Skip to content

Latest commit

 

History

History
43 lines (27 loc) · 3.03 KB

README.md

File metadata and controls

43 lines (27 loc) · 3.03 KB
Raw

This repository contains developer resources demonstrating how to deploy and use various Arcee models on Amazon SageMaker.

You're welcome to use them in your projects. Of course, we will always appreciate proper credit and a friendly message :)

For information or to discuss how Arcee can help your business, please contact [email protected].

Models

The following Arcee models are featured in the sample notebooks:

  1. Llama-Spark: A powerful 8B parameter model built on the foundation of Llama-3.1-8B, excelling in reasoning, creative writing, and coding tasks.

  2. Arcee-Scribe: A versatile 7B parameter chat model with strong reasoning abilities, particularly adept at creative writing tasks.

  3. Arcee-Nova: A large language model with performance approaching that of GPT-4 (as of May 2023), built on Qwen2-72B-Instruct.

  4. Arcee-Spark: A powerful 7B parameter model that outperforms many larger models, offering a 32K context size.

  5. Arcee-Lite: A compact yet powerful 1.5B parameter model, ideal for embedded systems and resource-constrained environments.

  6. Arcee-Agent: A 7B parameter model specifically designed for function calling and tool use, excelling at interpreting and executing function calls.

  7. Arcee-SuperNova: a 70B model outperforming Llama 70B-Instruct, as well Llama 405B, Claude Sonnet 3.5 and GPT-4o in many general benchmarks.

  8. Llama-3.1-SuperNova-Lite: an 8B distilled version of the larger Llama-3.1-405B-Instruct model.

  9. SuperNova-Medius: a 14B parameter language model developed by Arcee.ai, built on the Qwen2.5-14B-Instruct architecture.

Deployment options

You can deploy these models on Amazon SageMaker in different ways:

  1. Deploy an Arcee model hosted on theHugging Face hub
    • model_notebooks: sample notebooks
    • cloudformation/create_endpoint_from_hf_model*.yaml: sample AWS CloudFormation templates
  2. Deploy a model artifact you have stored in S3
    • cloudformation/create_endpoint_from_s3_model*.yaml: sample AWS CloudFormation templates
  3. Deploy a model package that you have built yourself (not an AWS Marketplace model package)
    • cloudformation/create_endpoint_from_model_package*.yaml: sample AWS CloudFormation templates
  4. Deploy an Arcee model listed on the AWS Marketplace
    • This is the recommended approach if you cannot access Hugging Face in your AWS account, or if you want to start from an official package built and tested by Arcee.
    • You first need to subscribe to the model
    • You can then deploy the model using the built-in option, or with the sample notebooks in model_package_notebooks