
Guides an agent through changing, refactoring or reviewing NVIDIA FastGen library code
- Open source
- 4K GitHub stars
- Developer tools

Guides an agent through changing, refactoring or reviewing NVIDIA FastGen library code

Guides an agent through generating and inspecting NVIDIA FastGen teacher and student

Sets up the execution context for NVIDIA FastGen commands: paths, config overrides

Sets up or repairs an NVIDIA FastGen environment, credentials and model inputs

Guides an agent through running, resuming and debugging NVIDIA FastGen training jobs

Gives agents a structured workflow for checking FastGen parity, training quality and tests

Runs and re-scores FoundationPose 6D pose pipelines on BOP-format datasets

Installs and repairs the FoundationPose perception pipeline and builds its depth engines

Lets your AI agent read and change NVIDIA display and GPU settings on this machine

Guides an agent through debugging a failing, hanging or wrong-output Holoscan CLI command

Guides Holoscan SDK 5.x Conda install requests and routes to supported alternatives

Guides installing and verifying the Holoscan SDK 5.x CUDA 13 development container

Guides installing Holoscan SDK 5.x C++ Debian packages on Ubuntu 24.04 with CUDA 13

Guides building and verifying Holoscan SDK 5.x from a chosen source ref

Guides installing Holoscan SDK 5.x CUDA 13 Python wheels into an isolated venv

Guides building, testing and packaging external Holoscan Modules with the Holoscan CLI

Assesses a host and picks a Holoscan SDK 5.x installation method

Generates fixed-format SystemVerilog FPGA_top wrappers that instantiate HOLOLINK_top from

Generates, validates and compares HSB HOLOLINK_def.svh macro files for FPGA IP

Guides choosing packetizer fields in HOLOLINK_def.svh for HSB Sensor RX data paths

Serves a converted LeRobot dataset locally so an agent can check its videos and timelines

Routes users of the Isaac for Healthcare workflow repo to the right stage skill

Scaffolds a blank, runnable Isaac for Healthcare Workflow with an idle run mode

Grades or filters workflow HDF5 episodes by visual success using a vision model

Converts recorded robot workflow HDF5 files into local LeRobot datasets for training

Expands workflow HDF5 demonstrations into action-jittered variants for training data

Replays a recorded robot workflow episode in its original simulation scene to verify it

Guides an agent to record human teleop demonstrations into workflow HDF5 datasets

Runs the full Isaac for Healthcare workflow pipeline from recording to checkpoint check

Guides an agent through fine-tuning GR00T or openpi robot policies on LeRobot data