in range [0.0, 100.0]).	✅
`persist_predictions`	`bool`	Boolean flag to decide if predictions should be registered along with images.	✅
`minutely_usage_limit`	`int`	Maximum number of data registration requests per minute accounted in scope of single server or whole Roboflow platform, depending on context of usage..	❌
`hourly_usage_limit`	`int`	Maximum number of data registration requests per hour accounted in scope of single server or whole Roboflow platform, depending on context of usage..	❌
`daily_usage_limit`	`int`	Maximum number of data registration requests per day accounted in scope of single server or whole Roboflow platform, depending on context of usage..	❌
`max_image_size`	`Tuple[int, int]`	Maximum size of the image to be registered - bigger images will be downsized preserving aspect ratio. Format of data: `(width, height)`.	❌
`compression_level`	`int`	Compression level for images registered.	❌
`registration_tags`	`List[str]`	Tags to be attached to registered datapoints.	✅
`disable_sink`	`bool`	boolean flag that can be also reference to input - to arbitrarily disable data collection for specific request.	✅
`fire_and_forget`	`bool`	Boolean flag dictating if sink is supposed to be executed in the background, not waiting on status of registration before end of workflow run. Use `True` if best-effort registration is needed, use `False` while debugging and if error handling is needed.	✅
`labeling_batch_prefix`	`str`	Prefix of the name for labeling batches that will be registered in Roboflow app.	✅
`labeling_batches_recreation_frequency`	`str`	Frequency in which new labeling batches are created in Roboflow app. New batches are created with name prefix provided in `labeling_batch_prefix` in given time intervals.Useful in organising labeling flow..	❌

The Refs column marks possibility to parametrise the property with dynamic values available in workflow runtime. See Bindings for more info.

Available Connections¶

Compatible Blocks

Check what blocks you can connect to Roboflow Dataset Upload in version v2.

Input and Output Bindings¶

The available connections depend on its binding kinds. Check what binding kinds Roboflow Dataset Upload in version v2 has.

Bindings

input
- images (image): The image to infer on.
- target_project (roboflow_project): name of Roboflow dataset / project to be used as target for collected data.
- predictions (Union[instance_segmentation_prediction, object_detection_prediction, keypoint_detection_prediction, classification_prediction]): Model predictions to be saved.
- data_percentage (float): Percent of data that will be saved (in range [0.0, 100.0]).
- persist_predictions (boolean): Boolean flag to decide if predictions should be registered along with images.
- registration_tags (string): Tags to be attached to registered datapoints.
- disable_sink (boolean): boolean flag that can be also reference to input - to arbitrarily disable data collection for specific request.
- fire_and_forget (boolean): Boolean flag dictating if sink is supposed to be executed in the background, not waiting on status of registration before end of workflow run. Use True if best-effort registration is needed, use False while debugging and if error handling is needed.
- labeling_batch_prefix (string): Prefix of the name for labeling batches that will be registered in Roboflow app.
output
- error_status (boolean): Boolean flag.
- message (string): String value.

Example JSON definition of step Roboflow Dataset Upload in version v2

{
    "name": "<your_step_name_here>",
    "type": "roboflow_core/roboflow_dataset_upload@v2",
    "images": "$inputs.image",
    "target_project": "my_dataset",
    "usage_quota_name": "quota-for-data-sampling-1",
    "predictions": "$steps.object_detection_model.predictions",
    "data_percentage": true,
    "persist_predictions": true,
    "minutely_usage_limit": 10,
    "hourly_usage_limit": 10,
    "daily_usage_limit": 10,
    "max_image_size": [
        1920,
        1080
    ],
    "compression_level": 95,
    "registration_tags": [
        "location-florida",
        "factory-name",
        "$inputs.dynamic_tag"
    ],
    "disable_sink": true,
    "fire_and_forget": "<block_does_not_provide_example>",
    "labeling_batch_prefix": "my_labeling_batch_name",
    "labeling_batches_recreation_frequency": "never"
}

v1¶

Class: RoboflowDatasetUploadBlockV1 (there are multiple versions of this block)

Source: inference.core.workflows.core_steps.sinks.roboflow.dataset_upload.v1.RoboflowDatasetUploadBlockV1