Skip to content

Build and deploy a Transformer based NLP model using the open source Ludwig AutoML deep learning framework with TensorFlow and Amazon SageMaker.

License

Notifications You must be signed in to change notification settings

aws-samples/amazon-sagemaker-ludwig-transformer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Amazon SageMaker Ludwig Transformer

This repository provides a notebook for training and deploying a NLP transformer deep learning model using the open source Ludwig AutoML framework and HuggingFace library built on TensorFlow.

Dataset

We will train a multi-class classification model to classify Amazon product reviews with a star rating in the range of 1-5. We will be using a HuggingFace pre-trained yelp sentiment model as part of our training and fine-tuning.

Prerequisites

This notebook will Build Your Own Docker container. So you will require some additional permissions for your notebook role to publish to Amazon Elastic Container Registry ECR.

The easiest way to add these permissions is to add the managed policy to the role that you used to start your notebook instance in the SageMaker dashboard. To this end, open your notebook instance, go to the Permissions and encryption section to to navigate to the IAM role.

Notebook Role

After you choose the role, you can attach the AmazonEC2ContainerRegistryFullAccess policy. There’s no need to restart your notebook instance when you do this, the new permissions will be available immediately.

Notebook Role

Notebooks

Launch the train-ludwig.ipynb notebook to begin.

This notebook will step you through the process of creating three containers.

  1. ludwig-training - The base container build on TensorFlow 2.3 which clones and installs the latest Ludwig source and Amazon SageMaker Training Toolkit
  2. ludwig-inference - The container for hosting Ludwig models extends ludwig-training with a model server leveraging the Amazon SageMaker Inference Toolkit
  3. ludwig-reviews - The reviews training container that extends ludwig-training to add model configuration.

You will be shown how to train and evaluate the reviews model before deploying in local mode and launching test-ludwig.ipynb to evaluate CSV and JSON test payloads before deploying to an Amazon SageMaker Endpoint.

License

This library is licensed under the MIT-0 License. See the LICENSE file.

About

Build and deploy a Transformer based NLP model using the open source Ludwig AutoML deep learning framework with TensorFlow and Amazon SageMaker.

Topics

Resources

License

Code of conduct

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published