docs: site config, about section by lbliii · Pull Request #257 · NVIDIA-NeMo/Gym

lbliii · 2025-10-30T16:04:11Z

This is the initial PR for:

updating site config with sphinx requirements
updating docs commands via makefile shortcuts
adding an About/ section
beginning updates to the site home page (links to about items)
minor cleanup of tutorials now that we have concept pages (further work will be done here in the next PR for Getting Gtarted)
moved terminology to a resource page

All feedback is welcome, please fine-tune the language/sections as desired.

you can quickly deploy local preview of the docs by running:

cd docs
make docs-live

copy-pr-bot · 2025-10-30T16:04:14Z

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

copy-pr-bot · 2025-11-03T16:21:56Z

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

heatherlxd

After this read through I recommend we change the IA to be more onboarding focused to:

Overview
Concepts
Architecture
Ecosystem
Key Features
Release Notes
This order builds in complexity and understanding across the entire about section reading and ends with the more abstract.

I also noticed we change the order between how we talk about Agents, Models, and Resources - this happens between pages, tabs, and even sentences. We should have a consistent way to introduce this trio and make it consistent to build the mental model. While concepts starts with models I wonder if we should start with Agent since all the example sequences start with the agent? Leave order choice to the team but would recommend that the trio follows the same pattern and aligns to how they work together when possible.

Other suggestions in-line

heatherlxd · 2025-11-07T18:34:37Z

A general overall comment would be if we want to keep this order for the glossary. It is helpful to have these terms organized by category - I'm just not certain if those categories are in the ideal order? This may be a workflow question - align with when you encounter in E2E workflow or decision points.

heatherlxd · 2025-11-07T18:37:10Z

nit Should orchestrates be capitalized?

shashank3959 · 2025-11-12T07:14:22Z

+
+## The Challenge: Scaling RL Training Data
+
+Training AI agents through reinforcement learning requires massive amounts of high-quality interaction data. Teams face several obstacles:


better to explain interaction data before this

lbliii · 2025-11-12T16:59:37Z

@cwing-nvidia @bxyu-nvidia could we get approval for merge? we can continue to make changes after

…e glossary from resources; update all cross-references Signed-off-by: Lawrence Lane <llane@nvidia.com>

…erability; update concepts/index to single-page format Signed-off-by: Lawrence Lane <llane@nvidia.com>

…e; strengthen opening value props with explicit answers to 'what/why/for' Signed-off-by: Lawrence Lane <llane@nvidia.com>

…ts, collapsible sections, and concise language Signed-off-by: Lawrence Lane <llane@nvidia.com>

Signed-off-by: Lawrence Lane <llane@nvidia.com>

…ediate audience; remove basic explanations, tighten prose, focus on NeMo Gym-specific patterns Signed-off-by: Lawrence Lane <llane@nvidia.com>

Signed-off-by: Lawrence Lane <llane@nvidia.com>

@lbliii

Credit to @lbliii and see #257 --------- Signed-off-by: Brian Yu <bxyu@nvidia.com>

@lbliii

) Credit to @lbliii and see #257 --------- Signed-off-by: Brian Yu <bxyu@nvidia.com>

@lbliii

Credit to @lbliii and see #257 --------- Signed-off-by: Brian Yu <bxyu@nvidia.com>

Signed-off-by: Lawrence Lane <llane@nvidia.com>

@lbliii