Hugging Face Model Deployment on AWS

This repository contains tools and scripts for deploying Hugging Face models on AWS using SageMaker and Lambda functions.

Repository Contents

• Deploy_HuggingFace_Model.ipynb: Jupyter notebook for deploying Hugging Face transformer models to AWS SageMaker.
• Lambda4SagemakerEndpoint.ipynb: Jupyter notebook demonstrating how to create and configure Lambda functions to interact with SageMaker endpoints.
• lambda_function.py: Python script containing the AWS Lambda function code for invoking SageMaker endpoints.
• export_iam.py: Utility script for exporting and managing IAM roles and policies required for the deployment.

Prerequisites

• AWS Account with appropriate permissions
• Python 3.x
• AWS CLI configured
• Required Python packages:
  • boto3
  • sagemaker
  • transformers
  • jupyter

Setup Instructions

1. Clone this repository
2. Install required dependencies:

   pip install boto3 sagemaker transformers jupyter

3. Configure your AWS credentials:

   aws configure

Deployment Process

1. Prepare your Hugging Face model:

   • Open Deploy_HuggingFace_Model.ipynb to prepare and deploy your model to SageMaker

2. Set up Lambda function:

   • Use Lambda4SagemakerEndpoint.ipynb to understand how to create a Lambda function
   • Customize lambda_function.py for your specific model and use case

3. Configure IAM permissions:

   • Run export_iam.py to set up the necessary IAM roles and policies

IAM Requirements

The deployment requires specific IAM permissions:

• SageMaker execution role with S3 access
• Lambda execution role with permissions to invoke SageMaker endpoints
• Trust relationships between services

Usage

After deployment, you can invoke your model through:

• Direct SageMaker endpoint calls
• Lambda function invocations
• API Gateway (if configured)

Troubleshooting

Common issues:

• IAM permission errors: Ensure your roles have the correct policies and trust relationships
• S3 access issues: Verify your model artifacts are in the correct region and accessible
• Lambda timeout: Adjust timeout settings for larger models

License

[Specify your license here]

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

    Too many current requests. Your queue position is 1. Please wait for a while or switch to other models for a smoother experience.# deploy_higgingface_model