grids¶

Introduction¶

Grids are based on a 12 column system, which can adapt to the size of the viewing screen.

A grid directive can be set:

with the number of default columns (1 to 12);
either a single number for all screen sizes,
or four numbers for extra-small (<576px), small (768px), medium (992px) and large screens (>1200px),

then child grid-item directives should be set for each item.

Placing a card in a grid¶

The grid-item-card directive is a short-hand for placing a card content container inside a grid item (see Cards). Most of the card directive’s options can be used also here:

Title 1

Title 2

Show some sphinx-extensions¶

https://sphinx-extensions.readthedocs.io/en/latest/

sphinx-copybutton

Adds a little « Copy » button to the top-right of your code blocks.

sphinx-favicon

Adds custom favicons to your Sphinx documentation.

myst-parser

Powerful Markdown flavor for your Sphinx documentation without loosing the power of Sphinx itself.

sphinx-autoapi

Auto documents your API code without executing the code itself (as sphinx.autodoc does).

sphinx-notfound-page

Renders nice 404 pages respecting all the look & feel of your documentation.

sphinx-version-warning

Adds a warning banner at the top if the reader is reading an old version of your documentation.

sphinx-hoverxref

Adds tooltips on cross references of the documentation with the content of the linked section.

sphinx-last-updated-by-git

Adds the « last updated » date at the bottom of each documentation page (obtained from the Git commit date).

sphinxext-opengraph

Generates Open Graph metadata ✨ for each page of your documentation.

grid link-type ref¶

sphinx-copybutton

Tuto sphinx

grid link-type ref¶

Tuto sphinx

Tuto sphinx.

Example from ray code use-cases.rst¶

LLMs and Gen AI¶

Large language models (LLMs) and generative AI are rapidly changing industries, and demand compute at an astonishing pace.

Ray provides a distributed compute framework for scaling these models, allowing developers to train and deploy models faster and more efficiently.

With specialized libraries for data streaming, training, fine-tuning, hyperparameter tuning, and serving, Ray simplifies the process of developing and deploying large-scale AI models.

Learn more about how Ray scales LLMs and generative AI with the following resources.

[Blog] How Ray solves common production challenges for generative AI infrastructure

[Blog] Training 175B Parameter Language Models at 1000 GPU scale with Alpa and Ray

[Blog] Faster stable diffusion fine-tuning with Ray AIR

[Blog] How to fine tune and serve LLMs simply, quickly and cost effectively using Ray + DeepSpeed + HuggingFace

[Blog] How OpenAI Uses Ray to Train Tools like ChatGPT