mcore-run-on-slurm
mcore-run-on-slurm is an MLOps & model serving skill for AI agents, published by NVIDIA in the Megatron-LM repository on GitHub. Skillspace rates mcore-run-on-slurm B (81/100) and it passed the safety scan; a judge model has also reviewed it.
How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.
- Grade
- B
- Quality
- 81 / 100
- Safety
- Passed
- GitHub stars
- 17,075
Placed #2 of 6 in the judged category final for MLOps & model serving.
Our two independent classifiers differ on this skill's category, so treat the placement as provisional.
This rating describes the version of the repository we crawled; it may have changed since, so go by its current files.
Our review
- Architecture
- 100 / 100
- Maturity
- Packaged: licensed, with a version or references
- Security scan
- No issues found
- License
- Not declared. Check the repository before reusing it.
- Adoption
- Adopted by 0 other repositories, in 0 collections
- Scope
- Vendor-specific: built around one product. Documents launching the public Megatron-LM framework on SLURM clusters.
Judge review
Placed #2 of 6 in the judged category final for MLOps & model serving.
- Trigger
82 82 - Instructions
93 93 - Specificity
95 95 - Safety
80 80 - Disclosure
68 68 - Maintenance
88 88
A fully runnable sbatch script plus a hardware-specific CUDA_DEVICE_MAX_CONNECTIONS table that correctly notes misconfiguration fails as an assertion, not a silent deadlock.
Not checked yet
- No blind task eval yet
- Sibling files (scripts, references) not fetched; security scan covers SKILL.md only
Skillspace indexes and rates this skill; it does not host or endorse it. Run third-party skills at your own risk.