%reload_ext autoreload
%autoreload 2
%matplotlib inline
import warnings
warnings.filterwarnings("ignore")
!pip install -q --upgrade xgboost==0.90 numpy==1.18.5 pandas==1.0.4 bentoml
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
!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
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']])
train[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch', 'Survived']].head()
Model Training¶
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)
y_test = model.predict(X_test)
test['pred'] = y_test
test[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch','pred']].iloc[10:].head(2)
Create BentoService for model serving¶
%%writefile xgboost_titanic_bento_service.py
import xgboost as xgb
import bentoml
from bentoml.artifact import XgboostModelArtifact
from bentoml.adapters import DataframeInput
@bentoml.env(auto_pip_dependencies=True)
@bentoml.artifacts([XgboostModelArtifact('model')])
class TitanicSurvivalPredictionXgBoost(bentoml.BentoService):
@bentoml.api(input=DataframeInput())
def predict(self, df):
data = xgb.DMatrix(data=df[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch']])
return self.artifacts.model.predict(data)
Create BentoService saved bundle¶
# 1) import the custom BentoService defined above
from xgboost_titanic_bento_service import TitanicSurvivalPredictionXgBoost
# 2) `pack` it with required artifacts
bento_service = TitanicSurvivalPredictionXgBoost()
bento_service.pack('model', model)
# 3) save your BentoSerivce
saved_path = bento_service.save()
Load BentoService saved bundle¶
import bentoml
loaded_svc = bentoml.load(saved_path)
result = loaded_svc.predict(test)
test['pred'] = result
test[['Pclass', 'Age', 'Fare', 'SibSp', 'Parch','pred']]
Working with BentoML CLI¶
bentoml get <BentoServiceName> is great for list all versions of the BentoService
!bentoml get TitanicSurvivalPredictionXgBoost
bentoml get <BentoService name>:<BentoService version> to access detailed information
!bentoml get TitanicSurvivalPredictionXgBoost:latest
Run predicition task from CLI:
!bentoml run TitanicSurvivalPredictionXgBoost:latest predict \
--input '[{"Pclass": 1, "Age": 30, "Fare": 200, "SibSp": 1, "Parch": 0}]'
Model Serving via REST API¶
Start local API model server with bentoml serve command:
!bentoml serve TitanicSurvivalPredictionXgBoost:latest
Copy following curl command to make a curl request to REST API server from another terminal window:
curl -i \
--header "Content-Type: application/json" \
--request POST \
--data '[{"Pclass": 1, "Age": 30, "Fare": 200, "SibSp": 1, "Parch": 0}]' \
localhost:5000/predict
Containerize REST API server with Docker¶
The BentoService saved bundle direcotry is structured as a docker build context, which can be used directly to build a docker image for the model server:
!cd {saved_path} && docker build --quiet -t titanic-survival-predict-server .
Next, you can docker push the image to your choice of registry for deployment, or run it locally for development and testing:
!docker run -p 5000:5000 titanic-survival-predict-server
Deploy BentoService as REST API server to the cloud¶
BentoML support deployment to multiply cloud provider services, such as AWS Lambda, AWS Sagemaker, Google Cloudrun and etc. You can find the full list and guide on the documentation site at https://docs.bentoml.org/en/latest/deployment/index.html
Here is an example of deploying this saved BentoService to AWS Lambda.
Before started:
- Install AWS CLI https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-install.html
- Configure AWS credentials with
aws configure - Install aws-sam-cli package:
pip install -U aws-sam-cli==0.31.1 - Ensure docker is available and running https://docs.docker.com/get-docker/
!bentoml lambda deploy xgb-titanic-survival-predict \
-b TitanicSurvivalPredictionXgBoost:{bento_service.version}
!bentoml lambda get xgb-titanic-survival-predict
To send request to your AWS Lambda deployment, grab the endpoint URL from the json output above:
!curl -i \
--header "Content-Type: application/json" \
--request POST \
--data '[{"Pclass": 1, "Age": 30, "Fare": 200, "SibSp": 1, "Parch": 0}]' \
$(bentoml lambda get xgb-titanic-survival-predict | jq -r ".state.infoJson.endpoints[0]")
Use bentoml lambda delete to remove AWS Lambda deployment:
!bentoml lambda delete xgb-titanic-survival-predict