ImageClassifierAPI/code/lambdaDeploymentPackageProcess.txt at master · scooter4j/ImageClassifierAPI

17 lines (10 loc) · 856 Bytes

Building Lambda deployment .zip file

1. Create a folder in which to build a clean and minimal package, e.g. /Users/scottqueen/temp/AI_for_lambda

2. create a `classification_model` folder under the build folder and copy the classification model and label files into it

3. copy the Tensorflow and numpy libraries into the top-level build folder, eg /Users/scottqueen/temp/AI_for_lambda. To do this, execute this:

cp -r /anaconda3/envs/image_classification/lib/python3.6/site-packages/tensorflow .

cp -r /anaconda3/envs/image_classification/lib/python3.6/site-packages/numpy

4. copy the Lambda function code into the top-level build folder, eg /Users/scottqueen/temp/AI_for_lambda

cp /Users/scottqueen/AI/code/classify_image_lambda.py .

5. from the top-level build folder, create a zip file of the deployment artifacts

zip * lambda_image_classifer.zip

Provide feedback