nemo-curator
From NousResearch
GPU-accelerated data curation for LLM training. Supports text/image/video/audio. Features fuzzy deduplication (16× faster), quality filtering (30+ heuristics), semantic deduplication, PII redaction, NSFW detection. Scales across GPUs with RAPIDS. Use for preparing high-quality training datasets, cleaning web data, or deduplicating large corpora.
Facts
- Repository
- NousResearch/hermes-agent
- Status
- Actively maintained
- Last commit
- Source file
- optional-skills/mlops/nemo-curator/SKILL.md
Source preview
The instructions Claude Code reads when this skill runs.
# NeMo Curator - GPU-Accelerated Data Curation
NVIDIA's toolkit for preparing high-quality training data for LLMs.
## When to use NeMo Curator
**Use NeMo Curator when:**
- Preparing LLM training data from web scrapes (Common Crawl)
- Need fast deduplication (16× faster than CPU)
- Curating multi-modal datasets (text, images, video, audio)
- Filtering low-quality or toxic content
- Scaling data processing across GPU cluster
**Performance**:
- **16× faster** fuzzy deduplication (8TB RedPajama v2)
- **40% lower TCO** vs CPU alternatives
- **Near-linear scaling** across GPU nodes
**Use alternatives instead**:
- **datatrove**: CPU-based, open-source data processing
- **dolma**: Allen AI's data toolkit
- **Ray Data**: General ML data processing (no curation focus)
## Quick start
### Installation
```bash
# Text curation (CUDA 12)
uv pip install "nemo-curator[text_cuda12]"
# All modalities
uv pip install "nemo-curator[all_cuda12]"
# CPU-only (slower)
uv pip install "nemo-curator[cpu]"
```
### Basic text curation pipeline
```python
from nemo_curator import ScoreFilter, Modify
from nemo_curator.datasets import DocumentDataset
import pandas as pd
# Load data
df = pd.DataFrame({"text": ["Good document", "Bad doc", "Excellent text"]})
dataset = DocumentDataset(df)
# Quality filtering
def quality_score(doc):
return len(doc["text"].split()) > 5 # Filter short docs
filteredView full source on GitHub →Other skills
django-tdd
★ 229,918Django testing strategies with pytest-django, TDD methodology, factory_boy, mocking, coverage, and testing Django REST Framework APIs.
affaan-mupdated 15d agoMITclickhouse-io
★ 229,918ClickHouse database patterns, query optimization, analytics, and data engineering best practices for high-performance analytical workloads.
affaan-mupdated 15d agoMITlaravel-patterns
★ 229,918Patrones de arquitectura Laravel, routing/controladores, Eloquent ORM, capas de servicio, colas, eventos, caché y API resources para aplicaciones en producción.
affaan-mupdated 15d agoMITverification-loop
★ 229,918Sistema de verificación completo para sesiones de Claude Code.
affaan-mupdated 15d agoMITstrategic-compact
★ 229,918Suggests manual context compaction at logical intervals to preserve context through task phases rather than arbitrary auto-compaction.
affaan-mupdated 15d agoMITfrontend-patterns
★ 229,918Patrones de desarrollo frontend para React, Next.js, gestión de estado, optimización de rendimiento y buenas prácticas de UI.
affaan-mupdated 15d agoMIT