Amazon SageMaker introduces one-click model inference and fine-tuning for Hugging Face models in Amazon SageMaker JumpStart
Building on the Hugging Face deep learning containers released earlier this year, Amazon SageMaker is now making it even easier to deploy and fine-tune the state-of-the-art natural language processing models (NLP) with just a few clicks using Amazon SageMaker JumpStart. Amazon SageMaker JumpStart helps you quickly and easily get started with machine learning (ML). SageMaker JumpStart provides a set of solutions for the most common use cases that can be deployed readily with just a few clicks, and supports one-click deployment and fine-tuning of popular open source models such as natural language processing, object detection, and image classification models. These solutions are fully customizable and showcase the use of AWS CloudFormation templates and reference architectures so you can accelerate your ML journey. SageMaker JumpStart is also integrated in Amazon SageMaker Studio, our fully integrated development environment (IDE) for ML, making it intuitive to discover models, solutions, and more.
Building on the Hugging Face deep learning containers released earlier this year, Amazon SageMaker is now making it even easier to deploy and fine-tune the state-of-the-art natural language processing models (NLP) with just a few clicks using Amazon SageMaker JumpStart. Amazon SageMaker JumpStart helps you quickly and easily get started with machine learning (ML). SageMaker JumpStart provides a set of solutions for the most common use cases that can be deployed readily with just a few clicks, and supports one-click deployment and fine-tuning of popular open source models such as natural language processing, object detection, and image classification models. These solutions are fully customizable and showcase the use of AWS CloudFormation templates and reference architectures so you can accelerate your ML journey. SageMaker JumpStart is also integrated in Amazon SageMaker Studio, our fully integrated development environment (IDE) for ML, making it intuitive to discover models, solutions, and more.
To make it easier to get started with NLP use cases, Amazon SageMaker JumpStart now supports one-click inference and one-click fine-tuning for the most popular NLP models from Hugging Face. Starting today, you can access more than 40 pre-trained Hugging Face Transformers models in SageMaker JumpStart for sentence pair classification, text generation, text summarization, named entity recognition, and machine translation. These models can be deployed as-is for running inference. Sentence pair classification models can also be fine-tuned on new datasets, and then deployed for inference. SageMaker JumpStart also supports popular model hubs such as PyTorch Hub and TensorFlow Hub while hosting more than 300 computer vision and natural language processing models, along with 16 pre-built end-to-end solutions, and 19 example notebooks.
To find these new Amazon SageMaker JumpStart integrations, you can open SageMaker Studio and navigate to the SageMaker JumpStart tab. Search for Hugging Face in the search bar at the top of the page, and you will be given the choice of all the Hugging Face models. Alternatively, you can scroll down and find the Hugging Face models in the ‘Text’ carousel to find hundreds of models that can be deployed and fine-tuned with one-click.
Amazon SageMaker JumpStart is available in all regions where Amazon SageMaker Studio is available. To get started with these Hugging Face models on SageMaker JumpStart, refer to the documentation.