1 unstable release

0.13.0 Nov 15, 2023

#12 in #llm-chain

23 downloads per month

MIT license

220KB
4K SLoC

llm-chain-sagemaker-endpoint

Amazon SageMaker Endppoint driver. Allows you to invoke a model hosted on Amazon SageMaker Endpoint, this includes Amazon SageMaker Jumpstart models.

Getting Started

  1. This crate uses the AWS SDK for Rust to communicate with Amazon SageMaker. You need to set up the credentials and a region following this guide
  2. Follow the SageMaker JumpStart documentation to find an LLM, then deploy it.
  3. Note down the SageMaker Endpoint name created by SageMaker JumpStart.
  4. Some models is included in this crate, see model::Model::<model_name>. Select one in your executor options's Model field. See examples/simple.rs for example.
  5. For custom models or models not included in model::Model, use model::Model::Other(<model_name>), where model_name is the SageMaker endpoint name.

Dependencies

~24–38MB
~635K SLoC