Working with Models
This page explains how you can score models.
Contents
Overview
After models are added and configured in TIBCO ModelOps, they become available for approval and can be deployed as a scoring service.
The Models page provides a model-centric view of artifacts in TIBCO ModelOps. You can bind schemas to models on this tab. In addition to this, the Models page displays all the artifacts that are models and provides the following details:
Supported Models
The table below shows the types of models supported in scoring flow.
| Model Name | Model Type Code | Extension(s) | Content Types | Description |
|---|---|---|---|---|
| Predictive Model Markup Language | PMML | pmml | application/xml | A Predictive Model Markup Language (PMML) file. |
| Python Binary | PYTHON_BINARY | pyb | application/octet-stream | A Python Binary file. |
| Python Script | PYTHON_SCRIPT | py | text/plain | A Python Script file. |
| Statistica Model | STATISTICA | sdm | application/octet-stream | A Statistica Model file. |
| TensorFlow Graph | TENSORFLOW_GRAPH | tfg, pb | application/octet-stream | A TensorFlow Graph file. Default extension: tfg. |
| TensorFlow Model | TENSORFLOW_MODEL | tfm, pb | application/octet-stream | A TensorFlow Model file. Default extension: tfm. |
The table below shows the types of models supported for Streaming nodes.
| Model Name | Model Type Code | Extension(s) | Content Types | Description |
|---|---|---|---|---|
| H2O POJO | H2O | pojo | text/plain | An H2O Plain Old Java Object file. |
| Predictive Model Markup Language | PMML | pmml | application/xml | A Predictive Model Markup Language (PMML) file. |
| Python Binary | PYTHON_BINARY | pyb | application/octet-stream | A Python Binary file. |
| Python Script | PYTHON_SCRIPT | py | text/plain | A Python Script file. |
| R Data | R_DATA | rdata,rda | application/octet-stream | An R Data file. Default extension: rdata. |
| R Object | R_OBJECT | rds | application/octet-stream | An R Object file. |
| R Script | R_SCRIPT | r | text/plain | An R script file. |
| Statistica Model | STATISTICA | sdm | application/octet-stream | A Statistica Model file. |
| TensorFlow Graph | TENSORFLOW_GRAPH | tfg, pb | application/octet-stream | A TensorFlow Graph file. Default extension: tfg. |
| TensorFlow Model | TENSORFLOW_MODEL | tfm, pb | application/octet-stream | A TensorFlow Model file. Default extension: tfm. |
Binding Schemas to Model
- Before binding the schemas to the model, make sure that you have created the schemas for that particular model.
- Click the contextual menu of the model that needs schema binding.
- You can know whether a model is ready for scoring or not by looking in the Comment column. Also, you need to bind a model with schemas to deploy it as a scoring service.
- Once you click the down arrow, select the input and output schema from the drop-down list.
- Click the SET BINDING option to bind the schemas to the model.
-
You can also create a schema or delete the current binding by clicking the small drop-down arrow next to the SET BINDING option.
