matreche/modelarts-umn
1
0
Fork 0
forked from docs/modelarts
Code Issues Pull Requests Packages Projects Releases Wiki Activity
modelarts-umn/umn/source/examples_of_custom_scripts/scikit_learn.rst
otcbot[bot] 2d9d363ec9
Update content (#29) 
Changes to ma_umn from doc-exports#1


This is an automatically created Pull Request for changes to ma_umn in opentelekomcloud-docs/doc-exports#1.
Please do not edit it manually, since update to the original PR will overwrite local changes.
Original patch file, as well as complete rst archive,  can be found in the artifacts of the opentelekomcloud-docs/doc-exports#1

Reviewed-by: kucerakk <kucerakk@gmail.com>
2022-05-09 12:59:46 +00:00

6.7 KiB
Raw Blame History

original_name

modelarts_23_0179.html

Scikit Learn

Training and Saving a Model

1
2
3
4
5
6
7
8
9
10
11
12
13
14
 
import json
import pandas as pd
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.externals import joblib
iris = pd.read_csv('/data/iris.csv')
X = iris.drop(['virginica'],axis=1)
y = iris[['virginica']]
# Create a LogisticRegression instance and train model
logisticRegression = LogisticRegression(C=1000.0, random_state=0)
logisticRegression.fit(X,y)
# Save model to local path
joblib.dump(logisticRegression, '/tmp/sklearn.m')

After the model is saved, it must be uploaded to the OBS directory before being published. The config.json and customize_service.py files must be contained during publishing. For details about the definition method, see Model Package Specifications <modelarts_23_0091>.

Inference Code

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
 
# coding:utf-8
import collections
import json
from sklearn.externals import joblib
from model_service.python_model_service import XgSklServingBaseService

class user_Service(XgSklServingBaseService):

    # request data preprocess
    def _preprocess(self, data):
        list_data = []
        json_data = json.loads(data, object_pairs_hook=collections.OrderedDict)
        for element in json_data["data"]["req_data"]:
            array = []
            for each in element:
                array.append(element[each])
                list_data.append(array)
        return list_data

    # predict
    def _inference(self, data):
        sk_model = joblib.load(self.model_path)
        pre_result = sk_model.predict(data)
        pre_result = pre_result.tolist()
        return pre_result

    # predict result process
    def _postprocess(self,data):
        resp_data = []
        for element in data:
            resp_data.append({"predictresult": element})
        return resp_data