to_onnx fail on sklearn.linear_model.ElasticNet for 12 inputs

[bardout]

Summary

after fit on a ElasticNet on a regression, convert onnx, based on test_x, default parameters, fail to convert_operators @ iteration 1 - n_vars=0 n_ops=1, for operator SklearnLinearRegressor : at most 1 input(s) is(are) supported but we got 12 input(s). Search issue "LinearRegressor" did not help

expected results

I expect to convert the model to use on onnx_engine. Maybe there is a mistake in the train or the convert, the code is naive.

data

from kaggle: the wine-quality-dataset, tabular data, 12 input variables (float, based on physicochemical tests), one Output variable (based on sensory data): the quality (score between 0 and 10 with 2 decimals).

code

import warnings
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.linear_model import ElasticNet
warnings.filterwarnings("ignore")
np.random.seed(40)

# Read the wine-quality csv file (make sure you're running this from the root of MLflow!)
data = pd.read_csv("WineQT.csv"))

# Split the data into training and test sets. (0.75, 0.25) split.
train, test = train_test_split(data)

# The predicted column is "quality" which is a scalar from [3, 9]
train_x = train.drop(["quality"], axis=1)
test_x = test.drop(["quality"], axis=1)
train_y = train[["quality"]]
test_y = test[["quality"]]
lr = ElasticNet(alpha=0.2, l1_ratio=0.2, random_state=42)
lr.fit(train_x, train_y)

import skl2onnx
try:
    print('--- convert onnx, based on test_x, default parameters ---')
    onx = skl2onnx.to_onnx(lr, test_x, model_optim=True, verbose=3) # optional parameters TBD
    with open("D00_DATA\D03_MODEL\model.onnx", "wb") as f:
        f.write(onx.SerializeToString())
except Exception as exc: # Error  to_onnx or IO
    print(exc)

log verbose=3

[to_onnx] initial_types=[('fixed acidity', DoubleTensorType(shape=[None, 1])), ('volatile acidity', DoubleTensorType(shape=[None, 1])), ('citric acid', DoubleTensorType(shape=[None, 1])), ('residual sugar', DoubleTensorType(shape=[None, 1])), ('chlorides', DoubleTensorType(shape=[None, 1])), ('free sulfur dioxide', DoubleTensorType(shape=[None, 1])), ('total sulfur dioxide', DoubleTensorType(shape=[None, 1])), ('density', DoubleTensorType(shape=[None, 1])), ('pH', DoubleTensorType(shape=[None, 1])), ('sulphates', DoubleTensorType(shape=[None, 1])), ('alcohol', DoubleTensorType(shape=[None, 1])), ('Id', Int64TensorType(shape=[None, 1]))]
[convert_sklearn] parse_sklearn_model
[convert_sklearn] convert_topology
[convert_operators] begin
[convert_operators] iteration 1 - n_vars=0 n_ops=1
For operator SklearnLinearRegressor (type: SklearnLinearRegressor), at most 1 input(s) is(are) supported but we got 12 input(s) which are ['fixed_acidity', 'volatile_acidity', 'citric_acid', 'residual_sugar', 'chlorides', 'free_sulfur_dioxide', 'total_sulfur_dioxide', 'density', 'pH', 'sulphates', 'alcohol', 'Id']