/covalent

Pythonic tool for orchestrating machine-learning/high performance/quantum-computing workflows in heterogeneous compute environments.

Primary LanguagePythonApache License 2.0Apache-2.0

hero

version Static Badge Static Badge Static Badge Static Badge apache

Run AI, ML, and Scientific Research Code on Any Cloud or On-Prem Cluster with a Single Line

divider    divider    divider    divider 

pip install covalent --upgrade

Check our Quick Start Guide for setup instructions or dive into your First Experiment. Learn more on the Concepts.

What is Covalent?

Covalent is a Python library for AI/ML engineers, developers, and researchers. It provides a straightforward approach to running compute jobs, like LLMs, generative AI, and scientific research, on various cloud platforms or on-prem clusters.

Run Code Anywhere: Execute Python functions in any cloud or on-prem cluster by changing just a single line of code.

It is as simple as swapping the decorator with our executor plugins. Choose from existing plugins or create custom ones for tailored interactions with any infrastructure.

Abstraction of Infrastructure Management: Abstract the complexities of cloud consoles, terraform, or IaC in the background.
Serverless Infrastructure: Automatically converts any infrastructure, including on-prem SLURM clusters or cloud compute, into a serverless setup.

If you find Covalent useful or interesting, feel free to give us a ⭐ on GitHub! Your support helps us to continue developing and improving this framework.


For AI/ML Practitioners and Developers For Researchers
  • Robust Compute Backend: Ideal as a backend compute framework for AI/ML applications, Large Language Models (LLMs), Generative AI, and more.
  • Cloud-Agnostic Execution: Execute high-compute tasks seamlessly across different cloud environments.
  • Infrastructure Abstraction: Directly use computing resources while keeping your business code independent from the infrastructure/resource definitions.
  • Local-Like Access: Effortlessly connect to compute resources from your laptop, eliminating the need for SSH or complex scripts.
  • Unified Interface Across Environments: Consistent experience with on-prem HPC clusters and cloud platforms like SLURM, PBS, LSF, AWS, GCP, Azure.
  • Real-Time Monitoring Monitoring: User-friendly UI for real-time monitoring, enabling cost-effective and iterative R&D.

Out-of-box observability - Try out the demo

If you find Covalent useful or interesting, feel free to give us a ⭐ on GitHub! Your support helps us to continue developing and improving this framework.

video

Explore Covalent Through Examples

Jump right into practical examples to see Covalent in action. These tutorials cover a range of applications, giving you a hands-on experience:

AI Tutorial MNIST Tutorial Quantum Tutorial

Explore Our Extensive Plugin Ecosystem

Covalent integrates seamlessly with a variety of platforms. Discover our range of plugins to enhance your Covalent experience:


divider divider divider divider
divider divider divider divider

Key Features at a Glance

Get a quick overview of what Covalent offers. Our infographic summarizes the main features, providing you with a snapshot of our capabilities:


development

Know More About Covalent

For a more in-depth description of Covalent's features and how they work, see the Concepts page in the documentation.


divider divider divider divider

Installation

Covalent is developed using Python on Linux and macOS. The easiest way to install Covalent is by using the PyPI package manager.

pip install covalent --upgrade

For other methods of installation, please check the docs.

Deployments

Covalent offers flexible deployment options, from Docker image/AMIs for self-hosting to pip package for local installations, accommodating various use cases

divider divider divider


Contributing

To contribute to Covalent, refer to the Contribution Guidelines. We use GitHub's issue tracking to manage known issues, bugs, and pull requests. Get started by forking the develop branch and submitting a pull request with your contributions. Improvements to the documentation, including tutorials and how-to guides, are also welcome from the community. For more information on adding tutorials, check the Tutorial Guidelines. Participation in the Covalent community is governed by the Code of Conduct.

Citation

Please use the following citation in any publications.

https://doi.org/10.5281/zenodo.5903364

License

Covalent is licensed under the Apache 2.0 License. See the LICENSE file or contact the support team for more details.

For a detailed history of changes and new features, see the Changelog.