<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=145304570664993&amp;ev=PageView&amp;noscript=1">
Skip to content

BUILD NEXT GENERATION MACHINE INTELLIGENCE WITH POPLAR®

Learn how to get your ML models up and running fast on IPU systems with the Poplar SDK. Get access to our open source libraries, model garden, framework support, analysis tools and Docker Hub containers.

Get Started

DOCUMENTATION

Explore our software documentation, user guides and technical notes.

See more
Graphcore Software Documentation

TUTORIALS + CODE EXAMPLES

Hands-on code tutorials, simple application and feature examples.

See more
Graphcore tutorials and code examples

GRAPHCORE GITHUB

Access our open source libraries, APIs, applications and code examples.

See more
github-01

DOCKER HUB

Access a selection of Poplar SDK container images via Docker Hub.

See more
docker

MODEL GARDEN

Access a repository of deployable ML applications on the IPU.

See more
Models

HOW-TO VIDEOS

Watch practical how-to videos and demos by Graphcore engineers.

See more
Graphcore How-To Videos, Webinars & Demos

WEBINARS

Register for upcoming Graphcore webinars or watch on-demand.

See more
Webinars

RESEARCH PAPERS

Read publications from Graphcore's Research team and IPU innovators.

See more
Graphcore research papers

POPVISION™ TOOLS

Download PopVision to analyse IPU performance and utilisation.

See more
Analysis Tools

FRAMEWORKS

  • HuggingFace
  • PyTorch
  • Lightning
  • TensorFlow
  • Keras
  • ONNX
  • PaddlePaddle
  • HALO

Train, fine-tune and accelerate state-of-the-art transformer models on IPU systems with Hugging Face. Graphcore's IPU-optimized transformer models allows developers to train models faster with minimal changes to your code.

Learn more about our partnership with Hugging Face.

Introducing the PyTorch API for the IPU.

With PopTorch™ - a simple Python wrapper for PyTorch programs, developers can easily run models directly on Graphcore IPUs with a few lines of extra code.

Learn how to build performant PyTorch applications for training and inference with our latest user guide, tutorials, and code examples.

To find out more about our announcement, read our guest blog with PyTorch.

Introducing PyTorch Lightning for the IPU.

This initial release enables developers to use PyTorch Lightning's ultra-fast, lightweight AI research framework to accelerate their training models with the performance advantages of the IPU.

Users of PyTorch Lightning can now take any PopTorch models and run them directly on the IPU with minimal code changes.

To learn more, read our announcement blog.

Full support for TensorFlow 1 & 2 with integrated XLA backend.

The TensorFlow API for IPU includes distributed IPU-specific estimators and optimisers to enhance application performance.

Flexible model building capability with Keras Sequential & Functional API support for TensorFlow 2.

Functional and sequential support for Keras is provided via our TensorFlow 2 integration.

Developers can port many standard Keras models to the IPU, ready for optimization and performance tuning.

Developers can easily import and export models in ONNX format using Poplar Advanced Runtime (PopART™).

PopART uses Python and C++ APIs to allow models built in other ML frameworks to be accelerated on Graphcore’s IPU hardware.

Full support for Baidu's PaddlePaddle deep learning framework for large-scale model training, as well as high-performance inference tasks through the PaddlePaddle inference library.

To learn more, read our announcement blog.

Poplar supports Alibaba Cloud's Heterogeneity-Aware Lowering and Optimization platform (HALO).

HALO is part of Alibaba Cloud's Open Deep Learning API (ODLA), which is designed to rapidly enable heterogeneous computing and meet the growing demand for AI compute acceleration in datacentres.

Featured Resources

Explore our latest and most popular software documentation, how-to videos, tutorials and webinars.

Fundamentals of the IPU + Poplar

Quickstart demo for IPU beginners showing a simple addition example running on the IPU using the Poplar framework.

Watch the Video

NEW

Poplar Software White Paper

Cambrian AI Research examines the growing momentum of the Poplar software stack and ecosystem.

Read the White Paper

IPU Programmer's Guide

An introduction to the IPU architecture, programming model and tools available.

Read the Guide

PyTorch for IPU User Guide

User guide and API reference for running PyTorch models on the IPU with PopTorch.

Read the Guide

TensorFlow for IPU User Guide

User guides and API reference for the IPU implementation of TensorFlow and Keras.

Read the Guide

Programming on the IPU 101

Learn the basics with this introduction to the Poplar SDK, IPU programming techniques and how to port models.

Watch the Video

HPC Cookbook for the IPU

NEW

HPC Cookbook v5

Read the University of Bristol's community guide on low-level programming for HPC workloads using Poplar C++.

Start the Tutorial

NEW

Memory + Performance Optimisation Guide

General techniques to develop and optimise high-performance ML models running on the IPU.

Read the Guide

NEW

Tutorial: BERT Fine-tuning

How to fine-tune a pre-trained BERT-L model with PyTorch on a Graphcore IPU-POD16 system.

Start the Tutorial
More Resources

Developer News

 
 
 
Visit our Medium

Support and Community

Github

Github

Visit public repos
Graphcore Support

Graphcore Support

Get support
Stack Overflow

Stack Overflow

Ask a question
docker

Docker Hub

Access containers
×