Building the future of AI innovation with our software partners and ecosystem integrations.
Integration with today’s most popular machine learning frameworks, AI software tools and MLOps platforms to drive ease of use for IPU developers. Together, with our partners, we're building an ecosystem to drive open-source innovation and deliver the future in machine intelligence.
.svg "Channel"){kind=icon}
Aleph Alpha, the developer of the 200bn parameter Luminous model, is collaborating with Graphcore to develop and deploy advanced, large multi-modal models on the IPU.
Docker Hub is a leading repository registry in the cloud that allows developers to pull Graphcore’s Poplar SDK container images as part of the Docker Verified Publisher Program.
HALO (Heterogeneity-Aware Lowering and Optimization) is a deep learning framework developed by Alibaba Cloud for heterogeneous computing acceleration supported by Graphcore IPUs.
Hugging Face Optimum is an ML optimisation toolkit for specialised AI hardware that enables developers to train a variety of Transformer models out the box on Graphcore IPUs covering NLP, Computer Vision and Speech.
ONNX (Open Neural Network Exchange) is the open standard for machine learning interoperability. It enables IPU developers to use models with a variety of frameworks, tools, runtimes, and compilers.
PaddlePaddle is Baidu’s open source parallel distributed deep learning framework integrated with Graphcore IPUs for large-scale ML model training and high-performance inference tasks.
Try IPUs in the cloud with a zero set-up, pre-configured Jupyter development environment on Paperspace – a high-performance cloud computing and ML development platform for building, training and deploying machine learning models.
Pienso’s no code/low code interface equips non-technical users to derive insights from their text data - structured and unstructured. Running on Graphcore’s made-for-AI processors, those insights arrive 4x faster.
PopTorch™ is a set of extensions for PyTorch to enable PyTorch models to run directly on Graphcore IPUs. PopTorch has been designed to require as few changes as possible to your models in order to run on the IPU.
PyG (PyTorch Geometric) is a library built upon PyTorch to easily write and train Graph Neural Networks (GNNs) for a wide range of applications related to structured data. Currently available in preview on IPUs.
PyTorch Lightning is an ultra-fast, lightweight AI research framework to accelerate training models now integrated with Graphcore’s Poplar SDK to harness the performance advantages of the IPU.
The TensorFlow API for IPU enables developers to run TensorFlow models with distributed IPU-specific estimators and optimisers to enhance application performance.
VMware is the leading provider of enterprise virtualization software and tools for application modernization in the cloud. VMware’s Project Radium will support Graphcore IPUs as part of its hardware disaggregation initiative.
Weights & Biases is leading MLOps platform designed to help AI practitioners build better models faster. IPU profiling support in available within the Weights & Biases tool to log various metrics such as utilisation, clock speed, temperature and power usage.
Apply to join Graphcore's Partner Program using the link below.
For other inquiries contact us at partners@graphcore.ai.