SageMaker deploy custom script

Question

I'm pretty new to SageMaker, so I'm sorry if I miss something obvious.

I've trained a DL model which uses frames from a video to make a prediction. The current script, that runs in the SageMaker jupyter-notebook, takes a video URL as an input and uses an FFMPEG subprocess pipe to extract the frames and predict them afterwards. This works fine, but now I want to start that script from Lambda.

As far as I understood, I could deploy my model with sagemaker and make predictions for every single frame from Lambda, unfortunately this is not an option, as ffprobe, ffmpeg and numpy are too large to fit into the limited lambda space.

tl;dr: Is it possible to run my custom script (ffmpeg frame extraction + tensorflow model prediction) as an endpoint in SageMaker?

Tuan Vo · Accepted Answer · 2020-02-05 00:36:40Z

1

Sagemaker allows you to use a custom Docker image (AWS document)

Build your own custom container image: If there is no pre-built Amazon SageMaker container image that you can use or modify for an advanced scenario, you can package your own script or algorithm to use with Amazon SageMaker.You can use any programming language or framework to develop your container

Create a docker image with your code (FFmpeg, TensorFlow)
Testing the docker container locally
Deploying the image on Amazon ECR (Elastic Container Repository)
Create a SageMaker model and point to the image

For details, you can learn more from this tutorial

answered Feb 5, 2020 at 0:36

Tuan Vo

2,09513 silver badges10 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

SageMaker deploy custom script

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related