DVA-C02 · Question #633
A company uses an AWS Lambda function to perform natural language processing (NLP) tasks. The company has attached a Lambda layer to the function. The Lambda layer contain scientific libraries that…
The correct answer is D. Create a Docker container that includes the scientific libraries and the pre-trained model. Update. Lambda layers have a maximum unzipped size limit of 250 MB, and total uncompressed size across all layers must not exceed this limit. The reported 8.7 GB far exceeds this, causing the RequestEntityTooLargeException. Using a Docker container image, which supports up to 10 GB…
Question
A company uses an AWS Lambda function to perform natural language processing (NLP) tasks. The company has attached a Lambda layer to the function. The Lambda layer contain scientific libraries that the function uses during processing. The company added a large, pre-trained text-classification model to the Lambda layer. The addition increased the size of the Lambda layer to 8.7 GB. After the addition and a recent deployment, the Lambda function returned a RequestEntityTooLargeException error. The company needs to update the Lambda function with a high-performing and portable solution to decrease the initialization time for the function. Which solution will meet these requirements?
Options
- AStore the large pre-trained model in an Amazon S3 bucket. Use the AWS SDK to access the
- BCreate an Amazon EFS file system to store the large pre-trained model. Mount the file system to
- CSplit the components of the Lambda layer into five new Lambda layers. Zip the new layers, and
- DCreate a Docker container that includes the scientific libraries and the pre-trained model. Update
How the community answered
(59 responses)- A27% (16)
- B10% (6)
- C7% (4)
- D56% (33)
Explanation
Lambda layers have a maximum unzipped size limit of 250 MB, and total uncompressed size across all layers must not exceed this limit. The reported 8.7 GB far exceeds this, causing the RequestEntityTooLargeException. Using a Docker container image, which supports up to 10 GB, allows bundling the pre-trained model and libraries together, offering a high-performing and portable solution with faster initialization and easier deployment.
Community Discussion
No community discussion yet for this question.