Home › Skills › Megatron-Bridge › NVIDIA/Megatron-Bridge/perf-cpu-offloading

NVIDIA/Megatron-Bridge/perf-cpu-offloading

NVIDIA/skills/skills/Megatron-Bridge/perf-cpu-offloading
Agent Skill Active Apache-2.0

Validate and use CPU offloading in Megatron Bridge, including layer-level activation offloading and fractional optimizer state offloading with HybridDeviceOptimizer.

A Quality 85/100 ★ 2.9k stars Updated today Python
View on GitHub →

Installation

Install this skill (Claude Code)

# Clone and copy the skill into your project
git clone https://github.com/NVIDIA/skills.git
mkdir -p .claude/skills
cp -r skills/skills/Megatron-Bridge/perf-cpu-offloading .claude/skills/
# Or for personal use: ~/.claude/skills/

Quality score breakdown

Transparent heuristic — same formula for every entry. Total 85/100.

GitHub stars (log scale)35/40
Maintenance activity25/25
License present10/10
Official project0/10
Not archived5/5
Meaningful description5/5
Repo topics set5/5

Related in Megatron-Bridge

Agent Skill Active

Dev environment setup for Megatron Bridge — container-based development, uv package management, lockfile regeneration, adding dependencies, Slurm container usage, and common build...

★ 2.9k Python today Score 85
Agent Skill Active

Bump a pinned dependency (TransformerEngine, Megatron-LM, NRX, etc.), regenerate the lockfile, open a PR, and drive it to green by attaching a watchdog to the "CICD NeMo" workflow...

★ 2.9k Python today Score 85
Agent Skill Active

CI/CD reference for Megatron Bridge — pipeline structure, commit and PR workflow, CI failure investigation, and common failure patterns.

★ 2.9k Python today Score 85
Agent Skill Active

Code style and quality rules for Megatron Bridge — ruff configuration, naming conventions, type hints, mypy rules, docstrings, copyright headers, logging, and the code review check...

★ 2.9k Python today Score 85