Amazon SageMaker JumpStart

Machine learning (ML) hub with built-in algorithms, foundation models, and prebuilt ML solutions that you can deploy with just a few clicks

Get started with Amazon SageMaker JumpStart

Hundreds of built-in algorithms with pretrained models from popular model hubs

Popular foundation models that can be deployed with just a few clicks

Fully customizable solutions for common use cases with reference architectures to accelerate your ML journey

Share ML models and notebooks across your organization to accelerate ML model building and deployment

Amazon SageMaker JumpStart is a machine learning (ML) hub that can help you accelerate your ML journey. With SageMaker JumpStart, you can access built-in algorithms with pretrained models from model hubs, pretrained foundation models to help you perform tasks such as article summarization and image generation, and prebuilt solutions to solve common use cases. In addition, you can share ML artifacts, including ML models and notebooks, within your organization to accelerate ML model building and deployment.

How it works

Built-in algorithms with pretrained models
Solutions
ML artifact sharing

Built-in algorithms with pretrained models
Solutions
ML artifact sharing

Built-in algorithms

SageMaker JumpStart provides hundreds of built-in algorithms with pretrained models from model hubs, including TensorFlow Hub, PyTorch Hub, HuggingFace, and MxNet GluonCV. You can also access built-in algorithms using the SageMaker Python SDK. Built-in algorithms cover common ML tasks, such as data classifications (image, text, tabular) and sentiment analysis.

Learn more about built-in algorithms »

Type	ML Tasks	Example Algorithms and Models
Vision	Image Classification Image Embedding Object Detection Sematic Segmentation	200+ models including ResNet, Inception, MobileNet, SSD, Faster RCNN, YOLO, Stable Diffusion
Text (natural language processing)	Sentence Segmentation Text Classification Embedding Pair Classification Question Answering Summarization Text Generation Translation Named Entity Recognition	100+ models including BERT, RoBERTa, DistilBERT, Distillbart xsum, GPT2, ELECTRA, Blazing Text, Sequence-to-sequence, Latent Dirichlet Allocation (LDA), Neural Topic Model (NTM), Bloom
Tabular	Classification Regression	LightGBM, CatBoost, XGBoost, Linear Learner, AutoGluon, TabTransformer, DeepAR, Factorization Machines, K-nearest Neighbors, Object2Vec, K-means, Random Cut Forest, IP Insights
Audio	Audio Embedding	TRILL, TRILL Distilled, FRILL

Foundation models

Foundation models are large-scale ML models that contain billions of parameters and are pretrained on terabytes of text and image data, so you can perform a wide range of tasks such as article summarization and text, image, or video generation. Because foundation models are pretrained, they can help lower training and infrastructure costs and enable customization for your use case.

Foundation models available through SageMaker.

Prebuilt solutions

Prebuilt solutions can be used for common use cases and are fully customizable.

Use case	Solution	Get Started
Credit rating prediction	Corporate credit rating prediction using multimodal ML for quality credit predictions Graph-based credit scoring Explain credit decisions	Documentation »
Predictive maintenance	Predictive maintenance for vehicle fleets Predictive maintenance for manufacturing	Documentation »
Computer vision	Product defect detection in images Handwriting recognition Object detection for bird species	Documentation »
Reinforced learning	Autonomous driving with visual perception and active learning Distributed reinforcement learning for Procgen challenge Reinforcement learning for Battlesnake AI competitions	Documentation »
Fraud detection	Detect malicious users and transactions Fraud detection in financial transactions using deep graph library Financial payment classification	Documentation »
Extract and analyze data from documents	Differential privacy for sentiment classification Document summarization, entity, and relationship extraction Handwriting recognition using Amazon SageMaker Filling in missing values in tabular records	Documentation »
Churn prediction	Churn prediction for mobile Churn prediction with text	Documentation »
Demand forecasting	Demand forecasting with deep learning	Documentation »
Personalized recommendations	Entity resolution in identity graphs with deep graph library Purchase modeling	Documentation »
Price optimization	Price optimization using Double Machine Learning (ML) and Prophet forecasting	Documentation »
Healthcare and life sciences analytics	Lung cancer survival prediction	Documentation »

Customers

Tyson
“At Tyson Foods, we continue to look for new ways to use machine learning (ML) in our production process to improve productivity. We use image classification models to identify products from the production line that require package labels. However, the image classification models need to be retrained with new images from the field on a recurring basis. Amazon SageMaker JumpStart enables our data scientists to share ML models with support engineers so they can train ML models with new data without writing any code. This accelerates the time-to-market of ML solutions, promotes continuous improvements, and increases productivity.”

Rahul Damineni, Specialist Data Scientist, Tyson Foods
Mission Automate
“Thanks to Amazon SageMaker JumpStart, we are able to launch ML solutions within days to fulfill machine learning prediction needs faster and more reliably.”

Alex Panait, CEO – Mission Automate
MyCase
“Thanks to Amazon SageMaker JumpStart, we can have better starting points which makes it so that we can deploy a ML solution for our own use cases in 4-6 weeks instead of 3-4 months.”

Gus Nguyen, Software Engineer – MyCase
pivotree
“With Amazon SageMaker JumpStart, we can build ML applications such as automatic anomaly detection or object classification faster and launch solutions from proof of concept to production within days.”

Milos Hanzel, Platform Architect – Pivotree