Kinds¶
In Workflows, some values can’t be set in advance and are only determined during execution. This is similar to writing a function where you don’t know the exact input values upfront — they’re only provided at runtime, either from user inputs or from other function outputs.
To manage this, Workflows use selectors, which act like references, pointing to data without containing it directly.
selectors
Selectors might refer to a named input - for example input image - like $inputs.image
or predictions generated by a previous step - like $steps.my_model.predictions
In the Workflows ecosystem, users focus on data purpose (e.g., “image”) without worrying about its exact format. Meanwhile, developers building workflow blocks need precise data formats. Kinds serve both needs - they simplify data handling for users while ensuring developers work with the correct data structure.
What are the Kinds?¶
Kinds is Workflows type system with each kind defining:
-
name - expressing semantic meaning of the underlying data - like
image
orpoint
; -
Python data representation - the data type and format that blocks creators should expect when handling the data within blocks;
-
optional serialized data representation - defining what is the format of the kind that external systems should use to integrate with Workflows ecosystem - when needed, custom kinds serializers and deserializers are provided to ensure seamless translation;
Using kinds streamlines compatibility: when a step outputs data of a certain kind and another step requires that same kind, the workflow engine assumes they’ll be compatible, reducing the need for compatibility checks and providing compile-time verification of Workflows definitions.
Note
As for now, kinds
are such simplistic that do not support types polymorphism - and developers
are asked to use unions of kinds to solve that problem. As defining extensive unions of kinds may be
problematic, this problem will probably be addressed in Execution Engine v2
.
Warning
In inference
release 0.18.0
we decided to make drastic move to heal the ecosystem
from the problem with ambiguous kinds names (Batch[X]
vs X
- see more
here).
The change is breaking only if there is remote Workflow plugin depending on imports
from inference.core.workflows.execution_engine.entities.types
module, which is
not the case to the best of our knowledge. We removed problematic kinds as if they
never existed in the ecosystem and fixed all blocks from roboflow_core
plugin.
If there is anyone impacted by the change - here is the
migration guide.
This warning will be removed end of Q1 2025.
Warning
Support for proper serialization and deserialization of any arbitrary kind was
introduced in Execution Engine v1.3.0
(released with inference 0.26.0
). Workflows
plugins created prior that change may be updated - see refreshed
Blocks Bundling page.
This warning will be removed end of Q1 2025.
Kinds declared in Roboflow plugins¶
top_class
: String value representing top class predicted by classification modelfloat
: Float valueinteger
: Integer valueroboflow_api_key
: Roboflow API keyqr_code_detection
: Prediction with QR code detection*
: Equivalent of any elementdictionary
: Dictionaryimage_metadata
: Dictionary with image metadata required by supervisionrgb_color
: RGB colorinstance_segmentation_prediction
: Prediction with detected bounding boxes and segmentation masks in form of sv.Detections(...) objectbar_code_detection
: Prediction with barcode detectionroboflow_model_id
: Roboflow model idobject_detection_prediction
: Prediction with detected bounding boxes in form of sv.Detections(...) objectkeypoint_detection_prediction
: Prediction with detected bounding boxes and detected keypoints in form of sv.Detections(...) objectparent_id
: Identifier of parent for step outputvideo_metadata
: Video image metadatabytes
: This kind represent byteszone
: Definition of polygon zoneserialised_payloads
: Serialised element that is usually accepted by sinkcontours
: List of numpy arrays where each array represents contour pointsprediction_type
: String value with type of predictionclassification_prediction
: Predictions from classifiersecret
: Secret valuenumpy_array
: Numpy arraylist_of_values
: List of values of any typelanguage_model_output
: LLM / VLM outputboolean
: Boolean flagfloat_zero_to_one
:float
value in range[0.0, 1.0]
image
: Image in workflowsimage_keypoints
: Image keypoints detected by classical Computer Vision methodstring
: String valueinference_id
: Inference identifierpoint
: Single point in 2Droboflow_project
: Roboflow project namedetection
: Single element of detections-based prediction (likeobject_detection_prediction
)