Developers - Cerebras

Develop for the Cerebras System

With the Cerebras Software Platform, CSoft, you’ll spend more time pushing the frontiers of AI instead of optimizing distributed implementations.

Easily continuously pre-train massive GPT-family models with up to an astonishing 20 billion parameters on a single device, then scale to Cerebras Clusters with just a parameter change.

Get up and running quickly without changing your workflows. Utilize existing frameworks for your work or dive deep developing custom kernels – the Cerebras Software Platform is designed for flexibility and optimization.

Learn about Cerebras-GPT open source models

See What's New in CSoft R1.8

Cerebras AI Model Studio Walkthrough

High-Resolution UNet Walkthrough

Read SDK Technical Overview

PyTorch

Our PyTorch interface library is a simple wrapper for PyTorch program exposed through API calls that is easy to add as few extra lines of code for an existing Pytorch implementation.

TensorFlow

Integration of TensorFlow is via Cerebras Estimator, which is a wrapper class we developed based on standard TensorFlow Estimator and standard TensorFlow semantics.

SDK

The Cerebras SDK allows researchers to extend the platform and develop custom kernels – empowering them to push the limits of AI and HPC innovation.

Cerebras Model Zoo

This repository contains examples of common deep learning models demonstrating best practices for coding for the Cerebras hardware.

Developer Resources

Documentation

Videos

Developer Blog

Partner Hardware Request

Whitepapers

Publications

Community

Contact Us!

Developer blogs

Efficient Large-Scale GPT Training Using a Cerebras Wafer-Scale Cluster

Cerebras has built a platform for push-button training of large language models that can accelerate time to insights without having to orchestrate across a…

Machine Learning,Software,Cloud,Blog,Developer Blog,Large Language Model,NLP,Deep Learning

Cerebras Architecture Deep Dive: First Look Inside the HW/SW Co-Design for Deep Learning [Updated]

Our ML-optimized architecture enables the largest models to run on a single device. With data parallel-only scale out and native unstructured sparsity…

Machine Learning,Software,Blog,Developer Blog

Fine-Tuning with Cerebras AI Model Studio Launchpad

Cerebras shares research showing smaller foundation models that are fine-tuned on domain-specific tasks outperform larger foundation models.

Machine Learning,Software,Cloud,Blog,Developer Blog,Foundation Model,Large Language Model,NLP

Cerebras-GPT: A Family of Open, Compute-efficient, Large Language Models

Cerebras open sources seven GPT-3 models from 111 million to 13 billion parameters. Trained using the Chinchilla formula, these models set new benchmarks for…

Machine Learning,Software,Cluster,Cloud,Blog,Developer Blog

FAQ

Don’t see your question?
Send us an email at developer@cerebras.net

What is the programming model for a CS system?

To learn more about our programming model, please see our documentation – https://docs.cerebras.net/en/latest/wsc/cerebras-basics/how-cerebras-works.html#

How can I program using an ML framework on a CS system?

To get started with PyTorch please use this documentation
To get started with TensorFlow please use this documentation

How do I learn more about the SDK?

Read the SDK technical overview white paper
Please request access to our SDK here

Where can I get an example of a model’s reference implementation that can be run on a CS System?

Please find our example reference model implementation here: https://github.com/Cerebras/cerebras_reference_implementations. To get access to our full list, please contact us at developer@cerebras.net

How do I get access to upcoming Cerebras events?

Please sign up for our newsletter!

How can I access Cerebras’ supported models?

Contact us at developer@cerebras.net.