Computer Vision model

Question

I am trying to implement a computer vision model that I made on Google collabs over into Sagemaker however, the issue that I run into is some type of memory issue possibly caused by the way i upload my training data straight in jyupter lab causing the code to thing it is examining an image with500 channels rather than 500px. The code runs fine on google collabs.

AKA · Answer

Hi Viraj,
From what I can understand you are trying to bring a model that you trained in Colab into SageMaker and run inference. If that is the case, you can follow one of the approaches below:
Bring your own model and use a pre built container as explained here: https://docs.aws.amazon.com/sagemaker/latest/dg/docker-containers-prebuilt.html

Or create your own inference container and bring it to SageMaker as explained here: https://docs.aws.amazon.com/sagemaker/latest/dg/docker-containers-adapt-your-own.html

In both cases you will be hosting the inference in SageMaker as a managed endpoint

Computer Vision model

Relevant content