Skip to content

Dataset · nvidia

Data for structured format generation

nvidia/Nemotron-RL-Instruction-Following-Structured-Outputs-v2

Nvidia English dataset for training models on strict JSON, YAML, XML, TOML, and CSV output.

What's inside

An English dataset containing between 10,000 and 100,000 records for text generation based on predefined schemas. It includes two sections: direct generation in JSON, YAML, and XML, alongside more complex tasks involving data extraction and cross-format translation incorporating TOML and CSV.

What you can do with it

  • Fine-tuning LLMs to strictly follow schemas when generating JSON, YAML, or XML
  • Training models to extract structured data from unstructured text
Tasks
Text generation
Languages
English
Size
10K–100K rows
Downloads this month
1.8K
Likes
10

How to load it

from datasets import load_dataset

ds = load_dataset("nvidia/Nemotron-RL-Instruction-Following-Structured-Outputs-v2")