Method #1: Deploying a mannequin saved on Amazon S3
Deploying a ML mannequin as a Python pickle file in an Amazon S3 bucket and utilizing it by a Lambda API makes mannequin deployment easy, scalable, and cost-effective. We arrange AWS Lambda to load this mannequin from S3 when wanted, enabling fast predictions with out requiring a devoted server. When somebody calls the API linked to the Lambda operate, the mannequin is fetched, run, and returns predictions primarily based on the enter information. This serverless setup ensures excessive availability, scales routinely, and saves prices since you solely pay when the API is used.
Step 1. Create a zipper archive for the Lambda layer
A Lambda layer is a zipper archive that accommodates libraries, a customized runtime, and different dependencies. I’ll exhibit the creation of a Lambda layer utilizing two Python libraries, Pandas and Scikit-learn, which are usually utilized in ML fashions. Beneath is the code for making a Lambda layer zip archive, containing Pandas and Scikit-learn, utilizing Docker. Create a file, identify it createlayer.sh, and replica the code into it.
if [ "$1" != "" ] || [$# -gt 1]; then
echo "Creating layer suitable with python model $1"
docker run -v "$PWD":/var/job "lambci/lambda:build-python$1" /bin/sh -c "pip set up -r necessities.txt -t python/lib/python$1/site-packages/; exit"
zip -r sklearn_pandas_layer.zip python > /dev/null
rm -r python
echo "Accomplished creating layer!"
ls -lah sklearn_pandas_layer.zip
else
echo "Enter python model as argument - ./createlayer.sh 3.6"
Now, in the identical listing, create a file named necessities.txt to retailer the names and variations of the libraries within the layer. On this case, our necessities.txt file will listing the names and variations of the Pandas and Scikit-learn libraries we’re utilizing.
pandas==0.23.4
scikit-learn==0.20.3
Subsequent, within the terminal, navigate to the listing the place you may have positioned the createlayer.sh and necessities.txt information and run the command beneath to generate the Lambda layer zip file.
