In [1]:
%reload_ext autoreload
%autoreload 2
%matplotlib inline
import warnings
warnings.filterwarnings("ignore")
In [ ]:
!pip install bentoml
!pip install xgboost numpy pandas
In [2]:
import pandas as pd
import numpy as np
import xgboost as xgb
import bentoml
Prepare Dataset¶
download dataset from https://www.kaggle.com/c/titanic/data
In [3]:
!mkdir data
!curl https://raw.githubusercontent.com/agconti/kaggle-titanic/master/data/train.csv -o ./data/train.csv
!curl https://raw.githubusercontent.com/agconti/kaggle-titanic/master/data/test.csv -o ./data/test.csv
In [4]:
train = pd.read_csv("./data/train.csv")
test = pd.read_csv("./data/test.csv")
X_y_train = xgb.DMatrix(data=train[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch']], label= train['Survived'])
X_test = xgb.DMatrix(data=test[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch']])
In [5]:
train[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch', 'Survived']].head()
Out[5]:
Model Training¶
In [6]:
params = {
'base_score': np.mean(train['Survived']),
'eta': 0.1,
'max_depth': 3,
'gamma' :3,
'objective' :'reg:linear',
'eval_metric' :'mae'
}
model = xgb.train(params=params,
dtrain=X_y_train,
num_boost_round=3)
In [7]:
y_test = model.predict(X_test)
test['pred'] = y_test
test[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch','pred']].iloc[10:].head(2)
Out[7]:
Create BentoService for model serving¶
In [8]:
%%writefile xgboost_titanic_bento_service.py
import xgboost as xgb
import bentoml
from bentoml.artifact import XgboostModelArtifact
from bentoml.handlers import DataframeHandler
@bentoml.artifacts([XgboostModelArtifact('model')])
@bentoml.env(pip_dependencies=['xgboost'])
class TitanicSurvivalPredictionService(bentoml.BentoService):
@bentoml.api(DataframeHandler)
def predict(self, df):
data = xgb.DMatrix(data=df[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch']])
return self.artifacts.model.predict(data)
Save BentoML service archive¶
In [9]:
# 1) import the custom BentoService defined above
from xgboost_titanic_bento_service import TitanicSurvivalPredictionService
# 2) `pack` it with required artifacts
bento_service = TitanicSurvivalPredictionService()
bento_service.pack('model', model)
# 3) save your BentoSerivce
saved_path = bento_service.save()
Load saved BentoService for serving¶
In [10]:
import bentoml
bento_model = bentoml.load(saved_path)
result = bento_model.predict(test)
test['pred'] = result
test[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch','pred']].iloc[10:].head(2)
Out[10]:
Model Serving via REST API¶
In your termnial, run the following command to start the REST API server:
In [14]:
!bentoml serve {saved_path}
Copy following command to make a curl request to Rest API server
curl -i \
--header "Content-Type: application/json" \
--request POST \
--data '[{"Pclass": 1, "Age": 30, "Fare": 200, "SibSp": 1, "Parch": 0}]' \
localhost:5000/predict
Deploy it as API endpoint on AWS Lambda¶
In order to run this as AWS Lambda function, make sure to configure your AWS credentials via either aws configure command or setting the environment variables below:
%env AWS_ACCESS_KEY_ID=
%env AWS_SECRET_ACCESS_KEY=
%env AWS_DEFAULT_REGION=Make sure you have AWS SAM CLI installed:
In [ ]:
%env AWS_ACCESS_KEY_ID=
%env AWS_SECRET_ACCESS_KEY=
%env AWS_DEFAULT_REGION=
Make sure you have AWS SAM CLI installed:
In [ ]:
!pip install -U aws-sam-cli==0.33.1
In [16]:
!bentoml deployment create titanic-prediction \
--bento=TitanicSurvivalPredictionService:{bento_service.version} \
--platform=aws-lambda \
In [13]:
!bentoml deployment describe titanic-prediction
To send request to your AWS Lambda deployment, grab the endpoint URL from the json output above:
In [18]:
!curl -i \
--header "Content-Type: application/json" \
--request POST \
--data '[{"Pclass": 1, "Age": 30, "Fare": 200, "SibSp": 1, "Parch": 0}]' \
https://u7nerrir6a.execute-api.us-west-2.amazonaws.com/Prod/predict
In [15]:
!bentoml deployment delete titanic-prediction
In [ ]: