home/categories/machine-learning/orchestra-research-ai-research-skills-11-evaluation-nemo-evaluator-skill-md
machine-learningdata-ai
nemo-evaluator-sdk
Evaluates LLMs across 100+ benchmarks from 18+ harnesses (MMLU, HumanEval, GSM8K, safety, VLM) with multi-backend execution. Use when needing scalable evaluation on local Docker, Slurm HPC, or cloud platforms. NVIDIA's enterprise-grade platform with container-first architecture for reproducible benchmarking.
maintainer
Orchestra-Research
์
๋ฐ์ดํธ๋จ 2026. 1. 15.
์คํ
6563
ํฌํฌ
515
quick start
Installation and usage
Evaluates LLMs across 100+ benchmarks from 18+ harnesses (MMLU, HumanEval, GSM8K, safety, VLM) with multi-backend execution. Use when needing scalable evaluation on local Docker, Slurm HPC, or cloud platforms. NVIDIA's enterprise-grade platform with container-first architecture for reproducible benchmarking.
์ค์น
$ install --globalskills.sh
์ฌ์ฉ๋ฒ
์ค์น ํ ํฐ๋ฏธ๋์์ ๋ค์ ๋ช ๋ น์ ์คํํ์ฌ ์ด ์คํฌ์ ์ฌ์ฉํ ์ ์์ต๋๋ค:
skills use nemo-evaluator-sdk