Select Artifact
Load saved model weights, label encoders, and other outputs from previous flow runs.
The Select Artifact node loads outputs saved from previous flow runs — most commonly trained model weights or label encoders — so they can be used in an inference flow without retraining.
This is the typical starting point for a deployment or inference flow: select the artifact from the training run you want to use, then connect it to a Use AI Model node.
Configuration
| Field | Description |
|---|---|
| Tag | Optional label for this artifact selection, useful when referencing it in downstream nodes |
After adding the node, use the property panel to browse and select the artifact from your previous flow runs.
Outputs
| Output | Description |
|---|---|
selected_artifacts | The artifact(s) selected (model weights, encoders, or other saved files) |
tag | The tag applied to this selection |
Tips
- Artifacts are created automatically when you run a flow containing a Train AI Model node
- Use the Tag field when loading multiple artifacts in the same flow (e.g. a model and its label encoder separately)
- Connect the artifact output directly into the
trained_modelinput of a Use AI Model node