home.social

#jupyterlab — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #jupyterlab, aggregated by home.social.

fetched live
  1. Ditch localhost AI!
    Setup JupyterLab & PyTorch on Ubuntu 24.04 Remote GPUs.

    SRE Blueprint:
    • Miniconda: Avoid Anaconda bloat; fix CommandNotFoundError via source ~/.bashrc.
    • PyTorch CUDA 12.1: Native wheel install without system CUDA toolkits.
    • Persistent Daemon: Run Jupyter in tmux with a hashed password.
    • Zero Open Ports: Never expose port 8888! Use SSH tunneling or VSCode Remote-SSH.

    servermo.com/howto/setup-jupyt

    #PyTorch #JupyterLab #AI #DevOps #GPU #ServerMO

  2. I've updated the documentation for Carto-Lab Docker [1]. Maintaining open source software in spatial data science requires documenting reliable research workflows. I find this even more important than package and feature updates.

    New sections cover specific setups for reproducible spatial data processing, such as connecting VS Code to remote Jupyter containers [2], managing long-running spatial jobs in the background [3], and configuring SSH agent support for automated Git authentication [4].

    #SpatialDataScience #OpenScience #Reproducibility #JupyterLab

    [1]: cartolab.fdz.ioer.info/
    [2]: cartolab.fdz.ioer.info/use-cas
    [3]: cartolab.fdz.ioer.info/use-cas
    [4]: cartolab.fdz.ioer.info/use-cas

  3. Как я собрала локальную MCP-платформу для мониторинга промышленных данных

    Что получится, если собрать PostgreSQL, Airflow, JupyterLab, MinIO, Superset, шесть MCP-сервисов и локальную модель Ollama в одном Docker Compose-стенде? Показываю полный путь синтетических промышленных данных: от витрин и проверок качества до Parquet-артефактов, MCP-инструментов и LLM-пояснений. Внутри — архитектура, воспроизводимый запуск, результаты проверок и открытый репозиторий.

    habr.com/ru/articles/1064802/

    #MCP #Model_Context_Protocol #Data_Engineering #Apache_Airflow #PostgreSQL #Docker_Compose #MinIO #Ollama #JupyterLab #ETL

  4. I use #RISE slideshows to present #Jupyter notebooks with #Python in computational statistics classes. Does anyone have a similar solution for #RStats? I've used Jupyter notebooks with R for asynchronous lessons but 1) R in Jupyter is kind of clunky and 2) I'm not sure how well RISE will work if I have to update things, since it's kind of incompatible with #jupyterlab

  5. RE: fosstodon.org/@krassowski/1166

    4.6 and 7.6 release candidates are now available for testing. One week to go before a final release. If you have a large deployment or maintain Jupyter extensions - this is the time to do a final test before the stable release goes out!

  6. JupyterLab 4.6 beta is now available for testing 🎉.
    It includes:
    - new layout customization options
    - navigation between recently edited cells
    - revamped keyboard shortcuts UI
    - editable breadcrumbs in file browser
    - debugger improvements

    Full changelog: jupyterlab.readthedocs.io/en/l

  7. Рейтинг языков программирования на GitHub: анализ 2024–2025 в JupyterLab и Anaconda

    Мы не стали спорить с TIOBE и RedMonk и собрали свой рейтинг языков программирования на основе GitHub. Данные за 2024–2025 показывают неожиданные вещи: JavaScript впереди, TypeScript резко растёт, а Rust и Go выигрывают по качеству проектов. Разбираем, что стоит за цифрами и где могут быть искажения.

    habr.com/ru/companies/hostkey/

    #Python #GitHub_API #JupyterLab #анализ_данных #визуализация_данных #pandas #рейтинг_языков_программирования #репозитории #Data_Engineering #hostkey

  8. JupyterLite ist nun Teil des Projekts Jupyter

    JupyterLite ist eine JupyterLab-Distribution, die vollständig im Webbrowser ausgeführt wird und ohne Backend auskommt. Die Kernel werden mithilfe von WebAssembly direkt im Browser gestartet, wodurch ein Anwendungsserver überflüssig wird. kantel.github.io/posts/2026021 #JupyterLite #Jupyter #JupyterLab #Python #PyScript #Pyodide #P5js #DataScience #StatischeSeiten

  9. Один Dockerfile, два make-таргета и семинар, где студенты наконец пишут код

    Я веду семинары по машинному обучению на ФКН ВШЭ. Чтобы понять, как работает градиентный спуск, нужно написать его руками. И вот тут начинается проблема. Классический формат семинара: я стою перед аудиторией, пишу код, объясняю. Студенты смотрят . Кто-то конспектирует, кто-то фотографирует экран. Но не делает . Мне хотелось, чтобы семинар был не демонстрацией, а совместным экспериментом . Чтобы я показывал идею, а студенты тут же пробовали: меняли параметры, ломали код, смотрели что получится. В том же ноутбуке, в реальном времени.

    habr.com/ru/articles/994086/

    #JupyterLab #Docker #collaborative_editing #преподавание #машинное_обучение #ВШЭ

  10. Ok, I'm weird, I know... but I can't quite get over my little bit of excitement after installing @postmarketos on an old phone (OnePlus6T).

    The thing now runs Linux, as well as Waydroid (no Google) so I can run Android Apps from F-droid.

    I managed to set up a virtual environment in #Python and install and run #jupyterlab

    All on an old phone! My BT keyboard also just works, so typing this post in Fedilab on said old phone is a breeze!
    Love #linux and the #opensource community!

  11. Schlangenfraß: Jupyter Updates

    Schon vor einigen Wochen trudelten Update-Meldungen zu JupyterLab, Jupyter Notebook und JupyterLite in meinen Feedreader. Auch wenn ich diese Python-Umgebungen in der letzten Zeit aufgrund anderer Interessen etwas vernachlässigt habe, möchte ich Euch die Updates natürlich nicht vorenthalten. kantel.github.io/posts/2025121 #Python #Jupyter #JupyterLab #JupyterLabDesktop #JupyterLite

  12. #30DayMapChallenge Day 17: A New Tool Tried #JupyterGIS, a #Python-based mapping tool in #JupyterLab, for mapping transboundary aquifers using open data from @[email protected] + #HydroSHEDS. Real-time map+code collaboration, live annotations, and reproducible workflows.

  13. Day 17: A New Tool

    🛠️ Tried , a -based mapping tool in , for mapping transboundary aquifers using open data from + basins on an raster backdrop. What’s great? Real-time map+code collaboration, live annotations, and reproducible workflows. A promising lightweight, open alternative to traditional !

  14. Die Wiederentdeckung von MyST Markdown: Mein neues Publikationstool?

    Als mir jüngst zwei Videos zu MyST Markdown von unser allwissenden Datenkrake in meinen Feedreader gespült wurden, fiel mir ein, daß mir das Tool vor etwa zwei Jahren schon einmal untergekommen war und ich es so interessant fand, daß ich eigentlich ein paar Tests damit durchführen wollte. kantel.github.io/posts/2025111 #MySTMarkdown #JupyterLab #JupyterLabDesktop #Markdown #JupyterLite #JupyterBook #VisualStudioCode #Sphinx #Quarto

  15. New exciting features are coming to 4.5.0. We just published the release candidate. The goal is to release the final during the JupyterCon (next week!).

    Expect major enhancements around:
    - performance and scroll stability
    - notebook editing
    - table of contents
    - debugger
    - terminal

    pip install --pre --upgrade jupyterlab==4.5.0rc0

    Full release notes:
    jupyterlab.readthedocs.io/en/l

    Thank you to Rosio Reyes for compiling the release notes and thank you to everyone who contributed!

  16. On Day 2 of European Galaxy Days, Yvan Le Bras presented Two-Step Approach for Galaxy Tool Development Using Jupyter (“Interactive JupyterLab Notebook”).

    @galaxyfreiburg
    #interactive #EGD2025 #jupyterlab #tool_development #galaxyproject #jupyter_notebook #eosc #fair #open_science

  17. At the #IOER_FDZ, we have been thinking about ways to improve the production of scientific knowledge and contribute solutions to the #reproducibility crisis.

    One initiative that we are now sharing with the wider scientific community is #Carto-Lab Docker, a computational environment for the #spatial sciences built on a versioned, reproducible Docker container stack that enables FAIR principles (findable, accessible, interoperable and reusable). @mcnesium and I started developing it seven years ago. Having moved to version v.1.0.0, the software now left its beta phase and became an official research infrastructure component of the IOER-FDZ. (🚀@ Marc!)

    It includes pre-configured, environments for #open-source #cartography in #Python and #R, and has been designed to support transparent geospatial data analysis and sound publication practices in science.

    Firstly, Carto-Lab Docker bundles existing open-source software together with #Jupyterlab for easy deployment and use, reducing the barriers to good scientific practice and efficiently working in teams. It also encompasses critical training materials, documentation, and best-practice code to facilitate collaborative learning and incremental progress in research.

    You can find the documentation at [1]. Users may also be interested in our NFDI4Biodiversity training materials [2], which were authored using Carto-Lab Docker and explain many of the underlying concepts. Developers and administrators can find our public container images in the Quay.io container registry [3]. Contributions and input are warmly welcomed on Github [4].

    @ioer

    [1]: cartolab.fdz.ioer.info/
    [2]: training.fdz.ioer.info/
    [3]: quay.io/repository/ioer-fdz/ca
    [4]: github.com/ioer-dresden/carto-