Skip to main content

torch-optimize

Three-role optimization loop for a PyTorch training workload: the Main Loop Agent owns the end-to-end run, authoritative profiling, and hotspot selection; a disposable Optimizer subagent reworks one hotspot per round in an isolated context; a Judge subagent renders an evidence-based verdict driving iterate/pass. All cross-role state lives in files.

Source facts

Repository
NVIDIA/recsys-examples
Last source activity
September 9, 2026 at 10:49
Detected SKILL.md language
English
Stars
363
Forks
93

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.

File Explorer
21 files

Showing SKILL.md

SKILL.md
Source instructions · Read-only preview
nametorch-optimize
descriptionThree-role optimization loop for a PyTorch training workload: the Main Loop Agent owns the end-to-end run, authoritative profiling, and hotspot selection; a disposable Optimizer subagent reworks one hotspot per round in an isolated context; a Judge subagent renders an evidence-based verdict driving iterate/pass. All cross-role state lives in files.