Supported frameworks (Watson Machine Learning)

You can use popular tools, libraries, and frameworks to train and deploy machine learning models using IBM Watson Machine Learning. This topic lists supported versions and features.

Python deprecation notice

Due to a security vulnerability, Python 3.6 is deprecated in favor of Python 3.7. If a framework is marked as deprecated, then support for the framework will be removed in a future release. For details on migrating a model or function, see Notes about upgrading models from a deprecated framework later in this topic.

Supported machine learning frameworks

Framework Versions Online Batch CoreML
Spark 2.3
2.4
Training requires 2.4
Yes Yes
Inline payload only
No
PMML 3.0 to 4.3 Yes Programmatic only
Inline payload only
No
Hybrid/AutoML 0.1 Yes Yes No
SPSS
18.1
18.2
Yes Yes No
Scikit-learn 0.20 (deprecated)
0.22 (deprecated)
0.23
Yes Yes Yes
XGBoost 0.80 (deprecated)
0.90
Training requires 0.90
Yes Yes Yes
TensorFlow 1.15 (deprecated)
2.1 (deprecated)
Training requires Watson Machine Learning Accelerator 1.2.1
Yes Yes No
Keras 2.2.5 (deprecated)
Training requires Watson Machine Learning Accelerator 1.2.1
Yes Yes No
Caffe 1.0 (deprecated) Yes Yes No
PyTorch 1.1 (deprecated)
1.2 (deprecated)
1.3
Training requires Watson Machine Learning Accelerator 1.2.1
Yes Yes No
Decision Optimization 12.9
12.10
No Yes No
Python function 0.1 yes Programmatic only
Inline payload only
no
Python scripts 1.0 No Yes No

Notes about upgrading models from a deprecated framework

If you upgrade from IBM Cloud Pak for Data 2.5, deployments based on frameworks that are not supported in Cloud Pak for Data 3.0 are removed as part of the upgrade. To address this, you have these options following the upgrade:

Option 1: Save the model with a compatible framework

  1. Download the model from the Watson Machine Learning repository.
  2. Save the model back to the Watson Machine Learning repository with a model type and version supported in Cloud Pak for data 3.0.
  3. Deploy the model.
  4. Score the model to generate predictions.

If there is a failure when you deploy or score the model, it means that the model is not compatible with the new version used for saving the model. In this case, use Option 2.

Option 2: Retrain the model with a compatible framework

  1. Retrain the model with a model type and version supported in Cloud Pak for Data 3.0.
  2. Save the model to the Watson Machine Learning repository with the supported model type and version.
  3. Deploy and score the model.

Notes about upgrading Python functions from a deprecated framework

These notes apply to IBM Cloud Pak for Data with Patch 5 applied.

Python function

Option 1: Save the python function with a compatible runtime or software specification:

  1. Download the Python function from the Watson Machine Learning repository.
  2. Save the Python function back to the Watson Machine Learning repository with a supported runtime or software specification version.
  3. Deploy the Python function.
  4. Score the Python function to generate predictions.

If there is a failure when you score the Python function, it means that the function is not compatible with the new runtime or software specification version used for saving the Python function. In this case, use Option 2.

Option 2: Modify the function code and save it with a compatible runtime or software specification

SPSS Modeler flows might need retraining

If your SPSS model uses any of those modeler nodes, take the following action: