PyRun Documentation

Appearance

Release Notes ​

Track the latest updates, features, and improvements to the PyRun platform.

Version 1.2.0 - Democratizing Data & AI ​

Release Date: [Date - September 17, 2025]

This landmark release introduces dedicated AI and Data Cockpit sections to the PyRun platform, embodying our core mission to make sophisticated cloud computing accessible and simple for everyone. Now, developers and researchers can discover and integrate public datasets and AI models into their workflows in just a few clicks.

✨ Major New Features ​

  • New "AI" Section: A dedicated hub for all AI-related activities in PyRun.

    • AI Code Assistant: An intelligent assistant integrated directly into the workspace editor, providing code suggestions, completions, and boilerplate for common AI/ML tasks.
    • HuggingFace Model Explorer: Browse, search, and select models from the HuggingFace Hub without leaving PyRun. A "Use in PyRun" button provides starter code to load the selected model in your workspace, streamlining the path from discovery to execution.
    • Enhanced AI Pipeline Support (MCP): Expanded support for Managed Cloud Pipelines (MCP) for AI, simplifying the setup and execution of complex, multi-stage workflows like fine-tuning and inference.

  • New "Data Cockpit" Section: The Data Cockpit has been expanded into a full-featured section for data discovery and preparation.

    • Public Dataset Explorer: Explore and use well-known public datasets directly within PyRun. Initial offerings include genomics data (e.g., COVID-19 sequences), satellite imagery, and more.
    • Integrated Data Explorer: This explorer allows you to browse datasets and seamlessly integrate them into your workspace for use in pipelines.
    • The existing powerful data partitioning features remain, now integrated into this new, more discoverable interface.

🚀 Improvements & Enhancements ​

  • The user experience has been updated to include the new top-level AI and Data Cockpit sections.
  • Documentation has been added for the new AI Code Assistant and HuggingFace Model Explorer features.
  • The core platform has been optimized to support the new data and model discovery workflows.
  • Core platform stability and performance optimizations.

💬 Feedback ​

  • We'd love to hear your feedback on the new AI and Data Cockpit features! Please report any issues or suggest improvements via GitHub Bug Report.

Version 1.1.0 - Multi-Cloud, New Frameworks, and AI Enhancements ​

Release Date: [Date - June 6, 2025]

This is a major update expanding PyRun's capabilities across new cloud providers, frameworks, and user experience features, significantly strengthening our AI/ML support.

✨ Major New Features ​

  • IBM Cloud Support: PyRun now officially supports IBM Cloud as a compute provider. You can configure IBM Code Engine as a backend for both Lithops and Dask workloads. IBM Code Engine Support
  • New Framework Support (via Pipelines):
    • Ray: Added a pipeline for distributed hyperparameter tuning using Ray Tune, showcasing support for this popular AI/ML framework.
    • Cubed: Added a pipeline for large-scale n-dimensional array processing using Cubed.
  • Enhanced AI/ML Capabilities: Solidified PyRun's position as an AI-ready platform with pre-built pipelines for Image Classification and Audio Recognition using TensorFlow.
  • User Interface (UI) Improvements:
    • Workspace Search: A new search bar in the "Workspaces" section allows you to quickly find your projects.
    • Improved GitHub Integration: Streamlined the process of connecting and syncing with GitHub repositories.

🚀 Improvements & Enhancements ​

  • Documentation has been comprehensively updated to reflect all new features, including guides for new frameworks and cloud providers.
  • The pipeline library has been expanded to include Ray, Cubed, and new AI examples. (Pipelines)
  • Core platform stability and performance optimizations.

⚠️ Known Issues ​

  • AWS Account Limitation: (Same as v1.0.1) Currently, each AWS account can only be linked to a single PyRun user due to CloudFormation stack naming conventions.

💬 Feedback ​

  • We'd love to hear your feedback! Please report any issues or suggest improvements via GitHub Bug Report.

Version 1.0.1 - Introducing AI Capabilities ​

Release Date: May 13, 2025

This update introduces foundational support for AI and Machine Learning workloads on PyRun, providing new pipelines and paving the way for more advanced AI features.

✨ Major New Features ​

  • AI Platform Support: PyRun now offers enhanced support for AI/ML workflows, including easier environment management for popular AI libraries.
  • New AI Pipelines:
    • Audio Recognition: An AI pipeline for audio keyword recognition using TensorFlow.
    • DaskML for Distributed Learning: Demonstrates distributed machine learning using Dask-ML.
    • Image Classification: An AI pipeline for image classification tasks using TensorFlow.
    • LLM Execution with Ollama: An interactive notebook environment to run and experiment with Large Language Models (LLMs) locally within your PyRun workspace using Ollama.
    • (Explore these in Pipelines and learn more about AI Features)

🚀 Improvements & Enhancements ​

  • Documentation updated to reflect new AI capabilities and future roadmap for AI/ML.
  • Ongoing work to expand the pipeline library with new examples, including initial explorations into Ray support and other advanced use cases.

⚠️ Known Issues ​

  • AWS Account Limitation: (Same as v1.0.0) Currently, each AWS account can only be linked to a single PyRun user due to CloudFormation stack naming conventions. We are exploring solutions for multi-user support within a single AWS account.

💬 Feedback ​

  • We'd love to hear your feedback! Please report any issues or suggest improvements via GitHub Bug Report.

Version 1.0.0 - Initial Public Release ​

Release Date: [Date of Release - April 01, 2025]

Welcome to the first public release of PyRun! 🎉 We're thrilled to introduce a platform designed to simplify and democratize cloud computing for Python developers, data scientists, and researchers. This initial version lays the foundation for effortless execution, management, and monitoring of your Python workloads in the cloud.

✨ Major New Features ​

  • Core PyRun Platform: Execute Python code seamlessly in the cloud without managing complex infrastructure. (What is PyRun?)
  • Integrated Web Interface: A unified, VS Code-like interface for managing workspaces, code, configurations, and monitoring.
  • Workspace Management: Create isolated environments for your projects.
  • Lithops Integration: First-class support for running massively parallel tasks using Lithops, with easy configuration via the dashboard. (Working with Lithops)
  • Dask Integration: Seamlessly scale your data science and machine learning workloads using Dask, with flexible cluster configuration within your code. (Working with Dask)
  • Automatic Runtime Management: Easily customize your Python environment using .pyrun/environment.yml (Conda) or Dockerfile. PyRun automatically detects changes and rebuilds your runtime. (Runtime Management)
  • Real-Time Monitoring: Comprehensive dashboard to track job progress, resource utilization (CPU, memory, network), and view detailed logs. (Real-Time Monitoring)
  • Templates: Jumpstart projects with pre-configured templates for Lithops and Dask. (Templates)
  • Pipelines: Explore ready-to-run, real-world examples demonstrating PyRun's capabilities with Lithops and Dask (FLOPS benchmark, METASPACE annotation, Mandelbrot, Hyperparameter Tuning). (Pipelines)
  • AWS Integration: Simple AWS account linking using CloudFormation for resource provisioning. (Account Setup)
  • Data Cockpit: Foundation for managing data sources within PyRun. (Data Cockpit)

🐛 Bug Fixes ​

  • Initial release, no previous bugs to report.

⚠️ Known Issues ​

  • AWS Account Limitation: Currently, each AWS account can only be linked to a single PyRun user due to CloudFormation stack naming conventions. We are exploring solutions for multi-user support within a single AWS account.

⬆️ Getting Started ​

💬 Feedback ​

We'd love to hear your feedback! Please report any issues or suggest improvements via GitHub Bug Report.