nemo-mbridge-perf-memory-tuning

nemo-mbridge-perf-memory-tuning is a Model training & evaluation skill for AI agents, published by NVIDIA-NeMo in the Megatron-Bridge repository on GitHub. Skillspace rates nemo-mbridge-perf-memory-tuning C (77/100) and it passed the safety scan; the rating is based on static analysis only.

Techniques for reducing peak GPU memory in Megatron Bridge — expandable segments, PEFT + SP input re-gather, parallelism resizing, activation recompute, CPU offloading constraints, and common OOM fixes.

Grade
C
Quality
77 / 100
Safety
Passed
GitHub stars
798

Unproven: rated by static analysis only, not yet compared head-to-head.

This rating describes the version of the repository we crawled; it may have changed since, so go by its current files.

View source on GitHubNVIDIA-NeMo/Megatron-Bridge

Our review

Architecture
100 / 100
Maturity
Packaged: licensed, with a version or references
Security scan
Low-risk patterns
License
Apache-2.0
Adoption
Adopted by 2 other repositories, in 3 collections
Scope
Vendor-specific: built around one product. Memory tuning for the Megatron Bridge framework

What the scanner flagged

  • mentions secrets in text

Not checked yet

  • Not judged on the rubric yet (not a contender in its category)
  • No blind task eval yet
  • Sibling files (scripts, references) not fetched; security scan covers SKILL.md only

More Model training & evaluation skills

Skillspace indexes and rates this skill; it does not host or endorse it. Run third-party skills at your own risk.