topic Re: Using ONNX Serialize a Traditional ML Model in SAS Data Science

Using ONNX Serialize a Traditional ML Model

DaisyQL — Thu, 10 Dec 2020 16:19:37 GMT

Hi,

I saw the example of deploying a dlpy model into ONNX format. But I am wondering how to save a regular ML model (not from dlpy) using ONNX.

Here is my model:

conn.loadActionSet('decisionTree')
model = conn.decisionTree.forestTrain(
     table = dict(name = 'titanic_train'),
     target = target,
     inputs = inputs,
     nominals = nominals,
     casOut = dict(name = 'titanic_forest_model', replace = True)
)

Thank you so much!

Re: Using ONNX Serialize a Traditional ML Model

ximeng — Sat, 12 Dec 2020 19:55:15 GMT

Hello, we currently only support generating ONNX format from the SAS deep learning models. We haven't support saving a SAS ML model such as tree, boosting, regressions, etc to ONNX format.

Re: Using ONNX Serialize a Traditional ML Model

BadshahMukherje — Wed, 03 Nov 2021 18:13:30 GMT

The new Open Source packages for Python, sas-scoring-translator-python (pysct), and for R, sas-scoring-translator-r (rsct), which translate (wrap) SAS scoring code to those languages.

https://communities.sas.com/t5/SAS-Communities-Library/SAS-Scoring-code-translator-for-Python-and-R/ta-p/759542