Examples and Tutorials
The Graphcore GitHub examples repository contains a catalogue of application examples that have been optimised to run on Graphcore IPUs for both training and inference. The available code covers a wide range of popular models including NLP, Computer Vision, Speech, Multimodal, GNNs, AI for Simulation, and Recommender Systems. This includes a selection of models that achieve state of the art performance on IPUs, as well as code examples for self-learning. The README file in the repo lists details of all the models and which ML framework models are applicable to.
You can also access these examples from the Graphcore Model Garden, where you can filter by model type and framework.
The examples repository also contains some simple example programs and tutorials. These cover PyTorch, TensorFlow 2, TensorFlow 1, the Poplar graph programming framework, and the PopVision graph and system analyser tools.
The repository includes:
Tutorials to help you get started using the Poplar SDK and Graphcore tools to run code on the IPU.
You can also access these tutorials from the Tutorials document.
Feature examples: small code examples showing you how to use various software features when developing for IPUs. The feature examples README contains details of the code available.
Simple application examples: basic applications written in different frameworks targeting the IPU. The simple applications README contains details of the code available.
Kernel benchmarks: code for benchmarking the performance of some selected types of neural network layers on the IPU, using TensorFlow or the Graphcore PopART framework.
The repository also contains code used in technical notes, videos and blogs.
If you are using a version of the Poplar SDK prior to version 3.2, you will need to use the old GitHub tutorials repo. This has branches corresponding to each version of the SDK; for example for SDK 3.1, checkout branch
For general help, discussions and announcements, please join our Graphcore Slack Community.